diff --git a/CMakeLists.txt b/CMakeLists.txt
index a7ff612..ab1b3f7 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -1,11 +1,11 @@
 cmake_minimum_required (VERSION 3.5)
 project (qcint C)
-set(qcint_VERSION_MAJOR "5")
-set(qcint_VERSION_MINOR "4")
+set(qcint_VERSION_MAJOR "6")
+set(qcint_VERSION_MINOR "0")
 set(qcint_VERSION_PATCH "0")
 set(qcint_VERSION_TWEAK "0")
 set(qcint_VERSION "${qcint_VERSION_MAJOR}.${qcint_VERSION_MINOR}.${qcint_VERSION_PATCH}")
-set(qcint_SOVERSION "${qcint_VERSION_MAJOR}")
+set(cint_SOVERSION "${qcint_VERSION_MAJOR}")
 
 #set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -g -O2 -DNDEBUG")
 if ("${CMAKE_BUILD_TYPE}" STREQUAL "")
@@ -71,7 +71,7 @@ set(cintSrc
   src/g1e.c src/g2e.c src/g2e_simd1.c src/g3c1e.c src/g3c2e.c
   src/gout2e.c src/misc.c src/optimizer.c
   src/fmt.c src/rys_wheeler.c src/eigh.c src/rys_roots.c src/find_roots.c
-  src/polyfits.c src/sr_rys_polyfits.c
+  src/polyfits.c
   src/cint1e_a.c src/cint3c1e_a.c
   src/cint1e_grids.c src/g1e_grids.c
   src/autocode/breit1.c src/autocode/dkb.c src/autocode/gaunt1.c
@@ -84,17 +84,17 @@ set(cintSrc
 if(WITH_RANGE_COULOMB)
 # defined in config.h
 #  add_definitions(-DWITH_RANGE_COULOMB)
-  message("Enabled WITH_RANGE_COULOMB")
-  if(WITH_POLYNOMIAL_FIT)
-    add_definitions(-DWITH_POLYNOMIAL_FIT)
-    message("Enabled WITH_POLYNOMIAL_FIT")
-  endif(WITH_POLYNOMIAL_FIT)
+#  message("Enabled WITH_RANGE_COULOMB")
 endif(WITH_RANGE_COULOMB)
 
+if(WITH_POLYNOMIAL_FIT)
+  set(cintSrc ${cintSrc} src/sr_rys_polyfits.c)
+  add_definitions(-DWITH_POLYNOMIAL_FIT)
+  message("Enabled WITH_POLYNOMIAL_FIT")
+endif(WITH_POLYNOMIAL_FIT)
+
 if(WITH_COULOMB_ERF)
-  set(cintSrc ${cintSrc} src/g2e_coulerf.c src/cint2e_coulerf.c)
-  add_definitions(-DWITH_COULOMB_ERF)
-  message("Enabled WITH_COULOMB_ERF")
+  message("WITH_COULOMB_ERF is deprecated since v6.0")
 endif(WITH_COULOMB_ERF)
 
 if(WITH_F12)
@@ -103,14 +103,6 @@ if(WITH_F12)
   message("Enabled WITH_F12")
 endif(WITH_F12)
 
-if(WITH_GTG)
-  set(cintSrc ${cintSrc} src/g2e_gtg.c src/cint2e_gtg.c src/cint3c2e_gtg.c
-    src/cint2c2e_gtg.c)
-  add_definitions(-DWITH_GTG)
-  message("Enabled WITH_GTG")
-  message("Enabled WITH_GTG. Note there are bugs in gtg type integrals")
-endif(WITH_GTG)
-
 if(PYPZPX)
   add_definitions(-DPYPZPX)
   message("P orbitals convention (py, pz, px)")
diff --git a/include/cint.h.in b/include/cint.h.in
index 7eeabd9..2c4c918 100644
--- a/include/cint.h.in
+++ b/include/cint.h.in
@@ -22,7 +22,8 @@
  * Parameters and function prototypes for libcint.
  */
 
-#define QCINT_VERSION    @qcint_VERSION@
+#define QCINT_VERSION           @qcint_VERSION@
+#define CINT_SOVERSION          @cint_SOVERSION
 
 #cmakedefine CACHE_SIZE_I8
 #ifdef CACHE_SIZE_I8
@@ -203,7 +204,7 @@ typedef struct {
         union {int nfk; int grids_offset;};
         union {int nfl; int ngrids;};
         int nf;  // = nfi*nfj*nfk*nfl;
-        int _padding;
+        int rys_order; // = nrys_roots for regular ERIs. can be nrys_roots/2 for SR ERIs
         int x_ctr[4];
 
         int gbits;
diff --git a/src/cint2c2e.c b/src/cint2c2e.c
index a005aab..6cf98aa 100644
--- a/src/cint2c2e.c
+++ b/src/cint2c2e.c
@@ -21,8 +21,8 @@
 #include <stdlib.h>
 #include <stdio.h>
 #include <string.h>
-#include <assert.h>
 #include <math.h>
+#include <assert.h>
 #include "cint_bas.h"
 #include "misc.h"
 #include "g2e.h"
@@ -65,15 +65,38 @@
                 (*(fp2c[i]))(gctr[it], gprim[i], coeff[it]+im, \
                              ngp[it], x_prim[it], x_ctr[it], \
                              non0ctr[it][im], non0idx[it]+im*x_ctr[it]); \
+                empty_overall = 0; \
+        } \
+        cum = 0; \
+        np2c = 0;
+
+#define POP_PRIM2CTR_AND_SET0 \
+        for (i = 0; i < np2c; i++) { \
+                it = shltyp[i]; \
+                if (it != SHLTYPi) { \
+                        im = iprim[i]; \
+                        (*(fp2c[i]))(gctr[i], gprim[i], coeff[it]+im, \
+                                     ngp[it], x_prim[it], x_ctr[it], \
+                                     non0ctr[it][im], non0idx[it]+im*x_ctr[it]); \
+                        empty_overall = 0; \
+                } else if (fp2c[i] == CINTiprim_to_ctr_0) { \
+                        double *pout = gctr[it]; \
+                        for (int k = 0; k < nf; k++) { \
+                                pout[k] = 0.; \
+                        } \
+                } \
         } \
         cum = 0; \
         np2c = 0;
 
 #define PUSH \
         if (cum == SIMDD) { \
-                (*envs->f_g0_2e)(g, cutoff, &bc, envs, cum); \
-                (*envs->f_gout)(gout, g, idx, envs); \
-                POP_PRIM2CTR; \
+                if ((*envs->f_g0_2e)(g, cutoff, &bc, envs, cum)) { \
+                        (*envs->f_gout)(gout, g, idx, envs); \
+                        POP_PRIM2CTR; \
+                } else { \
+                        POP_PRIM2CTR_AND_SET0; \
+                } \
         } \
         envs->ai[cum] = ai[ip]; \
         envs->ak[cum] = ak[kp]; \
@@ -111,26 +134,30 @@
 
 #define RUN_REST \
         if (cum == 1) { \
-                (*envs->f_g0_2e_simd1)(g, cutoff, &bc, envs, 0); \
-                (*envs->f_gout_simd1)(gout, g, idx, envs); \
+                if ((*envs->f_g0_2e_simd1)(g, cutoff, &bc, envs, 0)) { \
+                        (*envs->f_gout_simd1)(gout, g, idx, envs); \
+                        POP_PRIM2CTR; \
+                } else { \
+                        POP_PRIM2CTR_AND_SET0; \
+                } \
         } else if (cum > 1) { \
-                r1 = MM_SET1(1.); \
-                for (i = 0; i < envs->nrys_roots; i++) { \
-                        MM_STORE(bc.u+i*SIMDD, r1); \
-                        MM_STORE(bc.w+i*SIMDD, r1); \
+                if ((*envs->f_g0_2e)(g, cutoff, &bc, envs, cum)) { \
+                        (*envs->f_gout)(gout, g, idx, envs); \
+                        POP_PRIM2CTR; \
+                } else { \
+                        POP_PRIM2CTR_AND_SET0; \
                 } \
-                (*envs->f_g0_2e)(g, cutoff, &bc, envs, cum); \
-                (*envs->f_gout)(gout, g, idx, envs); \
-        } \
-        POP_PRIM2CTR;
+        } else { \
+                assert(np2c == 0); \
+        }
 
 #define TRANSPOSE(a) \
         if (*empty) { \
                 CINTdmat_transpose(out, a, nf*nc, n_comp); \
+                *empty = 0; \
         } else { \
                 CINTdplus_transpose(out, a, nf*nc, n_comp); \
-        } \
-        *empty = 0;
+        }
 
 int CINT2c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty)
 {
@@ -158,6 +185,7 @@ int CINT2c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
         int _empty[2] = {1, 1};
         int *iempty = _empty + 0;
         int *kempty = _empty + 1;
+        int empty_overall = 1;
         int ngp[2];
         ngp[0] = nf * n_comp;
         ngp[1] = ngp[0] * i_ctr;
@@ -171,7 +199,7 @@ int CINT2c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
         if (n_comp == 1) {
                 MALLOC_INSTACK(g1, lenk);
                 gctr[SHLTYPk] = out;
-                kempty = empty;
+                *kempty = *empty;
         } else {
                 MALLOC_INSTACK(gctr[SHLTYPk], lenk);
                 g1 = out;
@@ -195,8 +223,7 @@ int CINT2c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
         CINTOpt_non0coeff_byshell(non0idxk, non0ctrk, coeff[1], k_prim, k_ctr);
         int *non0ctr[2] = {non0ctri, non0ctrk};
         int *non0idx[2] = {non0idxi, non0idxk};
-        double common_factor = envs->common_factor * (M_PI*M_PI*M_PI)*2/SQRTPI
-                * CINTcommon_fac_sp(envs->i_l) * CINTcommon_fac_sp(envs->k_l);
+        double common_factor = envs->common_factor;
         INITSIMD;
 
         for (kp = 0; kp < k_prim; kp++) {
@@ -215,11 +242,12 @@ int CINT2c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
         } // end loop k_prim
         RUN_REST;
 
-        if (n_comp > 1 && !*kempty) {
+        if (n_comp > 1 && !empty_overall) {
                 int nc = i_ctr * k_ctr;
                 TRANSPOSE(gctr[SHLTYPk]);
         }
-        return !*empty;
+        *empty &= empty_overall;
+        return !empty_overall;
 }
 
 
@@ -249,6 +277,7 @@ int CINT2c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         int _empty[2] = {1, 1};
         int *iempty = _empty + 0;
         int *kempty = _empty + 1;
+        int empty_overall = 1;
         int ngp[2];
         ngp[0] = nf * n_comp;
         ngp[1] = ngp[0] * i_ctr;
@@ -262,7 +291,7 @@ int CINT2c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         if (n_comp == 1) {
                 MALLOC_INSTACK(g1, lenk);
                 gctr[SHLTYPk] = out;
-                kempty = empty;
+                *kempty = *empty;
         } else {
                 MALLOC_INSTACK(gctr[SHLTYPk], lenk);
                 g1 = out;
@@ -273,8 +302,7 @@ int CINT2c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
 
         ALIGNMM Rys2eT bc;
         ALIGNMM double cutoff[SIMDD];
-        double common_factor = envs->common_factor * (M_PI*M_PI*M_PI)*2/SQRTPI
-                * CINTcommon_fac_sp(envs->i_l) * CINTcommon_fac_sp(envs->k_l);
+        double common_factor = envs->common_factor;
 
         CINTOpt *opt = envs->opt;
         int *idx = opt->index_xyz_array[envs->i_l*LMAX1+envs->k_l];
@@ -299,11 +327,12 @@ int CINT2c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         } // end loop k_prim
         RUN_REST;
 
-        if (n_comp > 1 && !*kempty) {
+        if (n_comp > 1 && !empty_overall) {
                 int nc = i_ctr * k_ctr;
                 TRANSPOSE(gctr[SHLTYPk]);
         }
-        return !*empty;
+        *empty &= empty_overall;
+        return !empty_overall;
 }
 
 void CINTinit_int2c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
@@ -331,7 +360,12 @@ void CINTinit_int2c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
         envs->nfk = (envs->k_l+1)*(envs->k_l+2)/2;
         envs->nfl = 1;
         envs->nf = envs->nfi * envs->nfk;
-        envs->common_factor = 1;
+
+        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
+        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
+
+        envs->common_factor = (M_PI*M_PI*M_PI)*2/SQRTPI
+                * CINTcommon_fac_sp(envs->i_l) * CINTcommon_fac_sp(envs->k_l);
         if (env[PTR_EXPCUTOFF] == 0) {
                 envs->expcutoff = EXPCUTOFF;
         } else {
@@ -347,20 +381,21 @@ void CINTinit_int2c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
         envs->lj_ceil = 0;
         envs->lk_ceil = envs->k_l + ng[KINC];
         envs->ll_ceil = 0;
-
-        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
-        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
-
-        int nroots = (envs->li_ceil + envs->lk_ceil)/2 + 1;
-        envs->nrys_roots = nroots;
-        assert(nroots < MXRYSROOTS);
+        int rys_order =(envs->li_ceil + envs->lk_ceil)/2 + 1;
+        int nrys_roots = rys_order;
+        double omega = env[PTR_RANGE_OMEGA];
+        if (omega < 0 && rys_order <= 3) {
+                nrys_roots *= 2;
+        }
+        envs->rys_order = rys_order;
+        envs->nrys_roots = nrys_roots;
 
         int dli = envs->li_ceil + 1;
         int dlk = envs->lk_ceil + 1;
-        envs->g_stride_i = nroots;
-        envs->g_stride_k = nroots * dli;
+        envs->g_stride_i = nrys_roots;
+        envs->g_stride_k = nrys_roots * dli;
         envs->g_stride_l = envs->g_stride_k;
-        envs->g_size     = nroots * dli * dlk;
+        envs->g_size     = nrys_roots * dli * dlk;
 
         MM_STORE(envs->aj, MM_SET1(0.));
         MM_STORE(envs->al, MM_SET1(0.));
@@ -381,9 +416,13 @@ void CINTinit_int2c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
         envs->rx_in_rklrx = envs->rk;
         envs->rx_in_rijrx = envs->ri;
 
-        if (nroots <= 2) {
+        if (rys_order <= 2) {
                 envs->f_g0_2d4d = &CINTg0_2e_2d4d_unrolled;
                 envs->f_g0_2d4d_simd1 = &CINTg0_2e_2d4d_unrolled_simd1;
+                if (rys_order != nrys_roots) {
+                        envs->f_g0_2d4d = &CINTsrg0_2e_2d4d_unrolled;
+                        envs->f_g0_2d4d_simd1 = &CINTsrg0_2e_2d4d_unrolled_simd1;
+                }
         } else {
                 envs->f_g0_2d4d = &CINTg0_2e_2d;
                 envs->f_g0_2d4d_simd1 = &CINTg0_2e_2d_simd1;
diff --git a/src/cint2c2e_gtg.c b/src/cint2c2e_gtg.c
deleted file mode 100644
index 96ed138..0000000
--- a/src/cint2c2e_gtg.c
+++ /dev/null
@@ -1,139 +0,0 @@
-/*
- * Copyright (C) 2013-  Qiming Sun <osirpt.sun@gmail.com>
- */
-
-#include <stdlib.h>
-#include <stdio.h>
-#include "cint_bas.h"
-#include "g2e.h"
-#include "optimizer.h"
-#include "cint2e.h"
-#include "misc.h"
-#include "cart2sph.h"
-
-
-#ifdef WITH_GTG
-int CINTg0_2e_gtg(double *g, Rys2eT *bc, CINTEnvVars *envs, int count);
-int CINTg0_2e_gtg_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs, int idsimd);
-
-void CINTinit_int2c2e_gtg_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
-                                  int *atm, int natm, int *bas, int nbas, double *env)
-{
-        envs->natm = natm;
-        envs->nbas = nbas;
-        envs->atm = atm;
-        envs->bas = bas;
-        envs->env = env;
-        envs->shls = shls;
-
-        int i_sh = shls[0];
-        int k_sh = shls[1];
-        envs->i_l = bas(ANG_OF, i_sh);
-        envs->j_l = 0;
-        envs->k_l = bas(ANG_OF, k_sh);
-        envs->l_l = 0;
-        envs->x_ctr[0] = bas(NCTR_OF, i_sh);
-        envs->x_ctr[1] = bas(NCTR_OF, k_sh);
-        envs->x_ctr[2] = 1;
-        envs->x_ctr[3] = 1;
-        envs->nfi = (envs->i_l+1)*(envs->i_l+2)/2;
-        envs->nfj = 1;
-        envs->nfk = (envs->k_l+1)*(envs->k_l+2)/2;
-        envs->nfl = 1;
-        envs->nf = envs->nfi * envs->nfk;
-        envs->common_factor = SQRTPI * .5;
-        if (env[PTR_EXPCUTOFF] == 0) {
-                envs->expcutoff = EXPCUTOFF;
-        } else {
-                envs->expcutoff = MAX(MIN_EXPCUTOFF, env[PTR_EXPCUTOFF]);
-        }
-
-        envs->gbits = ng[GSHIFT];
-        envs->ncomp_e1 = ng[POS_E1];
-        envs->ncomp_e2 = ng[POS_E2];
-        envs->ncomp_tensor = ng[TENSOR];
-
-        envs->li_ceil = envs->i_l + ng[IINC];
-        envs->lj_ceil = 0;
-        envs->lk_ceil = envs->k_l + ng[KINC];
-        envs->ll_ceil = 0;
-
-        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
-        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
-
-        envs->nrys_roots = 1;
-
-        int dli = envs->li_ceil + 1;
-        int dlk = envs->lk_ceil + 1;
-        envs->g_stride_i = 1;
-        envs->g_stride_k = dli;
-        envs->g_stride_l = envs->g_stride_k;
-        envs->g_size     = dli * dlk;
-
-        MM_STORE(envs->aj, MM_SET1(0.));
-        MM_STORE(envs->al, MM_SET1(0.));
-        MM_STORE(envs->rij+0*SIMDD, MM_SET1(envs->ri[0]));
-        MM_STORE(envs->rij+1*SIMDD, MM_SET1(envs->ri[1]));
-        MM_STORE(envs->rij+2*SIMDD, MM_SET1(envs->ri[2]));
-        MM_STORE(envs->rkl+0*SIMDD, MM_SET1(envs->rk[0]));
-        MM_STORE(envs->rkl+1*SIMDD, MM_SET1(envs->rk[1]));
-        MM_STORE(envs->rkl+2*SIMDD, MM_SET1(envs->rk[2]));
-        envs->g2d_ijmax = envs->g_stride_i;
-        envs->g2d_klmax = envs->g_stride_k;
-        envs->rkrl[0] = envs->rk[0];
-        envs->rkrl[1] = envs->rk[1];
-        envs->rkrl[2] = envs->rk[2];
-        envs->rirj[0] = envs->ri[0];
-        envs->rirj[1] = envs->ri[1];
-        envs->rirj[2] = envs->ri[2];
-        envs->rx_in_rklrx = envs->rk;
-        envs->rx_in_rijrx = envs->ri;
-
-        envs->f_g0_2d4d = &CINTg0_2e_2d;
-        envs->f_g0_2d4d_simd1 = &CINTg0_2e_2d_simd1;
-        envs->f_g0_2e = &CINTg0_2e_gtg;
-        envs->f_g0_2e_simd1 = &CINTg0_2e_gtg_simd1;
-
-        // for CINTg2c_index_xyz and c2s_sph_1e function
-        envs->j_l = envs->k_l;
-        envs->nfj = envs->nfk;
-        envs->g_stride_j = envs->g_stride_k;
-}
-#endif
-
-CACHE_SIZE_T int2c2e_gtg_sph(double *out, int *dims, int *shls, int *atm, int natm,
-                  int *bas, int nbas, double *env, CINTOpt *opt, double *cache)
-{
-        int ng[] = {0, 0, 0, 0, 0, 1, 1, 1};
-        CINTEnvVars envs;
-        CINTinit_int2c2e_gtg_EnvVars(&envs, ng, shls, atm, natm, bas, nbas, env);
-        envs.f_gout = &CINTgout2e;
-        envs.f_gout_simd1 = &CINTgout2e_simd1;
-        return CINT2c2e_drv(out, dims, &envs, opt, cache, &c2s_sph_1e);
-}
-void int2c2e_gtg_optimizer(CINTOpt **opt, int *atm, int natm,
-                         int *bas, int nbas, double *env)
-{
-        int ng[] = {0, 0, 0, 0, 0, 1, 1, 1};
-        CINTall_2c2e_gtg_optimizer(opt, ng, atm, natm, bas, nbas, env);
-}
-
-CACHE_SIZE_T int2c2e_gtg_cart(double *out, int *dims, int *shls, int *atm, int natm,
-                     int *bas, int nbas, double *env, CINTOpt *opt, double *cache)
-{
-        int ng[] = {0, 0, 0, 0, 0, 1, 1, 1};
-        CINTEnvVars envs;
-        CINTinit_int2c2e_gtg_EnvVars(&envs, ng, shls, atm, natm, bas, nbas, env);
-        envs.f_gout = &CINTgout2e;
-        envs.f_gout_simd1 = &CINTgout2e_simd1;
-        return CINT2c2e_drv(out, dims, &envs, opt, cache, &c2s_cart_1e);
-}
-
-CACHE_SIZE_T int2c2e_gtg_spinor(double complex *out, int *dims, int *shls, int *atm, int natm,
-                     int *bas, int nbas, double *env, CINTOpt *opt, double *cache)
-{
-        fprintf(stderr, "int2c2e_gtg_spinor not implemented\n");
-}
-
-ALL_CINT(int2c2e_gtg)
-
diff --git a/src/cint2e.c b/src/cint2e.c
index 2e8f69e..6c6b4ed 100644
--- a/src/cint2e.c
+++ b/src/cint2e.c
@@ -22,11 +22,12 @@
 #include <stdio.h>
 #include <stdint.h>
 #include <math.h>
+#include <assert.h>
 #include "cint_bas.h"
-#include "misc.h"
 #include "g2e.h"
 #include "optimizer.h"
 #include "cint2e.h"
+#include "misc.h"
 #include "cart2sph.h"
 #include "c2f.h"
 
@@ -64,14 +65,12 @@
                         fp2c[np2c] = CINTprim_to_ctr_0; \
                         shltyp[np2c] = SHLTYP##csymb; \
                         gprim[np2c] = gctr[SHLTYP##psymb]; \
-                        gp2c [np2c] = gctr[SHLTYP##csymb]; \
                         iprim[np2c] = csymb##p; \
                         np2c++; \
                 } else if (non0ctr##csymb[csymb##p] > 1) { \
                         fp2c[np2c] = CINTprim_to_ctr_1; \
                         shltyp[np2c] = SHLTYP##csymb; \
                         gprim[np2c] = gctr[SHLTYP##psymb]; \
-                        gp2c [np2c] = gctr[SHLTYP##csymb]; \
                         iprim[np2c] = csymb##p; \
                         np2c++; \
                 } \
@@ -82,9 +81,29 @@
         for (i = 0; i < np2c; i++) { \
                 it = shltyp[i]; \
                 im = iprim[i]; \
-                (*(fp2c[i]))(gp2c[i], gprim[i], coeff[it]+im, \
+                (*(fp2c[i]))(gctr[it], gprim[i], coeff[it]+im, \
                              ngp[it], x_prim[it], x_ctr[it], \
                              non0ctr[it][im], non0idx[it]+im*x_ctr[it]); \
+                empty_overall = 0; \
+        } \
+        cum = 0; \
+        np2c = 0;
+
+#define POP_PRIM2CTR_AND_SET0 \
+        for (i = 0; i < np2c; i++) { \
+                it = shltyp[i]; \
+                if (it != SHLTYPi) { \
+                        im = iprim[i]; \
+                        (*(fp2c[i]))(gctr[it], gprim[i], coeff[it]+im, \
+                                     ngp[it], x_prim[it], x_ctr[it], \
+                                     non0ctr[it][im], non0idx[it]+im*x_ctr[it]); \
+                        empty_overall = 0; \
+                } else if (fp2c[i] == CINTiprim_to_ctr_0) { \
+                        double *pout = gctr[it]; \
+                        for (int k = 0; k < ngp[1]; k++) { \
+                                pout[k] = 0.; \
+                        } \
+                } \
         } \
         cum = 0; \
         np2c = 0;
@@ -99,9 +118,12 @@
 
 #define PUSH(RIJ, RKL) \
         if (cum == SIMDD) { \
-                (*envs->f_g0_2e)(g, cutoff, &bc, envs, cum); \
-                (*envs->f_gout)(gout, g, idx, envs); \
-                POP_PRIM2CTR; \
+                if ((*envs->f_g0_2e)(g, cutoff, &bc, envs, cum)) { \
+                        (*envs->f_gout)(gout, g, idx, envs); \
+                        POP_PRIM2CTR; \
+                } else { \
+                        POP_PRIM2CTR_AND_SET0; \
+                } \
         } \
         envs->ai[cum] = ai[ip]; \
         envs->aj[cum] = aj[jp]; \
@@ -123,9 +145,8 @@
                 fp2c[np2c] = CINTiprim_to_ctr_1; \
         } \
         gprim[np2c] = gout + cum * ngp[0]; \
-        gp2c [np2c] = gctr[SHLTYPi]; \
         iprim[np2c] = ip; \
-        shltyp[np2c] = 0; \
+        shltyp[np2c] = SHLTYPi; \
         cum++; \
         np2c++;
 
@@ -139,7 +160,6 @@
         } \
         int cum = 0; \
         int np2c = 0; \
-        double *gp2c [SIMDD*4]; \
         double *gprim[SIMDD*4]; \
         int shltyp[SIMDD*4]; \
         int iprim[SIMDD*4]; \
@@ -152,18 +172,22 @@
 
 #define RUN_REST \
         if (cum == 1) { \
-                (*envs->f_g0_2e_simd1)(g, cutoff, &bc, envs, 0); \
-                (*envs->f_gout_simd1)(gout, g, idx, envs); \
+                if ((*envs->f_g0_2e_simd1)(g, cutoff, &bc, envs, 0)) { \
+                        (*envs->f_gout_simd1)(gout, g, idx, envs); \
+                        POP_PRIM2CTR; \
+                } else { \
+                        POP_PRIM2CTR_AND_SET0; \
+                } \
         } else if (cum > 1) { \
-                r1 = MM_SET1(1.); \
-                for (i = 0; i < envs->nrys_roots; i++) { \
-                        MM_STORE(bc.u+i*SIMDD, r1); \
-                        MM_STORE(bc.w+i*SIMDD, r1); \
+                if ((*envs->f_g0_2e)(g, cutoff, &bc, envs, cum)) { \
+                        (*envs->f_gout)(gout, g, idx, envs); \
+                        POP_PRIM2CTR; \
+                } else { \
+                        POP_PRIM2CTR_AND_SET0; \
                 } \
-                (*envs->f_g0_2e)(g, cutoff, &bc, envs, cum); \
-                (*envs->f_gout)(gout, g, idx, envs); \
-        } \
-        POP_PRIM2CTR
+        } else { \
+                assert(np2c == 0); \
+        }
 
 int CINT2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty)
 {
@@ -194,9 +218,9 @@ int CINT2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty)
         double *cj = env + bas(PTR_COEFF, j_sh);
         double *ck = env + bas(PTR_COEFF, k_sh);
         double *cl = env + bas(PTR_COEFF, l_sh);
-        double *coeff[4] = {ci, cj, ck, cl};
         double expcutoff = envs->expcutoff;
         double rr_ij = SQUARE(envs->rirj);
+        double rr_kl = SQUARE(envs->rkrl);
         double *log_maxci, *log_maxcj, *log_maxck, *log_maxcl;
         PairData *pdata_base, *pdata_ij;
         MALLOC_DATA_INSTACK(log_maxci, i_prim+j_prim+k_prim+l_prim);
@@ -223,11 +247,47 @@ int CINT2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty)
         int *jempty = _empty + 1;
         int *kempty = _empty + 2;
         int *lempty = _empty + 3;
+        int empty_overall = 1;
+        int lkl = envs->lk_ceil + envs->ll_ceil;
+        double akl, ekl, expijkl, ccekl, log_rr_kl, eijcutoff;
+
+        akl = ak[k_prim-1] + al[l_prim-1];
+        log_rr_kl = 1.7 - 1.5 * approx_log(akl);
+        double omega = env[PTR_RANGE_OMEGA];
+        if (omega < 0) {
+                // Normally the factor
+                //    (aj*d/aij+theta*R)^li * (ai*d/aij+theta*R)^lj * pi^1.5/aij^{(li+lj+3)/2}
+                // is a good approximation for polynomial parts in SR-ERIs.
+                //    <~ (aj*d/aij+theta*R)^li * (ai*d/aij+theta*R)^lj * (pi/aij)^1.5
+                //    <~ (d+theta*R)^li * (d+theta*R)^lj * (pi/aij)^1.5
+                if (envs->rys_order > 1) {
+                        double r_guess = 8.;
+                        double omega2 = omega * omega;
+                        int lij = envs->li_ceil + envs->lj_ceil;
+                        if (lij > 0) {
+                                double aij = ai[i_prim-1] + aj[j_prim-1];
+                                double dist_ij = sqrt(rr_ij);
+                                double theta = omega2 / (omega2 + aij);
+                                expcutoff += lij * approx_log(
+                                        (dist_ij+theta*r_guess+1.)/(dist_ij+1.));
+                        }
+                        if (lkl > 0) {
+                                double theta = omega2 / (omega2 + akl);
+                                log_rr_kl += lkl * approx_log(
+                                        sqrt(rr_kl) + theta*r_guess + 1.);
+                        }
+                }
+        } else {
+                if (lkl > 0) {
+                        log_rr_kl += lkl * approx_log(sqrt(rr_kl) + 1.);
+                }
+        }
 
         int *idx;
         MALLOC_DATA_INSTACK(idx, nf * 3);
         CINTg4c_index_xyz(idx, envs);
 
+        double *coeff[4] = {ci, cj, ck, cl};
         int *non0ctri, *non0ctrj, *non0ctrk, *non0ctrl;
         int *non0idxi, *non0idxj, *non0idxk, *non0idxl;
         MALLOC_DATA_INSTACK(non0ctri, i_prim+j_prim+k_prim+l_prim+i_prim*i_ctr+j_prim*j_ctr+k_prim*k_ctr+l_prim*l_ctr);
@@ -244,9 +304,7 @@ int CINT2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty)
         CINTOpt_non0coeff_byshell(non0idxl, non0ctrl, coeff[3], l_prim, l_ctr);
         int *non0ctr[4] = {non0ctri, non0ctrj, non0ctrk, non0ctrl};
         int *non0idx[4] = {non0idxi, non0idxj, non0idxk, non0idxl};
-        double common_factor = envs->common_factor * (M_PI*M_PI*M_PI)*2/SQRTPI
-                * CINTcommon_fac_sp(envs->i_l) * CINTcommon_fac_sp(envs->j_l)
-                * CINTcommon_fac_sp(envs->k_l) * CINTcommon_fac_sp(envs->l_l);
+        double common_factor = envs->common_factor;
 
         size_t ngp[4];
         ngp[0] = nf * n_comp;
@@ -268,7 +326,7 @@ int CINT2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty)
                 // patch SIMDD*2 for leni, lenj, lenk with s functions
                 MALLOC_INSTACK(g1, lenl+SIMDD*2);
                 bufctr[SHLTYPl] = out;
-                lempty = empty;
+                *lempty = *empty;
         } else {
                 // enlarge sizeo of out by SIMDD*2, for leni, lenj, lenk with s functions
                 cache += SIMDD*2;
@@ -285,46 +343,6 @@ int CINT2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty)
         ALIGNMM Rys2eT bc;
         ALIGNMM double cutoff[SIMDD];
         ALIGNMM double rkl[4];
-        double rr_kl = SQUARE(envs->rkrl);
-        double log_rr_kl, akl, ekl, expijkl, ccekl, eijcutoff;
-        akl = ak[k_prim-1] + al[l_prim-1];
-        log_rr_kl = 1.7 - 1.5 * approx_log(akl);
-        int lkl = envs->lk_ceil + envs->ll_ceil;
-#ifdef WITH_RANGE_COULOMB
-        double omega = env[PTR_RANGE_OMEGA];
-        if (omega < 0) {
-                // Normally the factor
-                //    (aj*d/aij+theta*R)^li * (ai*d/aij+theta*R)^lj * pi^1.5/aij^{(li+lj+3)/2}
-                // is a good approximation for polynomial parts in SR-ERIs.
-                //    <~ (aj*d/aij+theta*R)^li * (ai*d/aij+theta*R)^lj * (pi/aij)^1.5
-                //    <~ (d+theta*R)^li * (d+theta*R)^lj * (pi/aij)^1.5
-                if (envs->nrys_roots > 1) {
-                        double r_guess = 8.;
-                        double omega2 = omega * omega;
-                        int lij = envs->li_ceil + envs->lj_ceil;
-                        if (lij > 0) {
-                                double aij = ai[i_prim-1] + aj[j_prim-1];
-                                double dist_ij = sqrt(rr_ij);
-                                double theta = omega2 / (omega2 + aij);
-                                expcutoff += lij * approx_log(
-                                        (dist_ij+theta*r_guess+1.)/(dist_ij+1.));
-                        }
-                        if (lkl > 0) {
-                                double theta = omega2 / (omega2 + akl);
-                                log_rr_kl += lkl * approx_log(
-                                        sqrt(rr_kl) + theta*r_guess + 1.);
-                        }
-                }
-        } else {
-                if (lkl > 0) {
-                        log_rr_kl += lkl * approx_log(sqrt(rr_kl) + 1.);
-                }
-        }
-#else
-        if (lkl > 0) {
-                log_rr_kl += lkl * approx_log(sqrt(rr_kl) + 1.);
-        }
-#endif
 
         INITSIMD;
 
@@ -370,11 +388,12 @@ k_contracted: ;
         } // end loop l_prim
         RUN_REST;
 
-        if (n_comp > 1 && !*lempty) {
+        if (n_comp > 1 && !empty_overall) {
                 int nc = i_ctr * j_ctr * k_ctr * l_ctr;
                 TRANSPOSE(gctr[SHLTYPl]);
         }
-        return !*empty;
+        *empty &= empty_overall;
+        return !empty_overall;
 }
 
 
@@ -411,7 +430,6 @@ int CINT2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         double *cj = env + bas(PTR_COEFF, j_sh);
         double *ck = env + bas(PTR_COEFF, k_sh);
         double *cl = env + bas(PTR_COEFF, l_sh);
-        double *coeff[4] = {ci, cj, ck, cl};
         double expcutoff = envs->expcutoff;
         PairData *_pdata_ij, *_pdata_kl, *pdata_kl, *pdata_ij;
         if (opt->pairdata != NULL) {
@@ -446,6 +464,7 @@ int CINT2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         int *jempty = _empty + 1;
         int *kempty = _empty + 2;
         int *lempty = _empty + 3;
+        int empty_overall = 1;
 
         int *idx = opt->index_xyz_array[envs->i_l*LMAX1*LMAX1*LMAX1
                                        +envs->j_l*LMAX1*LMAX1
@@ -475,7 +494,7 @@ int CINT2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
                 // patch SIMDD*2 for leni, lenj, lenk with s functions
                 MALLOC_INSTACK(g1, lenl+SIMDD*2);
                 bufctr[SHLTYPl] = out;
-                lempty = empty;
+                *lempty = *empty;
         } else {
                 // enlarge out by SIMDD*2, for leni, lenj, lenk with s functions
                 cache += SIMDD*2;
@@ -489,10 +508,9 @@ int CINT2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         MALLOC_INSTACK(gout, len0+MAX(len0,leng));
         g = gout + len0;  // for gx, gy, gz
 
+        double *coeff[4] = {ci, cj, ck, cl};
         ALIGNMM Rys2eT bc;
-        double common_factor = envs->common_factor * (M_PI*M_PI*M_PI)*2/SQRTPI
-                * CINTcommon_fac_sp(envs->i_l) * CINTcommon_fac_sp(envs->j_l)
-                * CINTcommon_fac_sp(envs->k_l) * CINTcommon_fac_sp(envs->l_l);
+        double common_factor = envs->common_factor;
         double expijkl, eijcutoff;
 
         int *non0ctri = opt->non0ctr[i_sh];
@@ -545,11 +563,12 @@ k_contracted: ;
         } // end loop l_prim
         RUN_REST;
 
-        if (n_comp > 1 && !*lempty) {
+        if (n_comp > 1 && !empty_overall) {
                 int nc = i_ctr * j_ctr * k_ctr * l_ctr;
                 TRANSPOSE(gctr[SHLTYPl]);
         }
-        return !*empty;
+        *empty &= empty_overall;
+        return !empty_overall;
 }
 
 int (*CINT2e_1111_loop)(double *, CINTEnvVars *, double *, int *) = &CINT2e_loop;
diff --git a/src/cint2e_coulerf.c b/src/cint2e_coulerf.c
deleted file mode 100644
index c6f98e8..0000000
--- a/src/cint2e_coulerf.c
+++ /dev/null
@@ -1,56 +0,0 @@
-/*
- * Qcint is a general GTO integral library for computational chemistry
- * Copyright (C) 2014- Qiming Sun <osirpt.sun@gmail.com>
- *
- * This file is part of Qcint.
- *
- * Qcint is free software: you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation, either version 3 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-
-#include <stdlib.h>
-#include "cint_bas.h"
-#include "g2e.h"
-#include "optimizer.h"
-#include "cint2e.h"
-#include "cart2sph.h"
-
-CACHE_SIZE_T int2e_coulerf_sph(double *out, int *dims, int *shls, int *atm, int natm,
-                      int *bas, int nbas, double *env, CINTOpt *opt, double *cache)
-{
-        int ng[] = {0, 0, 0, 0, 0, 1, 1, 1};
-        CINTEnvVars envs;
-        CINTinit_int2e_coulerf_EnvVars(&envs, ng, shls, atm, natm, bas, nbas, env);
-        envs.f_gout = &CINTgout2e;
-        envs.f_gout_simd1 = &CINTgout2e_simd1;
-        return CINT2e_drv(out, dims, &envs, opt, cache, &c2s_sph_2e1);
-}
-void int2e_coulerf_optimizer(CINTOpt **opt, int *atm, int natm,
-                             int *bas, int nbas, double *env)
-{
-        int ng[] = {0, 0, 0, 0, 0, 1, 1, 1};
-        CINTall_2e_optimizer(opt, ng, atm, natm, bas, nbas, env);
-}
-
-#define ALL_CINT(NAME) \
-int c##NAME##_sph(double *out, int *shls, int *atm, int natm, \
-            int *bas, int nbas, double *env, CINTOpt *opt) { \
-        return NAME##_sph(out, NULL, shls, atm, natm, bas, nbas, env, opt, NULL); \
-} \
-void c##NAME##_sph_optimizer(CINTOpt **opt, int *atm, int natm, \
-                         int *bas, int nbas, double *env) { \
-        NAME##_optimizer(opt, atm, natm, bas, nbas, env); \
-}
-
-ALL_CINT(int2e_coulerf)
diff --git a/src/cint2e_gtg.c b/src/cint2e_gtg.c
deleted file mode 100644
index 99baf26..0000000
--- a/src/cint2e_gtg.c
+++ /dev/null
@@ -1,57 +0,0 @@
-/*
- * Qcint is a general GTO integral library for computational chemistry
- * Copyright (C) 2014- Qiming Sun <osirpt.sun@gmail.com>
- *
- * This file is part of Qcint.
- *
- * Qcint is free software: you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation, either version 3 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-
-#include <stdlib.h>
-#include "cint_bas.h"
-#include "g2e.h"
-#include "optimizer.h"
-#include "cint2e.h"
-#include "cart2sph.h"
-
-CACHE_SIZE_T int2e_gtg_sph(double *out, int *dims, int *shls, int *atm, int natm,
-                  int *bas, int nbas, double *env, CINTOpt *opt, double *cache)
-{
-        int ng[] = {0, 0, 0, 0, 0, 1, 1, 1};
-        CINTEnvVars envs;
-        CINTinit_int2e_gtg_EnvVars(&envs, ng, shls, atm, natm, bas, nbas, env);
-        envs.f_gout = &CINTgout2e;
-        envs.f_gout_simd1 = &CINTgout2e_simd1;
-        return CINT2e_drv(out, dims, &envs, opt, cache, &c2s_sph_2e1);
-}
-void int2e_gtg_optimizer(CINTOpt **opt, int *atm, int natm,
-                         int *bas, int nbas, double *env)
-{
-        int ng[] = {0, 0, 0, 0, 0, 1, 1, 1};
-        CINTall_2e_gtg_optimizer(opt, ng, atm, natm, bas, nbas, env);
-}
-
-#define ALL_CINT(NAME) \
-int c##NAME##_sph(double *out, int *shls, int *atm, int natm, \
-            int *bas, int nbas, double *env, CINTOpt *opt) { \
-        return NAME##_sph(out, NULL, shls, atm, natm, bas, nbas, env, opt, NULL); \
-} \
-void c##NAME##_sph_optimizer(CINTOpt **opt, int *atm, int natm, \
-                         int *bas, int nbas, double *env) { \
-        NAME##_optimizer(opt, atm, natm, bas, nbas, env); \
-}
-
-ALL_CINT(int2e_gtg);
-
diff --git a/src/cint3c2e.c b/src/cint3c2e.c
index 6e681f5..a19b164 100644
--- a/src/cint3c2e.c
+++ b/src/cint3c2e.c
@@ -21,6 +21,7 @@
 #include <stdlib.h>
 #include <string.h>
 #include <math.h>
+#include <assert.h>
 #include "cint_bas.h"
 #include "misc.h"
 #include "g2e.h"
@@ -63,15 +64,38 @@
                 (*(fp2c[i]))(gctr[it], gprim[i], coeff[it]+im, \
                              ngp[it], x_prim[it], x_ctr[it], \
                              non0ctr[it][im], non0idx[it]+im*x_ctr[it]); \
+                empty_overall = 0; \
+        } \
+        cum = 0; \
+        np2c = 0;
+
+#define POP_PRIM2CTR_AND_SET0 \
+        for (i = 0; i < np2c; i++) { \
+                it = shltyp[i]; \
+                if (it != SHLTYPi) { \
+                        im = iprim[i]; \
+                        (*(fp2c[i]))(gctr[it], gprim[i], coeff[it]+im, \
+                                     ngp[it], x_prim[it], x_ctr[it], \
+                                     non0ctr[it][im], non0idx[it]+im*x_ctr[it]); \
+                        empty_overall = 0; \
+                } else if (fp2c[i] == CINTiprim_to_ctr_0) { \
+                        double *pout = gctr[it]; \
+                        for (int k = 0; k < nf; k++) { \
+                                pout[k] = 0.; \
+                        } \
+                } \
         } \
         cum = 0; \
         np2c = 0;
 
 #define PUSH(RIJ, EXPIJ) \
         if (cum == SIMDD) { \
-                (*envs->f_g0_2e)(g, cutoff, &bc, envs, cum); \
-                (*envs->f_gout)(gout, g, idx, envs); \
-                POP_PRIM2CTR; \
+                if ((*envs->f_g0_2e)(g, cutoff, &bc, envs, cum)) { \
+                        (*envs->f_gout)(gout, g, idx, envs); \
+                        POP_PRIM2CTR; \
+                } else { \
+                        POP_PRIM2CTR_AND_SET0; \
+                } \
         } \
         envs->ai[cum] = ai[ip]; \
         envs->aj[cum] = aj[jp]; \
@@ -115,26 +139,30 @@
 
 #define RUN_REST \
         if (cum == 1) { \
-                (*envs->f_g0_2e_simd1)(g, cutoff, &bc, envs, 0); \
-                (*envs->f_gout_simd1)(gout, g, idx, envs); \
+                if ((*envs->f_g0_2e_simd1)(g, cutoff, &bc, envs, 0)) { \
+                        (*envs->f_gout_simd1)(gout, g, idx, envs); \
+                        POP_PRIM2CTR; \
+                } else { \
+                        POP_PRIM2CTR_AND_SET0; \
+                } \
         } else if (cum > 1) { \
-                r1 = MM_SET1(1.); \
-                for (i = 0; i < envs->nrys_roots; i++) { \
-                        MM_STORE(bc.u+i*SIMDD, r1); \
-                        MM_STORE(bc.w+i*SIMDD, r1); \
+                if ((*envs->f_g0_2e)(g, cutoff, &bc, envs, cum)) { \
+                        (*envs->f_gout)(gout, g, idx, envs); \
+                        POP_PRIM2CTR; \
+                } else { \
+                        POP_PRIM2CTR_AND_SET0; \
                 } \
-                (*envs->f_g0_2e)(g, cutoff, &bc, envs, cum); \
-                (*envs->f_gout)(gout, g, idx, envs); \
-        } \
-        POP_PRIM2CTR
+        } else { \
+                assert(np2c == 0); \
+        }
 
 #define TRANSPOSE(a) \
         if (*empty) { \
                 CINTdmat_transpose(out, a, nf*nc, n_comp); \
+                *empty = 0; \
         } else { \
                 CINTdplus_transpose(out, a, nf*nc, n_comp); \
-        } \
-        *empty = 0;
+        }
 
 int CINT3c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty)
 {
@@ -161,6 +189,7 @@ int CINT3c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
         double *coeff[3] = {ci, cj, ck};
 
         double expcutoff = envs->expcutoff;
+        double rr_ij = SQUARE(envs->rirj);
         double *log_maxci, *log_maxcj;
         PairData *pdata_base;
         MALLOC_DATA_INSTACK(pdata_base, i_prim*j_prim);
@@ -170,7 +199,7 @@ int CINT3c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
         CINTOpt_log_max_pgto_coeff(log_maxcj, cj, j_prim, j_ctr);
         if (CINTset_pairdata(pdata_base, ai, aj, envs->ri, envs->rj,
                              log_maxci, log_maxcj, envs->li_ceil, envs->lj_ceil,
-                             i_prim, j_prim, SQUARE(envs->rirj), expcutoff, env)) {
+                             i_prim, j_prim, rr_ij, expcutoff, env)) {
                 return 0;
         }
 
@@ -182,6 +211,25 @@ int CINT3c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
         int *iempty = _empty + 0;
         int *jempty = _empty + 1;
         int *kempty = _empty + 2;
+        int empty_overall = 1;
+
+        double omega = env[PTR_RANGE_OMEGA];
+        if (omega < 0 && envs->rys_order > 1) {
+                double r_guess = 8.;
+                double omega2 = omega * omega;
+                int lij = envs->li_ceil + envs->lj_ceil;
+                if (lij > 0) {
+                        double dist_ij = sqrt(rr_ij);
+                        double aij = ai[i_prim-1] + aj[j_prim-1];
+                        double theta = omega2 / (omega2 + aij);
+                        expcutoff += lij * approx_log(
+                                (dist_ij+theta*r_guess+1.)/(dist_ij+1.));
+                }
+                if (envs->lk_ceil > 0) {
+                        double theta = omega2 / (omega2 + ak[k_prim-1]);
+                        expcutoff += envs->lk_ceil * approx_log(theta*r_guess+1.);
+                }
+        }
 
         int *idx;
         MALLOC_DATA_INSTACK(idx, nf * 3);
@@ -200,9 +248,7 @@ int CINT3c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
         CINTOpt_non0coeff_byshell(non0idxk, non0ctrk, coeff[2], k_prim, k_ctr);
         int *non0ctr[3] = {non0ctri, non0ctrj, non0ctrk};
         int *non0idx[3] = {non0idxi, non0idxj, non0idxk};
-        double common_factor = envs->common_factor * (M_PI*M_PI*M_PI)*2/SQRTPI
-                * CINTcommon_fac_sp(envs->i_l) * CINTcommon_fac_sp(envs->j_l)
-                * CINTcommon_fac_sp(envs->k_l);
+        double common_factor = envs->common_factor;
 
         int ngp[4];
         ngp[0] = nf * n_comp;
@@ -221,7 +267,7 @@ int CINT3c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
                 // patch SIMDD for leni, lenj with s functions
                 MALLOC_INSTACK(g1, lenk+SIMDD);
                 gctr[SHLTYPk] = out;
-                kempty = empty;
+                *kempty = *empty;
         } else {
                 // enlarge out by SIMDD, for leni, lenj with s functions
                 cache += SIMDD;
@@ -269,11 +315,12 @@ int CINT3c2e_loop_nopt(double *out, CINTEnvVars *envs, double *cache, int *empty
         } // end loop k_prim
         RUN_REST;
 
-        if (n_comp > 1 && !*kempty) {
+        if (n_comp > 1 && !empty_overall) {
                 int nc = i_ctr * j_ctr * k_ctr;
                 TRANSPOSE(gctr[SHLTYPk]);
         }
-        return !*empty;
+        *empty &= empty_overall;
+        return !empty_overall;
 }
 
 
@@ -306,6 +353,7 @@ int CINT3c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         double *ck = env + bas(PTR_COEFF, k_sh);
         double *coeff[3] = {ci, cj, ck};
         double expcutoff = envs->expcutoff;
+        double rr_ij = SQUARE(envs->rirj);
         PairData *pdata_base, *pdata_ij;
         if (opt->pairdata != NULL) {
                 pdata_base = opt->pairdata[i_sh*opt->nbas+j_sh];
@@ -315,7 +363,7 @@ int CINT3c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
                 MALLOC_DATA_INSTACK(pdata_base, i_prim*j_prim);
                 if (CINTset_pairdata(pdata_base, ai, aj, envs->ri, envs->rj,
                                      log_maxci, log_maxcj, envs->li_ceil, envs->lj_ceil,
-                                     i_prim, j_prim, SQUARE(envs->rirj), expcutoff, env)) {
+                                     i_prim, j_prim, rr_ij, expcutoff, env)) {
                         return 0;
                 }
         }
@@ -328,6 +376,25 @@ int CINT3c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         int *iempty = _empty + 0;
         int *jempty = _empty + 1;
         int *kempty = _empty + 2;
+        int empty_overall = 1;
+
+        double omega = env[PTR_RANGE_OMEGA];
+        if (omega < 0 && envs->rys_order > 1) {
+                double r_guess = 8.;
+                double omega2 = omega * omega;
+                int lij = envs->li_ceil + envs->lj_ceil;
+                if (lij > 0) {
+                        double dist_ij = sqrt(rr_ij);
+                        double aij = ai[i_prim-1] + aj[j_prim-1];
+                        double theta = omega2 / (omega2 + aij);
+                        expcutoff += lij * approx_log(
+                                (dist_ij+theta*r_guess+1.)/(dist_ij+1.));
+                }
+                if (envs->lk_ceil > 0) {
+                        double theta = omega2 / (omega2 + ak[k_prim-1]);
+                        expcutoff += envs->lk_ceil * approx_log(theta*r_guess+1.);
+                }
+        }
 
         int *idx = opt->index_xyz_array[envs->i_l*LMAX1*LMAX1
                                        +envs->j_l*LMAX1
@@ -343,9 +410,7 @@ int CINT3c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         CINTOpt_non0coeff_byshell(non0idxk, non0ctrk, ck, k_prim, k_ctr);
         int *non0ctr[3] = {opt->non0ctr[i_sh], opt->non0ctr[j_sh], non0ctrk};
         int *non0idx[3] = {opt->sortedidx[i_sh], opt->sortedidx[j_sh], non0idxk};
-        double common_factor = envs->common_factor * (M_PI*M_PI*M_PI)*2/SQRTPI
-                * CINTcommon_fac_sp(envs->i_l) * CINTcommon_fac_sp(envs->j_l)
-                * CINTcommon_fac_sp(envs->k_l);
+        double common_factor = envs->common_factor;
 
         int ngp[4];
         ngp[0] = nf * n_comp;
@@ -365,7 +430,7 @@ int CINT3c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
                 // patch SIMDD for leni, lenj with s functions
                 MALLOC_INSTACK(g1, lenk+SIMDD);
                 gctr[SHLTYPk] = out;
-                kempty = empty;
+                *kempty = *empty;
         } else {
                 // enlarge out by SIMDD, for leni, lenj with s functions
                 cache += SIMDD;
@@ -412,11 +477,12 @@ int CINT3c2e_loop(double *out, CINTEnvVars *envs, double *cache, int *empty)
         } // end loop k_prim
         RUN_REST;
 
-        if (n_comp > 1 && !*kempty) {
+        if (n_comp > 1 && !empty_overall) {
                 int nc = i_ctr * j_ctr * k_ctr;
                 TRANSPOSE(gctr[SHLTYPk]);
         }
-        return !*empty;
+        *empty &= empty_overall;
+        return !empty_overall;
 }
 
 
diff --git a/src/cint3c2e_gtg.c b/src/cint3c2e_gtg.c
deleted file mode 100644
index 9808a75..0000000
--- a/src/cint3c2e_gtg.c
+++ /dev/null
@@ -1,40 +0,0 @@
-/*
- * Copyright (C) 2013-  Qiming Sun <osirpt.sun@gmail.com>
- */
-
-#include <stdlib.h>
-#include "cint_bas.h"
-#include "g2e.h"
-#include "optimizer.h"
-#include "cint2e.h"
-#include "cart2sph.h"
-
-CACHE_SIZE_T int3c2e_gtg_sph(double *out, int *dims, int *shls, int *atm, int natm,
-                  int *bas, int nbas, double *env, CINTOpt *opt, double *cache)
-{
-        int ng[] = {0, 0, 0, 0, 0, 1, 1, 1};
-        CINTEnvVars envs;
-        CINTinit_int3c2e_gtg_EnvVars(&envs, ng, shls, atm, natm, bas, nbas, env);
-        envs.f_gout = &CINTgout2e;
-        envs.f_gout_simd1 = &CINTgout2e_simd1;
-        return CINT3c2e_drv(out, dims, &envs, opt, cache, &c2s_sph_3c2e1, 0);
-}
-void int3c2e_gtg_optimizer(CINTOpt **opt, int *atm, int natm,
-                         int *bas, int nbas, double *env)
-{
-        int ng[] = {0, 0, 0, 0, 0, 1, 1, 1};
-        CINTall_3c2e_gtg_optimizer(opt, ng, atm, natm, bas, nbas, env);
-}
-
-#define ALL_CINT(NAME) \
-int c##NAME##_sph(double *out, int *shls, int *atm, int natm, \
-            int *bas, int nbas, double *env, CINTOpt *opt) { \
-        return NAME##_sph(out, NULL, shls, atm, natm, bas, nbas, env, opt, NULL); \
-} \
-void c##NAME##_sph_optimizer(CINTOpt **opt, int *atm, int natm, \
-                         int *bas, int nbas, double *env) { \
-        NAME##_optimizer(opt, atm, natm, bas, nbas, env); \
-}
-
-ALL_CINT(int3c2e_gtg)
-
diff --git a/src/fmt.c b/src/fmt.c
index 018d418..33e1f3c 100644
--- a/src/fmt.c
+++ b/src/fmt.c
@@ -319,7 +319,6 @@ void fmt_erfc_like(double *f, double t, double lower, int m)
 
         int i;
         double lower2 = lower * lower;
-#ifdef WITH_RANGE_COULOMB
         // F[m] < .5*sqrt(pi/t) * erfc(low*tt)
         if (t * lower2 > ERFC_bound) {
                 for (i = 0; i <= m; i++) {
@@ -327,7 +326,6 @@ void fmt_erfc_like(double *f, double t, double lower, int m)
                 }
                 return;
         }
-#endif
 
         if (t < TURNOVER_POINT[m]) {
                 fmt1_erfc_like(f, t, lower, m);
@@ -357,7 +355,6 @@ void fmt_lerfc_like(long double *f, long double t, long double lower, int m)
 
         int i;
         long double lower2 = lower * lower;
-#ifdef WITH_RANGE_COULOMB
         // F[m] < .5*sqrt(pi/t) * erfc(low*tt)
         if (t * lower2 > ERFC_bound) {
                 for (i = 0; i <= m; i++) {
@@ -365,7 +362,6 @@ void fmt_lerfc_like(long double *f, long double t, long double lower, int m)
                 }
                 return;
         }
-#endif
 
         if (t < TURNOVER_POINT[m]) {
                 fmt1_lerfc_like(f, t, lower, m);
@@ -477,7 +473,6 @@ void fmt_qerfc_like(__float128 *f, __float128 t, __float128 lower, int m)
 
         int i;
         __float128 lower2 = lower * lower;
-#ifdef WITH_RANGE_COULOMB
         // F[m] < .5*sqrt(pi/t) * erfc(low*tt)
         if (t * lower2 > ERFC_bound) {
                 for (i = 0; i <= m; i++) {
@@ -485,7 +480,6 @@ void fmt_qerfc_like(__float128 *f, __float128 t, __float128 lower, int m)
                 }
                 return;
         }
-#endif
 
         if (t < TURNOVER_POINT[m]) {
                 fmt1_qerfc_like(f, t, lower, m);
diff --git a/src/g2e.c b/src/g2e.c
index ca79b07..7465b70 100644
--- a/src/g2e.c
+++ b/src/g2e.c
@@ -62,7 +62,15 @@ void CINTinit_int2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
         envs->nfk = (envs->k_l+1)*(envs->k_l+2)/2;
         envs->nfl = (envs->l_l+1)*(envs->l_l+2)/2;
         envs->nf = envs->nfi * envs->nfk * envs->nfl * envs->nfj;
-        envs->common_factor = 1;
+
+        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
+        envs->rj = env + atm(PTR_COORD, bas(ATOM_OF, j_sh));
+        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
+        envs->rl = env + atm(PTR_COORD, bas(ATOM_OF, l_sh));
+
+        envs->common_factor = (M_PI*M_PI*M_PI)*2/SQRTPI
+                * CINTcommon_fac_sp(envs->i_l) * CINTcommon_fac_sp(envs->j_l)
+                * CINTcommon_fac_sp(envs->k_l) * CINTcommon_fac_sp(envs->l_l);
         if (env[PTR_EXPCUTOFF] == 0) {
                 envs->expcutoff = EXPCUTOFF;
         } else {
@@ -79,42 +87,19 @@ void CINTinit_int2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
         envs->lj_ceil = envs->j_l + ng[JINC];
         envs->lk_ceil = envs->k_l + ng[KINC];
         envs->ll_ceil = envs->l_l + ng[LINC];
-
-        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
-        envs->rj = env + atm(PTR_COORD, bas(ATOM_OF, j_sh));
-        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
-        envs->rl = env + atm(PTR_COORD, bas(ATOM_OF, l_sh));
-
-        int nroots = (envs->li_ceil + envs->lj_ceil +
-                      envs->lk_ceil + envs->ll_ceil)/2 + 1;
-        envs->nrys_roots = nroots;
-        assert(nroots < MXRYSROOTS);
+        int rys_order =(envs->li_ceil + envs->lj_ceil
+                        + envs->lk_ceil + envs->ll_ceil)/2 + 1;
+        int nrys_roots = rys_order;
+        double omega = env[PTR_RANGE_OMEGA];
+        if (omega < 0 && rys_order <= 3) {
+                nrys_roots *= 2;
+        }
+        envs->rys_order = rys_order;
+        envs->nrys_roots = nrys_roots;
 
         int dli, dlj, dlk, dll;
         int ibase = envs->li_ceil > envs->lj_ceil;
         int kbase = envs->lk_ceil > envs->ll_ceil;
-        if (nroots <= 2) {
-                envs->f_g0_2d4d = &CINTg0_2e_2d4d_unrolled;
-                envs->f_g0_2d4d_simd1 = &CINTg0_2e_2d4d_unrolled_simd1;
-        } else if (kbase) {
-                if (ibase) {
-                        envs->f_g0_2d4d = &CINTg0_2e_ik2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_ik2d4d_simd1;
-                } else {
-                        envs->f_g0_2d4d = &CINTg0_2e_kj2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_kj2d4d_simd1;
-                }
-        } else {
-                if (ibase) {
-                        envs->f_g0_2d4d = &CINTg0_2e_il2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_il2d4d_simd1;
-                } else {
-                        envs->f_g0_2d4d = &CINTg0_2e_lj2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_lj2d4d_simd1;
-                }
-        }
-        envs->f_g0_2e = &CINTg0_2e;
-        envs->f_g0_2e_simd1 = &CINTg0_2e_simd1;
 
         if (kbase) {
                 dlk = envs->lk_ceil + envs->ll_ceil + 1;
@@ -131,11 +116,11 @@ void CINTinit_int2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
                 dli = envs->li_ceil + 1;
                 dlj = envs->li_ceil + envs->lj_ceil + 1;
         }
-        envs->g_stride_i = nroots;
-        envs->g_stride_k = nroots * dli;
-        envs->g_stride_l = nroots * dli * dlk;
-        envs->g_stride_j = nroots * dli * dlk * dll;
-        envs->g_size     = nroots * dli * dlk * dll * dlj;
+        envs->g_stride_i = nrys_roots;
+        envs->g_stride_k = nrys_roots * dli;
+        envs->g_stride_l = nrys_roots * dli * dlk;
+        envs->g_stride_j = nrys_roots * dli * dlk * dll;
+        envs->g_size     = nrys_roots * dli * dlk * dll * dlj;
 
         if (kbase) {
                 envs->g2d_klmax = envs->g_stride_k;
@@ -164,6 +149,33 @@ void CINTinit_int2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
                 envs->rirj[1] = envs->rj[1] - envs->ri[1];
                 envs->rirj[2] = envs->rj[2] - envs->ri[2];
         }
+
+        if (rys_order <= 2) {
+                envs->f_g0_2d4d = &CINTg0_2e_2d4d_unrolled;
+                envs->f_g0_2d4d_simd1 = &CINTg0_2e_2d4d_unrolled_simd1;
+                if (rys_order != nrys_roots) {
+                        envs->f_g0_2d4d = &CINTsrg0_2e_2d4d_unrolled;
+                        envs->f_g0_2d4d_simd1 = &CINTsrg0_2e_2d4d_unrolled_simd1;
+                }
+        } else if (kbase) {
+                if (ibase) {
+                        envs->f_g0_2d4d = &CINTg0_2e_ik2d4d;
+                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_ik2d4d_simd1;
+                } else {
+                        envs->f_g0_2d4d = &CINTg0_2e_kj2d4d;
+                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_kj2d4d_simd1;
+                }
+        } else {
+                if (ibase) {
+                        envs->f_g0_2d4d = &CINTg0_2e_il2d4d;
+                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_il2d4d_simd1;
+                } else {
+                        envs->f_g0_2d4d = &CINTg0_2e_lj2d4d;
+                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_lj2d4d_simd1;
+                }
+        }
+        envs->f_g0_2e = &CINTg0_2e;
+        envs->f_g0_2e_simd1 = &CINTg0_2e_simd1;
 }
 
 void CINTg4c_index_xyz(int *idx, CINTEnvVars *envs)
@@ -765,24 +777,8 @@ MM_STORE(gz+n*SIMDD, MM_FMA(rz, MM_LOAD(p1z+n*SIMDD), MM_LOAD(p2z+n*SIMDD)));
 }
 /************* some special g0_4d results *************/
 /* 4 digits stand for i_ceil, k_ceil, l_ceil, j_ceil */
-static inline void _make_rc(double *rc, double *cx, double *cy, double *cz, double *r)
-{
-        __MD r0 = MM_SET1(r[0]);
-        __MD r1 = MM_SET1(r[1]);
-        __MD r2 = MM_SET1(r[2]);
-        MM_STORE(rc+0*SIMDD, MM_ADD(r0, MM_LOAD(cx+0*SIMDD)));
-        MM_STORE(rc+1*SIMDD, MM_ADD(r0, MM_LOAD(cx+1*SIMDD)));
-        MM_STORE(rc+2*SIMDD, MM_ADD(r1, MM_LOAD(cy+0*SIMDD)));
-        MM_STORE(rc+3*SIMDD, MM_ADD(r1, MM_LOAD(cy+1*SIMDD)));
-        MM_STORE(rc+4*SIMDD, MM_ADD(r2, MM_LOAD(cz+0*SIMDD)));
-        MM_STORE(rc+5*SIMDD, MM_ADD(r2, MM_LOAD(cz+1*SIMDD)));
-}
-
 static inline void _g0_2d4d_0000(double *g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        //MM_STORE(g+0*SIMDD, MM_SET1(1.));
-        //MM_STORE(g+1*SIMDD, MM_SET1(1.));
-        //g[2] = w[0];
 }
 
 static inline void _g0_2d4d_0001(double *g, Rys2eT *bc, CINTEnvVars *envs)
@@ -831,29 +827,35 @@ static inline void _g0_2d4d_0003(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cy = bc->c0py;
         double *cz = bc->c0pz;
         double *b =  bc->b01;
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
         __MD b0 = MM_LOAD(b+0*SIMDD);
         __MD b1 = MM_LOAD(b+1*SIMDD);
         __MD g16 = MM_LOAD(g+16*SIMDD);
         __MD g17 = MM_LOAD(g+17*SIMDD);
         __MD i3 = MM_SET1(3.);
-        MM_STORE(g+2 *SIMDD, MM_LOAD(cx+0*SIMDD));
-        MM_STORE(g+3 *SIMDD, MM_LOAD(cx+1*SIMDD));
-        MM_STORE(g+10*SIMDD, MM_LOAD(cy+0*SIMDD));
-        MM_STORE(g+11*SIMDD, MM_LOAD(cy+1*SIMDD));
-        MM_STORE(g+18*SIMDD, MM_LOAD(cz+0*SIMDD) * g16);
-        MM_STORE(g+19*SIMDD, MM_LOAD(cz+1*SIMDD) * g17);
-        MM_STORE(g+4 *SIMDD, MM_LOAD(cx+0*SIMDD) * MM_LOAD(cx+0*SIMDD) + b0);
-        MM_STORE(g+5 *SIMDD, MM_LOAD(cx+1*SIMDD) * MM_LOAD(cx+1*SIMDD) + b1);
-        MM_STORE(g+12*SIMDD, MM_LOAD(cy+0*SIMDD) * MM_LOAD(cy+0*SIMDD) + b0);
-        MM_STORE(g+13*SIMDD, MM_LOAD(cy+1*SIMDD) * MM_LOAD(cy+1*SIMDD) + b1);
-        MM_STORE(g+20*SIMDD,(MM_LOAD(cz+0*SIMDD) * MM_LOAD(cz+0*SIMDD) + b0)* g16);
-        MM_STORE(g+21*SIMDD,(MM_LOAD(cz+1*SIMDD) * MM_LOAD(cz+1*SIMDD) + b1)* g17);
-        MM_STORE(g+6 *SIMDD, MM_LOAD(cx+0*SIMDD) *(MM_LOAD(cx+0*SIMDD) * MM_LOAD(cx+0*SIMDD) + i3 * b0));
-        MM_STORE(g+7 *SIMDD, MM_LOAD(cx+1*SIMDD) *(MM_LOAD(cx+1*SIMDD) * MM_LOAD(cx+1*SIMDD) + i3 * b1));
-        MM_STORE(g+14*SIMDD, MM_LOAD(cy+0*SIMDD) *(MM_LOAD(cy+0*SIMDD) * MM_LOAD(cy+0*SIMDD) + i3 * b0));
-        MM_STORE(g+15*SIMDD, MM_LOAD(cy+1*SIMDD) *(MM_LOAD(cy+1*SIMDD) * MM_LOAD(cy+1*SIMDD) + i3 * b1));
-        MM_STORE(g+22*SIMDD,(MM_LOAD(cz+0*SIMDD) * MM_LOAD(cz+0*SIMDD) + i3 * b0)* MM_LOAD(g+18*SIMDD));
-        MM_STORE(g+23*SIMDD,(MM_LOAD(cz+1*SIMDD) * MM_LOAD(cz+1*SIMDD) + i3 * b1)* MM_LOAD(g+19*SIMDD));
+        MM_STORE(g+2 *SIMDD, cx0);
+        MM_STORE(g+3 *SIMDD, cx1);
+        MM_STORE(g+10*SIMDD, cy0);
+        MM_STORE(g+11*SIMDD, cy1);
+        MM_STORE(g+18*SIMDD, cz0 * g16);
+        MM_STORE(g+19*SIMDD, cz1 * g17);
+        MM_STORE(g+4 *SIMDD, cx0 * cx0 + b0);
+        MM_STORE(g+5 *SIMDD, cx1 * cx1 + b1);
+        MM_STORE(g+12*SIMDD, cy0 * cy0 + b0);
+        MM_STORE(g+13*SIMDD, cy1 * cy1 + b1);
+        MM_STORE(g+20*SIMDD,(cz0 * cz0 + b0)* g16);
+        MM_STORE(g+21*SIMDD,(cz1 * cz1 + b1)* g17);
+        MM_STORE(g+6 *SIMDD, cx0 *(cx0 * cx0 + i3 * b0));
+        MM_STORE(g+7 *SIMDD, cx1 *(cx1 * cx1 + i3 * b1));
+        MM_STORE(g+14*SIMDD, cy0 *(cy0 * cy0 + i3 * b0));
+        MM_STORE(g+15*SIMDD, cy1 *(cy1 * cy1 + i3 * b1));
+        MM_STORE(g+22*SIMDD,(cz0 * cz0 + i3 * b0)* MM_LOAD(g+18*SIMDD));
+        MM_STORE(g+23*SIMDD,(cz1 * cz1 + i3 * b1)* MM_LOAD(g+19*SIMDD));
 }
 
 static inline void _g0_2d4d_0010(double *g, Rys2eT *bc, CINTEnvVars *envs)
@@ -873,20 +875,21 @@ static inline void _g0_2d4d_0011(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cz = bc->c0pz;
         double *b  = bc->b01;
         double *r  = envs->rkrl;
-        ALIGNMM double rc[6*SIMDD];
-        _make_rc(rc, cx, cy, cz, r);
-        __MD r0 = MM_LOAD(rc+0*SIMDD);
-        __MD r1 = MM_LOAD(rc+1*SIMDD);
-        __MD r2 = MM_LOAD(rc+2*SIMDD);
-        __MD r3 = MM_LOAD(rc+3*SIMDD);
-        __MD r4 = MM_LOAD(rc+4*SIMDD);
-        __MD r5 = MM_LOAD(rc+5*SIMDD);
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
         __MD cx0 = MM_LOAD(cx+0*SIMDD);
         __MD cx1 = MM_LOAD(cx+1*SIMDD);
         __MD cy0 = MM_LOAD(cy+0*SIMDD);
         __MD cy1 = MM_LOAD(cy+1*SIMDD);
         __MD cz0 = MM_LOAD(cz+0*SIMDD);
         __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD r0 = MM_ADD(rx, cx0);
+        __MD r1 = MM_ADD(rx, cx1);
+        __MD r2 = MM_ADD(ry, cy0);
+        __MD r3 = MM_ADD(ry, cy1);
+        __MD r4 = MM_ADD(rz, cz0);
+        __MD r5 = MM_ADD(rz, cz1);
         __MD b0 = MM_LOAD(b+0*SIMDD);
         __MD b1 = MM_LOAD(b+1*SIMDD);
         __MD g24 = MM_LOAD(g+24*SIMDD);
@@ -918,20 +921,21 @@ static inline void _g0_2d4d_0012(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cz = bc->c0pz;
         double *b =  bc->b01;
         double *r  = envs->rkrl;
-        ALIGNMM double rc[6*SIMDD];
-        _make_rc(rc, cx, cy, cz, r);
-        __MD r0 = MM_LOAD(rc+0*SIMDD);
-        __MD r1 = MM_LOAD(rc+1*SIMDD);
-        __MD r2 = MM_LOAD(rc+2*SIMDD);
-        __MD r3 = MM_LOAD(rc+3*SIMDD);
-        __MD r4 = MM_LOAD(rc+4*SIMDD);
-        __MD r5 = MM_LOAD(rc+5*SIMDD);
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
         __MD cx0 = MM_LOAD(cx+0*SIMDD);
         __MD cx1 = MM_LOAD(cx+1*SIMDD);
         __MD cy0 = MM_LOAD(cy+0*SIMDD);
         __MD cy1 = MM_LOAD(cy+1*SIMDD);
         __MD cz0 = MM_LOAD(cz+0*SIMDD);
         __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD r0 = MM_ADD(rx, cx0);
+        __MD r1 = MM_ADD(rx, cx1);
+        __MD r2 = MM_ADD(ry, cy0);
+        __MD r3 = MM_ADD(ry, cy1);
+        __MD r4 = MM_ADD(rz, cz0);
+        __MD r5 = MM_ADD(rz, cz1);
         __MD b0 = MM_LOAD(b+0*SIMDD);
         __MD b1 = MM_LOAD(b+1*SIMDD);
         __MD i2 = MM_SET1(2.);
@@ -1005,26 +1009,27 @@ static inline void _g0_2d4d_0021(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cy = bc->c0py;
         double *cz = bc->c0pz;
         double *b1 = bc->b01;
-        double *rkl = envs->rkrl;
-        ALIGNMM double rc[6*SIMDD];
-        _make_rc(rc, cx, cy, cz, rkl);
-        __MD b10 = MM_LOAD(b1+0*SIMDD);
-        __MD b11 = MM_LOAD(b1+1*SIMDD);
-        __MD g32 = MM_LOAD(g+32*SIMDD);
-        __MD g33 = MM_LOAD(g+33*SIMDD);
-        __MD i2 = MM_SET1(2.);
-        __MD r0 = MM_LOAD(rc+0*SIMDD);
-        __MD r1 = MM_LOAD(rc+1*SIMDD);
-        __MD r2 = MM_LOAD(rc+2*SIMDD);
-        __MD r3 = MM_LOAD(rc+3*SIMDD);
-        __MD r4 = MM_LOAD(rc+4*SIMDD);
-        __MD r5 = MM_LOAD(rc+5*SIMDD);
+        double *r = envs->rkrl;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
         __MD s0 = MM_LOAD(cx+0*SIMDD);
         __MD s1 = MM_LOAD(cx+1*SIMDD);
         __MD s2 = MM_LOAD(cy+0*SIMDD);
         __MD s3 = MM_LOAD(cy+1*SIMDD);
         __MD s4 = MM_LOAD(cz+0*SIMDD);
         __MD s5 = MM_LOAD(cz+1*SIMDD);
+        __MD r0 = MM_ADD(rx, s0);
+        __MD r1 = MM_ADD(rx, s1);
+        __MD r2 = MM_ADD(ry, s2);
+        __MD r3 = MM_ADD(ry, s3);
+        __MD r4 = MM_ADD(rz, s4);
+        __MD r5 = MM_ADD(rz, s5);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD i2 = MM_SET1(2.);
         MM_STORE(g+2 *SIMDD, s0);
         MM_STORE(g+3 *SIMDD, s1);
         MM_STORE(g+8 *SIMDD, r0);
@@ -1218,8 +1223,8 @@ static inline void _g0_2d4d_0110(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b   = bc->b00;
-        __MD b0  = MM_LOAD(b+0*SIMDD);
-        __MD b1  = MM_LOAD(b+1*SIMDD);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
         __MD r0 = MM_LOAD(c0x+0*SIMDD);
         __MD r1 = MM_LOAD(c0x+1*SIMDD);
         __MD r2 = MM_LOAD(c0y+0*SIMDD);
@@ -1264,57 +1269,76 @@ static inline void _g0_2d4d_0111(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cpz = bc->c0pz;
         double *b0  = bc->b00;
         double *b1  = bc->b01;
-        double *rkl = envs->rkrl;
-        ALIGNMM double rcp[6*SIMDD];
-        _make_rc(rcp, cpx, cpy, cpz, rkl);
+        double *r   = envs->rkrl;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD rcp0 = MM_ADD(rx, cpx0);
+        __MD rcp1 = MM_ADD(rx, cpx1);
+        __MD rcp2 = MM_ADD(ry, cpy0);
+        __MD rcp3 = MM_ADD(ry, cpy1);
+        __MD rcp4 = MM_ADD(rz, cpz0);
+        __MD rcp5 = MM_ADD(rz, cpz1);
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
         __MD b00 = MM_LOAD(b0+0*SIMDD);
         __MD b01 = MM_LOAD(b0+1*SIMDD);
         __MD b10 = MM_LOAD(b1+0*SIMDD);
         __MD b11 = MM_LOAD(b1+1*SIMDD);
         __MD g48 = MM_LOAD(g+48*SIMDD);
         __MD g49 = MM_LOAD(g+49*SIMDD);
-        MM_STORE(g+2 *SIMDD, MM_LOAD(rcp+0*SIMDD));
-        MM_STORE(g+3 *SIMDD, MM_LOAD(rcp+1*SIMDD));
-        MM_STORE(g+4 *SIMDD, MM_LOAD(cpx+0*SIMDD));
-        MM_STORE(g+5 *SIMDD, MM_LOAD(cpx+1*SIMDD));
-        MM_STORE(g+12*SIMDD, MM_LOAD(c0x+0*SIMDD));
-        MM_STORE(g+13*SIMDD, MM_LOAD(c0x+1*SIMDD));
-        MM_STORE(g+26*SIMDD, MM_LOAD(rcp+2*SIMDD));
-        MM_STORE(g+27*SIMDD, MM_LOAD(rcp+3*SIMDD));
-        MM_STORE(g+28*SIMDD, MM_LOAD(cpy+0*SIMDD));
-        MM_STORE(g+29*SIMDD, MM_LOAD(cpy+1*SIMDD));
-        MM_STORE(g+36*SIMDD, MM_LOAD(c0y+0*SIMDD));
-        MM_STORE(g+37*SIMDD, MM_LOAD(c0y+1*SIMDD));
-        MM_STORE(g+50*SIMDD, MM_LOAD(rcp+4*SIMDD) * g48);
-        MM_STORE(g+51*SIMDD, MM_LOAD(rcp+5*SIMDD) * g49);
-        MM_STORE(g+52*SIMDD, MM_LOAD(cpz+0*SIMDD) * g48);
-        MM_STORE(g+53*SIMDD, MM_LOAD(cpz+1*SIMDD) * g49);
-        MM_STORE(g+60*SIMDD, MM_LOAD(c0z+0*SIMDD) * g48);
-        MM_STORE(g+61*SIMDD, MM_LOAD(c0z+1*SIMDD) * g49);
-        MM_STORE(g+14*SIMDD, MM_LOAD(c0x+0*SIMDD) * MM_LOAD(rcp+0*SIMDD) + b00);
-        MM_STORE(g+15*SIMDD, MM_LOAD(c0x+1*SIMDD) * MM_LOAD(rcp+1*SIMDD) + b01);
-        MM_STORE(g+16*SIMDD, MM_LOAD(c0x+0*SIMDD) * MM_LOAD(cpx+0*SIMDD) + b00);
-        MM_STORE(g+17*SIMDD, MM_LOAD(c0x+1*SIMDD) * MM_LOAD(cpx+1*SIMDD) + b01);
-        MM_STORE(g+6 *SIMDD, MM_LOAD(cpx+0*SIMDD) * MM_LOAD(rcp+0*SIMDD) + b10);
-        MM_STORE(g+7 *SIMDD, MM_LOAD(cpx+1*SIMDD) * MM_LOAD(rcp+1*SIMDD) + b11);
-        MM_STORE(g+30*SIMDD, MM_LOAD(cpy+0*SIMDD) * MM_LOAD(rcp+2*SIMDD) + b10);
-        MM_STORE(g+31*SIMDD, MM_LOAD(cpy+1*SIMDD) * MM_LOAD(rcp+3*SIMDD) + b11);
-        MM_STORE(g+38*SIMDD, MM_LOAD(c0y+0*SIMDD) * MM_LOAD(rcp+2*SIMDD) + b00);
-        MM_STORE(g+39*SIMDD, MM_LOAD(c0y+1*SIMDD) * MM_LOAD(rcp+3*SIMDD) + b01);
-        MM_STORE(g+40*SIMDD, MM_LOAD(c0y+0*SIMDD) * MM_LOAD(cpy+0*SIMDD) + b00);
-        MM_STORE(g+41*SIMDD, MM_LOAD(c0y+1*SIMDD) * MM_LOAD(cpy+1*SIMDD) + b01);
-        MM_STORE(g+54*SIMDD,(MM_LOAD(cpz+0*SIMDD) * MM_LOAD(rcp+4*SIMDD) + b10) * g48);
-        MM_STORE(g+55*SIMDD,(MM_LOAD(cpz+1*SIMDD) * MM_LOAD(rcp+5*SIMDD) + b11) * g49);
-        MM_STORE(g+62*SIMDD,(MM_LOAD(c0z+0*SIMDD) * MM_LOAD(rcp+4*SIMDD) + b00) * g48);
-        MM_STORE(g+63*SIMDD,(MM_LOAD(c0z+1*SIMDD) * MM_LOAD(rcp+5*SIMDD) + b01) * g49);
-        MM_STORE(g+64*SIMDD,(MM_LOAD(c0z+0*SIMDD) * MM_LOAD(cpz+0*SIMDD) + b00) * g48);
-        MM_STORE(g+65*SIMDD,(MM_LOAD(c0z+1*SIMDD) * MM_LOAD(cpz+1*SIMDD) + b01) * g49);
-        MM_STORE(g+18*SIMDD, MM_LOAD(c0x+0*SIMDD) * MM_LOAD(g+6 *SIMDD) + b00 * (MM_LOAD(rcp+0*SIMDD) + MM_LOAD(cpx+0*SIMDD)));
-        MM_STORE(g+19*SIMDD, MM_LOAD(c0x+1*SIMDD) * MM_LOAD(g+7 *SIMDD) + b01 * (MM_LOAD(rcp+1*SIMDD) + MM_LOAD(cpx+1*SIMDD)));
-        MM_STORE(g+42*SIMDD, MM_LOAD(c0y+0*SIMDD) * MM_LOAD(g+30*SIMDD) + b00 * (MM_LOAD(rcp+2*SIMDD) + MM_LOAD(cpy+0*SIMDD)));
-        MM_STORE(g+43*SIMDD, MM_LOAD(c0y+1*SIMDD) * MM_LOAD(g+31*SIMDD) + b01 * (MM_LOAD(rcp+3*SIMDD) + MM_LOAD(cpy+1*SIMDD)));
-        MM_STORE(g+66*SIMDD, MM_LOAD(c0z+0*SIMDD) * MM_LOAD(g+54*SIMDD) + b00 * (MM_LOAD(g+50*SIMDD) + MM_LOAD(g+52*SIMDD)));
-        MM_STORE(g+67*SIMDD, MM_LOAD(c0z+1*SIMDD) * MM_LOAD(g+55*SIMDD) + b01 * (MM_LOAD(g+51*SIMDD) + MM_LOAD(g+53*SIMDD)));
+        MM_STORE(g+2 *SIMDD, rcp0);
+        MM_STORE(g+3 *SIMDD, rcp1);
+        MM_STORE(g+4 *SIMDD, cpx0);
+        MM_STORE(g+5 *SIMDD, cpx1);
+        MM_STORE(g+12*SIMDD, c0x0);
+        MM_STORE(g+13*SIMDD, c0x1);
+        MM_STORE(g+26*SIMDD, rcp2);
+        MM_STORE(g+27*SIMDD, rcp3);
+        MM_STORE(g+28*SIMDD, cpy0);
+        MM_STORE(g+29*SIMDD, cpy1);
+        MM_STORE(g+36*SIMDD, c0y0);
+        MM_STORE(g+37*SIMDD, c0y1);
+        MM_STORE(g+50*SIMDD, rcp4 * g48);
+        MM_STORE(g+51*SIMDD, rcp5 * g49);
+        MM_STORE(g+52*SIMDD, cpz0 * g48);
+        MM_STORE(g+53*SIMDD, cpz1 * g49);
+        MM_STORE(g+60*SIMDD, c0z0 * g48);
+        MM_STORE(g+61*SIMDD, c0z1 * g49);
+        MM_STORE(g+14*SIMDD, c0x0 * rcp0 + b00);
+        MM_STORE(g+15*SIMDD, c0x1 * rcp1 + b01);
+        MM_STORE(g+16*SIMDD, c0x0 * cpx0 + b00);
+        MM_STORE(g+17*SIMDD, c0x1 * cpx1 + b01);
+        MM_STORE(g+6 *SIMDD, cpx0 * rcp0 + b10);
+        MM_STORE(g+7 *SIMDD, cpx1 * rcp1 + b11);
+        MM_STORE(g+30*SIMDD, cpy0 * rcp2 + b10);
+        MM_STORE(g+31*SIMDD, cpy1 * rcp3 + b11);
+        MM_STORE(g+38*SIMDD, c0y0 * rcp2 + b00);
+        MM_STORE(g+39*SIMDD, c0y1 * rcp3 + b01);
+        MM_STORE(g+40*SIMDD, c0y0 * cpy0 + b00);
+        MM_STORE(g+41*SIMDD, c0y1 * cpy1 + b01);
+        MM_STORE(g+54*SIMDD,(cpz0 * rcp4 + b10) * g48);
+        MM_STORE(g+55*SIMDD,(cpz1 * rcp5 + b11) * g49);
+        MM_STORE(g+62*SIMDD,(c0z0 * rcp4 + b00) * g48);
+        MM_STORE(g+63*SIMDD,(c0z1 * rcp5 + b01) * g49);
+        MM_STORE(g+64*SIMDD,(c0z0 * cpz0 + b00) * g48);
+        MM_STORE(g+65*SIMDD,(c0z1 * cpz1 + b01) * g49);
+        MM_STORE(g+18*SIMDD, c0x0 * MM_LOAD(g+6 *SIMDD) + b00 * (rcp0 + cpx0));
+        MM_STORE(g+19*SIMDD, c0x1 * MM_LOAD(g+7 *SIMDD) + b01 * (rcp1 + cpx1));
+        MM_STORE(g+42*SIMDD, c0y0 * MM_LOAD(g+30*SIMDD) + b00 * (rcp2 + cpy0));
+        MM_STORE(g+43*SIMDD, c0y1 * MM_LOAD(g+31*SIMDD) + b01 * (rcp3 + cpy1));
+        MM_STORE(g+66*SIMDD, c0z0 * MM_LOAD(g+54*SIMDD) + b00 * (MM_LOAD(g+50*SIMDD) + MM_LOAD(g+52*SIMDD)));
+        MM_STORE(g+67*SIMDD, c0z1 * MM_LOAD(g+55*SIMDD) + b01 * (MM_LOAD(g+51*SIMDD) + MM_LOAD(g+53*SIMDD)));
 }
 
 static inline void _g0_2d4d_0120(double *g, Rys2eT *bc, CINTEnvVars *envs)
@@ -1417,6 +1441,18 @@ static inline void _g0_2d4d_0201(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cpz = bc->c0pz;
         double *b0  = bc->b00;
         double *b1  = bc->b10;
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
         __MD b00 = MM_LOAD(b0+0*SIMDD);
         __MD b01 = MM_LOAD(b0+1*SIMDD);
         __MD b10 = MM_LOAD(b1+0*SIMDD);
@@ -1424,36 +1460,36 @@ static inline void _g0_2d4d_0201(double *g, Rys2eT *bc, CINTEnvVars *envs)
         __MD g24 = MM_LOAD(g+24*SIMDD);
         __MD g25 = MM_LOAD(g+25*SIMDD);
         __MD i2 = MM_SET1(2.);
-        MM_STORE(g+2 *SIMDD, MM_LOAD(cpx+0*SIMDD));
-        MM_STORE(g+3 *SIMDD, MM_LOAD(cpx+1*SIMDD));
-        MM_STORE(g+4 *SIMDD, MM_LOAD(c0x+0*SIMDD));
-        MM_STORE(g+5 *SIMDD, MM_LOAD(c0x+1*SIMDD));
-        MM_STORE(g+14*SIMDD, MM_LOAD(cpy+0*SIMDD));
-        MM_STORE(g+15*SIMDD, MM_LOAD(cpy+1*SIMDD));
-        MM_STORE(g+16*SIMDD, MM_LOAD(c0y+0*SIMDD));
-        MM_STORE(g+17*SIMDD, MM_LOAD(c0y+1*SIMDD));
-        MM_STORE(g+26*SIMDD, MM_LOAD(cpz+0*SIMDD) * g24);
-        MM_STORE(g+27*SIMDD, MM_LOAD(cpz+1*SIMDD) * g25);
-        MM_STORE(g+28*SIMDD, MM_LOAD(c0z+0*SIMDD) * g24);
-        MM_STORE(g+29*SIMDD, MM_LOAD(c0z+1*SIMDD) * g25);
-        MM_STORE(g+6 *SIMDD, MM_LOAD(cpx+0*SIMDD) * MM_LOAD(c0x+0*SIMDD) + b00);
-        MM_STORE(g+7 *SIMDD, MM_LOAD(cpx+1*SIMDD) * MM_LOAD(c0x+1*SIMDD) + b01);
-        MM_STORE(g+8 *SIMDD, MM_LOAD(c0x+0*SIMDD) * MM_LOAD(c0x+0*SIMDD) + b10);
-        MM_STORE(g+9 *SIMDD, MM_LOAD(c0x+1*SIMDD) * MM_LOAD(c0x+1*SIMDD) + b11);
-        MM_STORE(g+18*SIMDD, MM_LOAD(cpy+0*SIMDD) * MM_LOAD(c0y+0*SIMDD) + b00);
-        MM_STORE(g+19*SIMDD, MM_LOAD(cpy+1*SIMDD) * MM_LOAD(c0y+1*SIMDD) + b01);
-        MM_STORE(g+20*SIMDD, MM_LOAD(c0y+0*SIMDD) * MM_LOAD(c0y+0*SIMDD) + b10);
-        MM_STORE(g+21*SIMDD, MM_LOAD(c0y+1*SIMDD) * MM_LOAD(c0y+1*SIMDD) + b11);
-        MM_STORE(g+30*SIMDD,(MM_LOAD(cpz+0*SIMDD) * MM_LOAD(c0z+0*SIMDD) + b00) * g24);
-        MM_STORE(g+31*SIMDD,(MM_LOAD(cpz+1*SIMDD) * MM_LOAD(c0z+1*SIMDD) + b01) * g25);
-        MM_STORE(g+32*SIMDD,(MM_LOAD(c0z+0*SIMDD) * MM_LOAD(c0z+0*SIMDD) + b10) * g24);
-        MM_STORE(g+33*SIMDD,(MM_LOAD(c0z+1*SIMDD) * MM_LOAD(c0z+1*SIMDD) + b11) * g25);
-        MM_STORE(g+10*SIMDD, MM_LOAD(cpx+0*SIMDD) * MM_LOAD(g+8 *SIMDD ) + i2 * b00 * MM_LOAD(c0x+0*SIMDD));
-        MM_STORE(g+11*SIMDD, MM_LOAD(cpx+1*SIMDD) * MM_LOAD(g+9 *SIMDD ) + i2 * b01 * MM_LOAD(c0x+1*SIMDD));
-        MM_STORE(g+22*SIMDD, MM_LOAD(cpy+0*SIMDD) * MM_LOAD(g+20*SIMDD ) + i2 * b00 * MM_LOAD(c0y+0*SIMDD));
-        MM_STORE(g+23*SIMDD, MM_LOAD(cpy+1*SIMDD) * MM_LOAD(g+21*SIMDD ) + i2 * b01 * MM_LOAD(c0y+1*SIMDD));
-        MM_STORE(g+34*SIMDD, MM_LOAD(cpz+0*SIMDD) * MM_LOAD(g+32*SIMDD ) + i2 * b00 * MM_LOAD(g+28*SIMDD));
-        MM_STORE(g+35*SIMDD, MM_LOAD(cpz+1*SIMDD) * MM_LOAD(g+33*SIMDD ) + i2 * b01 * MM_LOAD(g+29*SIMDD));
+        MM_STORE(g+2 *SIMDD, cpx0);
+        MM_STORE(g+3 *SIMDD, cpx1);
+        MM_STORE(g+4 *SIMDD, c0x0);
+        MM_STORE(g+5 *SIMDD, c0x1);
+        MM_STORE(g+14*SIMDD, cpy0);
+        MM_STORE(g+15*SIMDD, cpy1);
+        MM_STORE(g+16*SIMDD, c0y0);
+        MM_STORE(g+17*SIMDD, c0y1);
+        MM_STORE(g+26*SIMDD, cpz0 * g24);
+        MM_STORE(g+27*SIMDD, cpz1 * g25);
+        MM_STORE(g+28*SIMDD, c0z0 * g24);
+        MM_STORE(g+29*SIMDD, c0z1 * g25);
+        MM_STORE(g+6 *SIMDD, cpx0 * c0x0 + b00);
+        MM_STORE(g+7 *SIMDD, cpx1 * c0x1 + b01);
+        MM_STORE(g+8 *SIMDD, c0x0 * c0x0 + b10);
+        MM_STORE(g+9 *SIMDD, c0x1 * c0x1 + b11);
+        MM_STORE(g+18*SIMDD, cpy0 * c0y0 + b00);
+        MM_STORE(g+19*SIMDD, cpy1 * c0y1 + b01);
+        MM_STORE(g+20*SIMDD, c0y0 * c0y0 + b10);
+        MM_STORE(g+21*SIMDD, c0y1 * c0y1 + b11);
+        MM_STORE(g+30*SIMDD,(cpz0 * c0z0 + b00) * g24);
+        MM_STORE(g+31*SIMDD,(cpz1 * c0z1 + b01) * g25);
+        MM_STORE(g+32*SIMDD,(c0z0 * c0z0 + b10) * g24);
+        MM_STORE(g+33*SIMDD,(c0z1 * c0z1 + b11) * g25);
+        MM_STORE(g+10*SIMDD, cpx0 * MM_LOAD(g+8 *SIMDD ) + i2 * b00 * c0x0);
+        MM_STORE(g+11*SIMDD, cpx1 * MM_LOAD(g+9 *SIMDD ) + i2 * b01 * c0x1);
+        MM_STORE(g+22*SIMDD, cpy0 * MM_LOAD(g+20*SIMDD ) + i2 * b00 * c0y0);
+        MM_STORE(g+23*SIMDD, cpy1 * MM_LOAD(g+21*SIMDD ) + i2 * b01 * c0y1);
+        MM_STORE(g+34*SIMDD, cpz0 * MM_LOAD(g+32*SIMDD ) + i2 * b00 * MM_LOAD(g+28*SIMDD));
+        MM_STORE(g+35*SIMDD, cpz1 * MM_LOAD(g+33*SIMDD ) + i2 * b01 * MM_LOAD(g+29*SIMDD));
 }
 
 static inline void _g0_2d4d_0210(double *g, Rys2eT *bc, CINTEnvVars *envs)
@@ -1465,7 +1501,7 @@ static inline void _g0_2d4d_0210(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b0 = bc->b00;
-        double *b1 = bc->b01;
+        double *b1 = bc->b10;
         __MD r0 = MM_LOAD(c0x+0*SIMDD);
         __MD r1 = MM_LOAD(c0x+1*SIMDD);
         __MD r2 = MM_LOAD(c0y+0*SIMDD);
@@ -1522,29 +1558,35 @@ static inline void _g0_2d4d_0300(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cy = bc->c00y;
         double *cz = bc->c00z;
         double *b =  bc->b10;
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
         __MD b0 = MM_LOAD(b+0*SIMDD);
         __MD b1 = MM_LOAD(b+1*SIMDD);
         __MD g16 = MM_LOAD(g+16*SIMDD);
         __MD g17 = MM_LOAD(g+17*SIMDD);
         __MD i3 = MM_SET1(3.);
-        MM_STORE(g+2 *SIMDD, MM_LOAD(cx+0*SIMDD));
-        MM_STORE(g+3 *SIMDD, MM_LOAD(cx+1*SIMDD));
-        MM_STORE(g+10*SIMDD, MM_LOAD(cy+0*SIMDD));
-        MM_STORE(g+11*SIMDD, MM_LOAD(cy+1*SIMDD));
-        MM_STORE(g+18*SIMDD, MM_LOAD(cz+0*SIMDD) * g16);
-        MM_STORE(g+19*SIMDD, MM_LOAD(cz+1*SIMDD) * g17);
-        MM_STORE(g+4 *SIMDD, MM_LOAD(cx+0*SIMDD) * MM_LOAD(cx+0*SIMDD) + b0);
-        MM_STORE(g+5 *SIMDD, MM_LOAD(cx+1*SIMDD) * MM_LOAD(cx+1*SIMDD) + b1);
-        MM_STORE(g+12*SIMDD, MM_LOAD(cy+0*SIMDD) * MM_LOAD(cy+0*SIMDD) + b0);
-        MM_STORE(g+13*SIMDD, MM_LOAD(cy+1*SIMDD) * MM_LOAD(cy+1*SIMDD) + b1);
-        MM_STORE(g+20*SIMDD,(MM_LOAD(cz+0*SIMDD) * MM_LOAD(cz+0*SIMDD) + b0)* g16);
-        MM_STORE(g+21*SIMDD,(MM_LOAD(cz+1*SIMDD) * MM_LOAD(cz+1*SIMDD) + b1)* g17);
-        MM_STORE(g+6 *SIMDD, MM_LOAD(cx+0*SIMDD) *(MM_LOAD(cx+0*SIMDD) * MM_LOAD(cx+0*SIMDD) + i3 * b0));
-        MM_STORE(g+7 *SIMDD, MM_LOAD(cx+1*SIMDD) *(MM_LOAD(cx+1*SIMDD) * MM_LOAD(cx+1*SIMDD) + i3 * b1));
-        MM_STORE(g+14*SIMDD, MM_LOAD(cy+0*SIMDD) *(MM_LOAD(cy+0*SIMDD) * MM_LOAD(cy+0*SIMDD) + i3 * b0));
-        MM_STORE(g+15*SIMDD, MM_LOAD(cy+1*SIMDD) *(MM_LOAD(cy+1*SIMDD) * MM_LOAD(cy+1*SIMDD) + i3 * b1));
-        MM_STORE(g+22*SIMDD,(MM_LOAD(cz+0*SIMDD) * MM_LOAD(cz+0*SIMDD) + i3 * b0)* MM_LOAD(g+18*SIMDD));
-        MM_STORE(g+23*SIMDD,(MM_LOAD(cz+1*SIMDD) * MM_LOAD(cz+1*SIMDD) + i3 * b1)* MM_LOAD(g+19*SIMDD));
+        MM_STORE(g+2 *SIMDD, cx0);
+        MM_STORE(g+3 *SIMDD, cx1);
+        MM_STORE(g+10*SIMDD, cy0);
+        MM_STORE(g+11*SIMDD, cy1);
+        MM_STORE(g+18*SIMDD, cz0 * g16);
+        MM_STORE(g+19*SIMDD, cz1 * g17);
+        MM_STORE(g+4 *SIMDD, cx0 * cx0 + b0);
+        MM_STORE(g+5 *SIMDD, cx1 * cx1 + b1);
+        MM_STORE(g+12*SIMDD, cy0 * cy0 + b0);
+        MM_STORE(g+13*SIMDD, cy1 * cy1 + b1);
+        MM_STORE(g+20*SIMDD,(cz0 * cz0 + b0)* g16);
+        MM_STORE(g+21*SIMDD,(cz1 * cz1 + b1)* g17);
+        MM_STORE(g+6 *SIMDD, cx0 *(cx0 * cx0 + i3 * b0));
+        MM_STORE(g+7 *SIMDD, cx1 *(cx1 * cx1 + i3 * b1));
+        MM_STORE(g+14*SIMDD, cy0 *(cy0 * cy0 + i3 * b0));
+        MM_STORE(g+15*SIMDD, cy1 *(cy1 * cy1 + i3 * b1));
+        MM_STORE(g+22*SIMDD,(cz0 * cz0 + i3 * b0)* MM_LOAD(g+18*SIMDD));
+        MM_STORE(g+23*SIMDD,(cz1 * cz1 + i3 * b1)* MM_LOAD(g+19*SIMDD));
 }
 
 static inline void _g0_2d4d_1000(double *g, Rys2eT *bc, CINTEnvVars *envs)
@@ -1578,22 +1620,22 @@ static inline void _g0_2d4d_1001(double *g, Rys2eT *bc, CINTEnvVars *envs)
         __MD py1 = MM_LOAD(cpy+1*SIMDD);
         __MD pz0 = MM_LOAD(cpz+0*SIMDD);
         __MD pz1 = MM_LOAD(cpz+1*SIMDD);
-        __MD b0 = MM_LOAD(b+0*SIMDD);
-        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b0  = MM_LOAD(b+0*SIMDD);
+        __MD b1  = MM_LOAD(b+1*SIMDD);
         __MD g16 = MM_LOAD(g+16*SIMDD);
         __MD g17 = MM_LOAD(g+17*SIMDD);
-        MM_STORE(g+2 *SIMDD, px0);
-        MM_STORE(g+3 *SIMDD, px1);
-        MM_STORE(g+4 *SIMDD, cx0);
-        MM_STORE(g+5 *SIMDD, cx1);
-        MM_STORE(g+10*SIMDD, py0);
-        MM_STORE(g+11*SIMDD, py1);
-        MM_STORE(g+12*SIMDD, cy0);
-        MM_STORE(g+13*SIMDD, cy1);
-        MM_STORE(g+18*SIMDD, MM_MUL(pz0, g16));
-        MM_STORE(g+19*SIMDD, MM_MUL(pz1, g17));
-        MM_STORE(g+20*SIMDD, MM_MUL(cz0, g16));
-        MM_STORE(g+21*SIMDD, MM_MUL(cz1, g17));
+        MM_STORE(g+2 *SIMDD, cx0);
+        MM_STORE(g+3 *SIMDD, cx1);
+        MM_STORE(g+4 *SIMDD, px0);
+        MM_STORE(g+5 *SIMDD, px1);
+        MM_STORE(g+10*SIMDD, cy0);
+        MM_STORE(g+11*SIMDD, cy1);
+        MM_STORE(g+12*SIMDD, py0);
+        MM_STORE(g+13*SIMDD, py1);
+        MM_STORE(g+18*SIMDD, MM_MUL(cz0, g16));
+        MM_STORE(g+19*SIMDD, MM_MUL(cz1, g17));
+        MM_STORE(g+20*SIMDD, MM_MUL(pz0, g16));
+        MM_STORE(g+21*SIMDD, MM_MUL(pz1, g17));
         MM_STORE(g+6 *SIMDD, MM_FMA(px0, cx0, b0));
         MM_STORE(g+7 *SIMDD, MM_FMA(px1, cx1, b1));
         MM_STORE(g+14*SIMDD, MM_FMA(py0, cy0, b0));
@@ -1683,22 +1725,22 @@ static inline void _g0_2d4d_1010(double *g, Rys2eT *bc, CINTEnvVars *envs)
         __MD py1 = MM_LOAD(cpy+1*SIMDD);
         __MD pz0 = MM_LOAD(cpz+0*SIMDD);
         __MD pz1 = MM_LOAD(cpz+1*SIMDD);
-        __MD b0 = MM_LOAD(b+0*SIMDD);
-        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b0  = MM_LOAD(b+0*SIMDD);
+        __MD b1  = MM_LOAD(b+1*SIMDD);
         __MD g16 = MM_LOAD(g+16*SIMDD);
         __MD g17 = MM_LOAD(g+17*SIMDD);
-        MM_STORE(g+2 *SIMDD, px0);
-        MM_STORE(g+3 *SIMDD, px1);
-        MM_STORE(g+4 *SIMDD, cx0);
-        MM_STORE(g+5 *SIMDD, cx1);
-        MM_STORE(g+10*SIMDD, py0);
-        MM_STORE(g+11*SIMDD, py1);
-        MM_STORE(g+12*SIMDD, cy0);
-        MM_STORE(g+13*SIMDD, cy1);
-        MM_STORE(g+18*SIMDD, MM_MUL(pz0, g16));
-        MM_STORE(g+19*SIMDD, MM_MUL(pz1, g17));
-        MM_STORE(g+20*SIMDD, MM_MUL(cz0, g16));
-        MM_STORE(g+21*SIMDD, MM_MUL(cz1, g17));
+        MM_STORE(g+2 *SIMDD, cx0);
+        MM_STORE(g+3 *SIMDD, cx1);
+        MM_STORE(g+4 *SIMDD, px0);
+        MM_STORE(g+5 *SIMDD, px1);
+        MM_STORE(g+10*SIMDD, cy0);
+        MM_STORE(g+11*SIMDD, cy1);
+        MM_STORE(g+12*SIMDD, py0);
+        MM_STORE(g+13*SIMDD, py1);
+        MM_STORE(g+18*SIMDD, MM_MUL(cz0, g16));
+        MM_STORE(g+19*SIMDD, MM_MUL(cz1, g17));
+        MM_STORE(g+20*SIMDD, MM_MUL(pz0, g16));
+        MM_STORE(g+21*SIMDD, MM_MUL(pz1, g17));
         MM_STORE(g+6 *SIMDD, MM_FMA(px0, cx0, b0));
         MM_STORE(g+7 *SIMDD, MM_FMA(px1, cx1, b1));
         MM_STORE(g+14*SIMDD, MM_FMA(py0, cy0, b0));
@@ -1717,57 +1759,76 @@ static inline void _g0_2d4d_1011(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cpz = bc->c0pz;
         double *b0  = bc->b00;
         double *b1  = bc->b01;
-        double *rkl = envs->rkrl;
-        ALIGNMM double rcp[6*SIMDD];
-        _make_rc(rcp, cpx, cpy, cpz, rkl);
+        double *r   = envs->rkrl;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD rcp0 = MM_ADD(rx, cpx0);
+        __MD rcp1 = MM_ADD(rx, cpx1);
+        __MD rcp2 = MM_ADD(ry, cpy0);
+        __MD rcp3 = MM_ADD(ry, cpy1);
+        __MD rcp4 = MM_ADD(rz, cpz0);
+        __MD rcp5 = MM_ADD(rz, cpz1);
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
         __MD b00 = MM_LOAD(b0+0*SIMDD);
         __MD b01 = MM_LOAD(b0+1*SIMDD);
         __MD b10 = MM_LOAD(b1+0*SIMDD);
         __MD b11 = MM_LOAD(b1+1*SIMDD);
         __MD g48 = MM_LOAD(g+48*SIMDD);
         __MD g49 = MM_LOAD(g+49*SIMDD);
-        MM_STORE(g+2 *SIMDD, MM_LOAD(c0x+0*SIMDD));
-        MM_STORE(g+3 *SIMDD, MM_LOAD(c0x+1*SIMDD));
-        MM_STORE(g+4 *SIMDD, MM_LOAD(rcp+0*SIMDD));
-        MM_STORE(g+5 *SIMDD, MM_LOAD(rcp+1*SIMDD));
-        MM_STORE(g+8 *SIMDD, MM_LOAD(cpx+0*SIMDD));
-        MM_STORE(g+9 *SIMDD, MM_LOAD(cpx+1*SIMDD));
-        MM_STORE(g+26*SIMDD, MM_LOAD(c0y+0*SIMDD));
-        MM_STORE(g+27*SIMDD, MM_LOAD(c0y+1*SIMDD));
-        MM_STORE(g+28*SIMDD, MM_LOAD(rcp+2*SIMDD));
-        MM_STORE(g+29*SIMDD, MM_LOAD(rcp+3*SIMDD));
-        MM_STORE(g+32*SIMDD, MM_LOAD(cpy+0*SIMDD));
-        MM_STORE(g+33*SIMDD, MM_LOAD(cpy+1*SIMDD));
-        MM_STORE(g+50*SIMDD, MM_LOAD(c0z+0*SIMDD) * g48);
-        MM_STORE(g+51*SIMDD, MM_LOAD(c0z+1*SIMDD) * g49);
-        MM_STORE(g+52*SIMDD, MM_LOAD(rcp+4*SIMDD) * g48);
-        MM_STORE(g+53*SIMDD, MM_LOAD(rcp+5*SIMDD) * g49);
-        MM_STORE(g+56*SIMDD, MM_LOAD(cpz+0*SIMDD) * g48);
-        MM_STORE(g+57*SIMDD, MM_LOAD(cpz+1*SIMDD) * g49);
-        MM_STORE(g+6 *SIMDD , MM_LOAD(rcp+0*SIMDD) * MM_LOAD(c0x+0*SIMDD) + b00);
-        MM_STORE(g+7 *SIMDD , MM_LOAD(rcp+1*SIMDD) * MM_LOAD(c0x+1*SIMDD) + b01);
-        MM_STORE(g+10*SIMDD , MM_LOAD(cpx+0*SIMDD) * MM_LOAD(c0x+0*SIMDD) + b00);
-        MM_STORE(g+11*SIMDD , MM_LOAD(cpx+1*SIMDD) * MM_LOAD(c0x+1*SIMDD) + b01);
-        MM_STORE(g+12*SIMDD , MM_LOAD(cpx+0*SIMDD) * MM_LOAD(rcp+0*SIMDD) + b10);
-        MM_STORE(g+13*SIMDD , MM_LOAD(cpx+1*SIMDD) * MM_LOAD(rcp+1*SIMDD) + b11);
-        MM_STORE(g+30*SIMDD , MM_LOAD(rcp+2*SIMDD) * MM_LOAD(c0y+0*SIMDD) + b00);
-        MM_STORE(g+31*SIMDD , MM_LOAD(rcp+3*SIMDD) * MM_LOAD(c0y+1*SIMDD) + b01);
-        MM_STORE(g+34*SIMDD , MM_LOAD(cpy+0*SIMDD) * MM_LOAD(c0y+0*SIMDD) + b00);
-        MM_STORE(g+35*SIMDD , MM_LOAD(cpy+1*SIMDD) * MM_LOAD(c0y+1*SIMDD) + b01);
-        MM_STORE(g+36*SIMDD , MM_LOAD(cpy+0*SIMDD) * MM_LOAD(rcp+2*SIMDD) + b10);
-        MM_STORE(g+37*SIMDD , MM_LOAD(cpy+1*SIMDD) * MM_LOAD(rcp+3*SIMDD) + b11);
-        MM_STORE(g+54*SIMDD,(MM_LOAD(rcp+4*SIMDD) * MM_LOAD(c0z+0*SIMDD) + b00)* g48);
-        MM_STORE(g+55*SIMDD,(MM_LOAD(rcp+5*SIMDD) * MM_LOAD(c0z+1*SIMDD) + b01)* g49);
-        MM_STORE(g+58*SIMDD,(MM_LOAD(cpz+0*SIMDD) * MM_LOAD(c0z+0*SIMDD) + b00)* g48);
-        MM_STORE(g+59*SIMDD,(MM_LOAD(cpz+1*SIMDD) * MM_LOAD(c0z+1*SIMDD) + b01)* g49);
-        MM_STORE(g+60*SIMDD,(MM_LOAD(cpz+0*SIMDD) * MM_LOAD(rcp+4*SIMDD) + b10)* g48);
-        MM_STORE(g+61*SIMDD,(MM_LOAD(cpz+1*SIMDD) * MM_LOAD(rcp+5*SIMDD) + b11)* g49);
-        MM_STORE(g+14*SIMDD, MM_LOAD(rcp+0*SIMDD) * MM_LOAD(g+10*SIMDD) + b00 * MM_LOAD(cpx+0*SIMDD) + b10 * MM_LOAD(c0x+0*SIMDD));
-        MM_STORE(g+15*SIMDD, MM_LOAD(rcp+1*SIMDD) * MM_LOAD(g+11*SIMDD) + b01 * MM_LOAD(cpx+1*SIMDD) + b11 * MM_LOAD(c0x+1*SIMDD));
-        MM_STORE(g+38*SIMDD, MM_LOAD(rcp+2*SIMDD) * MM_LOAD(g+34*SIMDD) + b00 * MM_LOAD(cpy+0*SIMDD) + b10 * MM_LOAD(c0y+0*SIMDD));
-        MM_STORE(g+39*SIMDD, MM_LOAD(rcp+3*SIMDD) * MM_LOAD(g+35*SIMDD) + b01 * MM_LOAD(cpy+1*SIMDD) + b11 * MM_LOAD(c0y+1*SIMDD));
-        MM_STORE(g+62*SIMDD, MM_LOAD(rcp+4*SIMDD) * MM_LOAD(g+58*SIMDD) + b00 * MM_LOAD(g+56*SIMDD) + b10 * MM_LOAD(g+50*SIMDD));
-        MM_STORE(g+63*SIMDD, MM_LOAD(rcp+5*SIMDD) * MM_LOAD(g+59*SIMDD) + b01 * MM_LOAD(g+57*SIMDD) + b11 * MM_LOAD(g+51*SIMDD));
+        MM_STORE(g+2 *SIMDD, c0x0);
+        MM_STORE(g+3 *SIMDD, c0x1);
+        MM_STORE(g+4 *SIMDD, rcp0);
+        MM_STORE(g+5 *SIMDD, rcp1);
+        MM_STORE(g+8 *SIMDD, cpx0);
+        MM_STORE(g+9 *SIMDD, cpx1);
+        MM_STORE(g+26*SIMDD, c0y0);
+        MM_STORE(g+27*SIMDD, c0y1);
+        MM_STORE(g+28*SIMDD, rcp2);
+        MM_STORE(g+29*SIMDD, rcp3);
+        MM_STORE(g+32*SIMDD, cpy0);
+        MM_STORE(g+33*SIMDD, cpy1);
+        MM_STORE(g+50*SIMDD, c0z0 * g48);
+        MM_STORE(g+51*SIMDD, c0z1 * g49);
+        MM_STORE(g+52*SIMDD, rcp4 * g48);
+        MM_STORE(g+53*SIMDD, rcp5 * g49);
+        MM_STORE(g+56*SIMDD, cpz0 * g48);
+        MM_STORE(g+57*SIMDD, cpz1 * g49);
+        MM_STORE(g+6 *SIMDD, rcp0 * c0x0 + b00);
+        MM_STORE(g+7 *SIMDD, rcp1 * c0x1 + b01);
+        MM_STORE(g+10*SIMDD, cpx0 * c0x0 + b00);
+        MM_STORE(g+11*SIMDD, cpx1 * c0x1 + b01);
+        MM_STORE(g+12*SIMDD, cpx0 * rcp0 + b10);
+        MM_STORE(g+13*SIMDD, cpx1 * rcp1 + b11);
+        MM_STORE(g+30*SIMDD, rcp2 * c0y0 + b00);
+        MM_STORE(g+31*SIMDD, rcp3 * c0y1 + b01);
+        MM_STORE(g+34*SIMDD, cpy0 * c0y0 + b00);
+        MM_STORE(g+35*SIMDD, cpy1 * c0y1 + b01);
+        MM_STORE(g+36*SIMDD, cpy0 * rcp2 + b10);
+        MM_STORE(g+37*SIMDD, cpy1 * rcp3 + b11);
+        MM_STORE(g+54*SIMDD,(rcp4 * c0z0 + b00)* g48);
+        MM_STORE(g+55*SIMDD,(rcp5 * c0z1 + b01)* g49);
+        MM_STORE(g+58*SIMDD,(cpz0 * c0z0 + b00)* g48);
+        MM_STORE(g+59*SIMDD,(cpz1 * c0z1 + b01)* g49);
+        MM_STORE(g+60*SIMDD,(cpz0 * rcp4 + b10)* g48);
+        MM_STORE(g+61*SIMDD,(cpz1 * rcp5 + b11)* g49);
+        MM_STORE(g+14*SIMDD, rcp0 * MM_LOAD(g+10*SIMDD) + b00 * cpx0 + b10 * c0x0);
+        MM_STORE(g+15*SIMDD, rcp1 * MM_LOAD(g+11*SIMDD) + b01 * cpx1 + b11 * c0x1);
+        MM_STORE(g+38*SIMDD, rcp2 * MM_LOAD(g+34*SIMDD) + b00 * cpy0 + b10 * c0y0);
+        MM_STORE(g+39*SIMDD, rcp3 * MM_LOAD(g+35*SIMDD) + b01 * cpy1 + b11 * c0y1);
+        MM_STORE(g+62*SIMDD, rcp4 * MM_LOAD(g+58*SIMDD) + b00 * MM_LOAD(g+56*SIMDD) + b10 * MM_LOAD(g+50*SIMDD));
+        MM_STORE(g+63*SIMDD, rcp5 * MM_LOAD(g+59*SIMDD) + b01 * MM_LOAD(g+57*SIMDD) + b11 * MM_LOAD(g+51*SIMDD));
 }
 
 static inline void _g0_2d4d_1020(double *g, Rys2eT *bc, CINTEnvVars *envs)
@@ -1837,20 +1898,21 @@ static inline void _g0_2d4d_1100(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cz = bc->c00z;
         double *b  = bc->b10;
         double *r  = envs->rirj;
-        ALIGNMM double rc[6*SIMDD];
-        _make_rc(rc, cx, cy, cz, r);
-        __MD r0 = MM_LOAD(rc+0*SIMDD);
-        __MD r1 = MM_LOAD(rc+1*SIMDD);
-        __MD r2 = MM_LOAD(rc+2*SIMDD);
-        __MD r3 = MM_LOAD(rc+3*SIMDD);
-        __MD r4 = MM_LOAD(rc+4*SIMDD);
-        __MD r5 = MM_LOAD(rc+5*SIMDD);
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
         __MD cx0 = MM_LOAD(cx+0*SIMDD);
         __MD cx1 = MM_LOAD(cx+1*SIMDD);
         __MD cy0 = MM_LOAD(cy+0*SIMDD);
         __MD cy1 = MM_LOAD(cy+1*SIMDD);
         __MD cz0 = MM_LOAD(cz+0*SIMDD);
         __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD r0 = MM_ADD(rx, cx0);
+        __MD r1 = MM_ADD(rx, cx1);
+        __MD r2 = MM_ADD(ry, cy0);
+        __MD r3 = MM_ADD(ry, cy1);
+        __MD r4 = MM_ADD(rz, cz0);
+        __MD r5 = MM_ADD(rz, cz1);
         __MD b0 = MM_LOAD(b+0*SIMDD);
         __MD b1 = MM_LOAD(b+1*SIMDD);
         __MD g24 = MM_LOAD(g+24*SIMDD);
@@ -1885,57 +1947,76 @@ static inline void _g0_2d4d_1101(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cpz = bc->c0pz;
         double *b0  = bc->b00;
         double *b1  = bc->b10;
-        double *rij = envs->rirj;
-        ALIGNMM double rc0[6*SIMDD];
-        _make_rc(rc0, c0x, c0y, c0z, rij);
+        double *r   = envs->rirj;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
+        __MD rc00 = MM_ADD(rx, c0x0);
+        __MD rc01 = MM_ADD(rx, c0x1);
+        __MD rc02 = MM_ADD(ry, c0y0);
+        __MD rc03 = MM_ADD(ry, c0y1);
+        __MD rc04 = MM_ADD(rz, c0z0);
+        __MD rc05 = MM_ADD(rz, c0z1);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
         __MD b00 = MM_LOAD(b0+0*SIMDD);
         __MD b01 = MM_LOAD(b0+1*SIMDD);
         __MD b10 = MM_LOAD(b1+0*SIMDD);
         __MD b11 = MM_LOAD(b1+1*SIMDD);
         __MD g48 = MM_LOAD(g+48*SIMDD);
         __MD g49 = MM_LOAD(g+49*SIMDD);
-        MM_STORE(g+2 *SIMDD, MM_LOAD(rc0+0*SIMDD));
-        MM_STORE(g+3 *SIMDD, MM_LOAD(rc0+1*SIMDD));
-        MM_STORE(g+4 *SIMDD, MM_LOAD(cpx+0*SIMDD));
-        MM_STORE(g+5 *SIMDD, MM_LOAD(cpx+1*SIMDD));
-        MM_STORE(g+8 *SIMDD, MM_LOAD(c0x+0*SIMDD));
-        MM_STORE(g+9 *SIMDD, MM_LOAD(c0x+1*SIMDD));
-        MM_STORE(g+26*SIMDD, MM_LOAD(rc0+2*SIMDD));
-        MM_STORE(g+27*SIMDD, MM_LOAD(rc0+3*SIMDD));
-        MM_STORE(g+28*SIMDD, MM_LOAD(cpy+0*SIMDD));
-        MM_STORE(g+29*SIMDD, MM_LOAD(cpy+1*SIMDD));
-        MM_STORE(g+32*SIMDD, MM_LOAD(c0y+0*SIMDD));
-        MM_STORE(g+33*SIMDD, MM_LOAD(c0y+1*SIMDD));
-        MM_STORE(g+50*SIMDD, MM_LOAD(rc0+4*SIMDD) * g48);
-        MM_STORE(g+51*SIMDD, MM_LOAD(rc0+5*SIMDD) * g49);
-        MM_STORE(g+52*SIMDD, MM_LOAD(cpz+0*SIMDD) * g48);
-        MM_STORE(g+53*SIMDD, MM_LOAD(cpz+1*SIMDD) * g49);
-        MM_STORE(g+56*SIMDD, MM_LOAD(c0z+0*SIMDD) * g48);
-        MM_STORE(g+57*SIMDD, MM_LOAD(c0z+1*SIMDD) * g49);
-        MM_STORE(g+6 *SIMDD, MM_LOAD(cpx+0*SIMDD) * MM_LOAD(rc0+0*SIMDD) + b00);
-        MM_STORE(g+7 *SIMDD, MM_LOAD(cpx+1*SIMDD) * MM_LOAD(rc0+1*SIMDD) + b01);
-        MM_STORE(g+10*SIMDD, MM_LOAD(c0x+0*SIMDD) * MM_LOAD(rc0+0*SIMDD) + b10);
-        MM_STORE(g+11*SIMDD, MM_LOAD(c0x+1*SIMDD) * MM_LOAD(rc0+1*SIMDD) + b11);
-        MM_STORE(g+12*SIMDD, MM_LOAD(c0x+0*SIMDD) * MM_LOAD(cpx+0*SIMDD) + b00);
-        MM_STORE(g+13*SIMDD, MM_LOAD(c0x+1*SIMDD) * MM_LOAD(cpx+1*SIMDD) + b01);
-        MM_STORE(g+30*SIMDD, MM_LOAD(cpy+0*SIMDD) * MM_LOAD(rc0+2*SIMDD) + b00);
-        MM_STORE(g+31*SIMDD, MM_LOAD(cpy+1*SIMDD) * MM_LOAD(rc0+3*SIMDD) + b01);
-        MM_STORE(g+34*SIMDD, MM_LOAD(c0y+0*SIMDD) * MM_LOAD(rc0+2*SIMDD) + b10);
-        MM_STORE(g+35*SIMDD, MM_LOAD(c0y+1*SIMDD) * MM_LOAD(rc0+3*SIMDD) + b11);
-        MM_STORE(g+36*SIMDD, MM_LOAD(c0y+0*SIMDD) * MM_LOAD(cpy+0*SIMDD) + b00);
-        MM_STORE(g+37*SIMDD, MM_LOAD(c0y+1*SIMDD) * MM_LOAD(cpy+1*SIMDD) + b01);
-        MM_STORE(g+54*SIMDD,(MM_LOAD(cpz+0*SIMDD) * MM_LOAD(rc0+4*SIMDD) + b00)* g48);
-        MM_STORE(g+55*SIMDD,(MM_LOAD(cpz+1*SIMDD) * MM_LOAD(rc0+5*SIMDD) + b01)* g49);
-        MM_STORE(g+58*SIMDD,(MM_LOAD(c0z+0*SIMDD) * MM_LOAD(rc0+4*SIMDD) + b10)* g48);
-        MM_STORE(g+59*SIMDD,(MM_LOAD(c0z+1*SIMDD) * MM_LOAD(rc0+5*SIMDD) + b11)* g49);
-        MM_STORE(g+60*SIMDD,(MM_LOAD(c0z+0*SIMDD) * MM_LOAD(cpz+0*SIMDD) + b00)* g48);
-        MM_STORE(g+61*SIMDD,(MM_LOAD(c0z+1*SIMDD) * MM_LOAD(cpz+1*SIMDD) + b01)* g49);
-        MM_STORE(g+14*SIMDD, MM_LOAD(cpx+0*SIMDD) * MM_LOAD(g+10*SIMDD ) + b00 *(MM_LOAD(rc0+0*SIMDD) + MM_LOAD(c0x+0*SIMDD)));
-        MM_STORE(g+15*SIMDD, MM_LOAD(cpx+1*SIMDD) * MM_LOAD(g+11*SIMDD ) + b01 *(MM_LOAD(rc0+1*SIMDD) + MM_LOAD(c0x+1*SIMDD)));
-        MM_STORE(g+38*SIMDD, MM_LOAD(cpy+0*SIMDD) * MM_LOAD(g+34*SIMDD ) + b00 *(MM_LOAD(rc0+2*SIMDD) + MM_LOAD(c0y+0*SIMDD)));
-        MM_STORE(g+39*SIMDD, MM_LOAD(cpy+1*SIMDD) * MM_LOAD(g+35*SIMDD ) + b01 *(MM_LOAD(rc0+3*SIMDD) + MM_LOAD(c0y+1*SIMDD)));
-        MM_STORE(g+62*SIMDD, MM_LOAD(cpz+0*SIMDD) * MM_LOAD(g+58*SIMDD ) + b00 *(MM_LOAD(g+50*SIMDD ) + MM_LOAD(g+56*SIMDD )));
-        MM_STORE(g+63*SIMDD, MM_LOAD(cpz+1*SIMDD) * MM_LOAD(g+59*SIMDD ) + b01 *(MM_LOAD(g+51*SIMDD ) + MM_LOAD(g+57*SIMDD )));
+        MM_STORE(g+2 *SIMDD, rc00);
+        MM_STORE(g+3 *SIMDD, rc01);
+        MM_STORE(g+4 *SIMDD, cpx0);
+        MM_STORE(g+5 *SIMDD, cpx1);
+        MM_STORE(g+8 *SIMDD, c0x0);
+        MM_STORE(g+9 *SIMDD, c0x1);
+        MM_STORE(g+26*SIMDD, rc02);
+        MM_STORE(g+27*SIMDD, rc03);
+        MM_STORE(g+28*SIMDD, cpy0);
+        MM_STORE(g+29*SIMDD, cpy1);
+        MM_STORE(g+32*SIMDD, c0y0);
+        MM_STORE(g+33*SIMDD, c0y1);
+        MM_STORE(g+50*SIMDD, rc04 * g48);
+        MM_STORE(g+51*SIMDD, rc05 * g49);
+        MM_STORE(g+52*SIMDD, cpz0 * g48);
+        MM_STORE(g+53*SIMDD, cpz1 * g49);
+        MM_STORE(g+56*SIMDD, c0z0 * g48);
+        MM_STORE(g+57*SIMDD, c0z1 * g49);
+        MM_STORE(g+6 *SIMDD, cpx0 * rc00 + b00);
+        MM_STORE(g+7 *SIMDD, cpx1 * rc01 + b01);
+        MM_STORE(g+10*SIMDD, c0x0 * rc00 + b10);
+        MM_STORE(g+11*SIMDD, c0x1 * rc01 + b11);
+        MM_STORE(g+12*SIMDD, c0x0 * cpx0 + b00);
+        MM_STORE(g+13*SIMDD, c0x1 * cpx1 + b01);
+        MM_STORE(g+30*SIMDD, cpy0 * rc02 + b00);
+        MM_STORE(g+31*SIMDD, cpy1 * rc03 + b01);
+        MM_STORE(g+34*SIMDD, c0y0 * rc02 + b10);
+        MM_STORE(g+35*SIMDD, c0y1 * rc03 + b11);
+        MM_STORE(g+36*SIMDD, c0y0 * cpy0 + b00);
+        MM_STORE(g+37*SIMDD, c0y1 * cpy1 + b01);
+        MM_STORE(g+54*SIMDD,(cpz0 * rc04 + b00)* g48);
+        MM_STORE(g+55*SIMDD,(cpz1 * rc05 + b01)* g49);
+        MM_STORE(g+58*SIMDD,(c0z0 * rc04 + b10)* g48);
+        MM_STORE(g+59*SIMDD,(c0z1 * rc05 + b11)* g49);
+        MM_STORE(g+60*SIMDD,(c0z0 * cpz0 + b00)* g48);
+        MM_STORE(g+61*SIMDD,(c0z1 * cpz1 + b01)* g49);
+        MM_STORE(g+14*SIMDD, cpx0 * MM_LOAD(g+10*SIMDD ) + b00 *(rc00 + c0x0));
+        MM_STORE(g+15*SIMDD, cpx1 * MM_LOAD(g+11*SIMDD ) + b01 *(rc01 + c0x1));
+        MM_STORE(g+38*SIMDD, cpy0 * MM_LOAD(g+34*SIMDD ) + b00 *(rc02 + c0y0));
+        MM_STORE(g+39*SIMDD, cpy1 * MM_LOAD(g+35*SIMDD ) + b01 *(rc03 + c0y1));
+        MM_STORE(g+62*SIMDD, cpz0 * MM_LOAD(g+58*SIMDD ) + b00 *(MM_LOAD(g+50*SIMDD ) + MM_LOAD(g+56*SIMDD )));
+        MM_STORE(g+63*SIMDD, cpz1 * MM_LOAD(g+59*SIMDD ) + b01 *(MM_LOAD(g+51*SIMDD ) + MM_LOAD(g+57*SIMDD )));
 }
 
 static inline void _g0_2d4d_1110(double *g, Rys2eT *bc, CINTEnvVars *envs)
@@ -1948,57 +2029,76 @@ static inline void _g0_2d4d_1110(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cpz = bc->c0pz;
         double *b0  = bc->b00;
         double *b1  = bc->b10;
-        double *r0  = envs->rirj;
-        ALIGNMM double rc0[6*SIMDD];
-        _make_rc(rc0, c0x, c0y, c0z, r0);
+        double *r   = envs->rirj;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
+        __MD rc00 = MM_ADD(rx, c0x0);
+        __MD rc01 = MM_ADD(rx, c0x1);
+        __MD rc02 = MM_ADD(ry, c0y0);
+        __MD rc03 = MM_ADD(ry, c0y1);
+        __MD rc04 = MM_ADD(rz, c0z0);
+        __MD rc05 = MM_ADD(rz, c0z1);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
         __MD b00 = MM_LOAD(b0+0*SIMDD);
         __MD b01 = MM_LOAD(b0+1*SIMDD);
         __MD b10 = MM_LOAD(b1+0*SIMDD);
         __MD b11 = MM_LOAD(b1+1*SIMDD);
         __MD g48 = MM_LOAD(g+48*SIMDD);
         __MD g49 = MM_LOAD(g+49*SIMDD);
-        MM_STORE(g+2 *SIMDD, MM_LOAD(rc0+0*SIMDD));
-        MM_STORE(g+3 *SIMDD, MM_LOAD(rc0+1*SIMDD));
-        MM_STORE(g+4 *SIMDD, MM_LOAD(cpx+0*SIMDD));
-        MM_STORE(g+5 *SIMDD, MM_LOAD(cpx+1*SIMDD));
-        MM_STORE(g+8 *SIMDD, MM_LOAD(c0x+0*SIMDD));
-        MM_STORE(g+9 *SIMDD, MM_LOAD(c0x+1*SIMDD));
-        MM_STORE(g+26*SIMDD, MM_LOAD(rc0+2*SIMDD));
-        MM_STORE(g+27*SIMDD, MM_LOAD(rc0+3*SIMDD));
-        MM_STORE(g+28*SIMDD, MM_LOAD(cpy+0*SIMDD));
-        MM_STORE(g+29*SIMDD, MM_LOAD(cpy+1*SIMDD));
-        MM_STORE(g+32*SIMDD, MM_LOAD(c0y+0*SIMDD));
-        MM_STORE(g+33*SIMDD, MM_LOAD(c0y+1*SIMDD));
-        MM_STORE(g+50*SIMDD, MM_LOAD(rc0+4*SIMDD) * g48);
-        MM_STORE(g+51*SIMDD, MM_LOAD(rc0+5*SIMDD) * g49);
-        MM_STORE(g+52*SIMDD, MM_LOAD(cpz+0*SIMDD) * g48);
-        MM_STORE(g+53*SIMDD, MM_LOAD(cpz+1*SIMDD) * g49);
-        MM_STORE(g+56*SIMDD, MM_LOAD(c0z+0*SIMDD) * g48);
-        MM_STORE(g+57*SIMDD, MM_LOAD(c0z+1*SIMDD) * g49);
-        MM_STORE(g+6 *SIMDD, MM_LOAD(cpx+0*SIMDD) * MM_LOAD(rc0+0*SIMDD) + b00);
-        MM_STORE(g+7 *SIMDD, MM_LOAD(cpx+1*SIMDD) * MM_LOAD(rc0+1*SIMDD) + b01);
-        MM_STORE(g+10*SIMDD, MM_LOAD(c0x+0*SIMDD) * MM_LOAD(rc0+0*SIMDD) + b10);
-        MM_STORE(g+11*SIMDD, MM_LOAD(c0x+1*SIMDD) * MM_LOAD(rc0+1*SIMDD) + b11);
-        MM_STORE(g+12*SIMDD, MM_LOAD(c0x+0*SIMDD) * MM_LOAD(cpx+0*SIMDD) + b00);
-        MM_STORE(g+13*SIMDD, MM_LOAD(c0x+1*SIMDD) * MM_LOAD(cpx+1*SIMDD) + b01);
-        MM_STORE(g+30*SIMDD, MM_LOAD(cpy+0*SIMDD) * MM_LOAD(rc0+2*SIMDD) + b00);
-        MM_STORE(g+31*SIMDD, MM_LOAD(cpy+1*SIMDD) * MM_LOAD(rc0+3*SIMDD) + b01);
-        MM_STORE(g+34*SIMDD, MM_LOAD(c0y+0*SIMDD) * MM_LOAD(rc0+2*SIMDD) + b10);
-        MM_STORE(g+35*SIMDD, MM_LOAD(c0y+1*SIMDD) * MM_LOAD(rc0+3*SIMDD) + b11);
-        MM_STORE(g+36*SIMDD, MM_LOAD(c0y+0*SIMDD) * MM_LOAD(cpy+0*SIMDD) + b00);
-        MM_STORE(g+37*SIMDD, MM_LOAD(c0y+1*SIMDD) * MM_LOAD(cpy+1*SIMDD) + b01);
-        MM_STORE(g+54*SIMDD,(MM_LOAD(cpz+0*SIMDD) * MM_LOAD(rc0+4*SIMDD) + b00)* g48);
-        MM_STORE(g+55*SIMDD,(MM_LOAD(cpz+1*SIMDD) * MM_LOAD(rc0+5*SIMDD) + b01)* g49);
-        MM_STORE(g+58*SIMDD,(MM_LOAD(c0z+0*SIMDD) * MM_LOAD(rc0+4*SIMDD) + b10)* g48);
-        MM_STORE(g+59*SIMDD,(MM_LOAD(c0z+1*SIMDD) * MM_LOAD(rc0+5*SIMDD) + b11)* g49);
-        MM_STORE(g+60*SIMDD,(MM_LOAD(c0z+0*SIMDD) * MM_LOAD(cpz+0*SIMDD) + b00)* g48);
-        MM_STORE(g+61*SIMDD,(MM_LOAD(c0z+1*SIMDD) * MM_LOAD(cpz+1*SIMDD) + b01)* g49);
-        MM_STORE(g+14*SIMDD, MM_LOAD(cpx+0*SIMDD) * MM_LOAD(g+10*SIMDD ) + b00 *(MM_LOAD(rc0+0*SIMDD) + MM_LOAD(c0x+0*SIMDD)));
-        MM_STORE(g+15*SIMDD, MM_LOAD(cpx+1*SIMDD) * MM_LOAD(g+11*SIMDD ) + b01 *(MM_LOAD(rc0+1*SIMDD) + MM_LOAD(c0x+1*SIMDD)));
-        MM_STORE(g+38*SIMDD, MM_LOAD(cpy+0*SIMDD) * MM_LOAD(g+34*SIMDD ) + b00 *(MM_LOAD(rc0+2*SIMDD) + MM_LOAD(c0y+0*SIMDD)));
-        MM_STORE(g+39*SIMDD, MM_LOAD(cpy+1*SIMDD) * MM_LOAD(g+35*SIMDD ) + b01 *(MM_LOAD(rc0+3*SIMDD) + MM_LOAD(c0y+1*SIMDD)));
-        MM_STORE(g+62*SIMDD, MM_LOAD(cpz+0*SIMDD) * MM_LOAD(g+58*SIMDD ) + b00 *(MM_LOAD(g+50*SIMDD ) + MM_LOAD(g+56*SIMDD )));
-        MM_STORE(g+63*SIMDD, MM_LOAD(cpz+1*SIMDD) * MM_LOAD(g+59*SIMDD ) + b01 *(MM_LOAD(g+51*SIMDD ) + MM_LOAD(g+57*SIMDD )));
+        MM_STORE(g+2 *SIMDD, rc00);
+        MM_STORE(g+3 *SIMDD, rc01);
+        MM_STORE(g+4 *SIMDD, cpx0);
+        MM_STORE(g+5 *SIMDD, cpx1);
+        MM_STORE(g+8 *SIMDD, c0x0);
+        MM_STORE(g+9 *SIMDD, c0x1);
+        MM_STORE(g+26*SIMDD, rc02);
+        MM_STORE(g+27*SIMDD, rc03);
+        MM_STORE(g+28*SIMDD, cpy0);
+        MM_STORE(g+29*SIMDD, cpy1);
+        MM_STORE(g+32*SIMDD, c0y0);
+        MM_STORE(g+33*SIMDD, c0y1);
+        MM_STORE(g+50*SIMDD, rc04 * g48);
+        MM_STORE(g+51*SIMDD, rc05 * g49);
+        MM_STORE(g+52*SIMDD, cpz0 * g48);
+        MM_STORE(g+53*SIMDD, cpz1 * g49);
+        MM_STORE(g+56*SIMDD, c0z0 * g48);
+        MM_STORE(g+57*SIMDD, c0z1 * g49);
+        MM_STORE(g+6 *SIMDD, cpx0 * rc00 + b00);
+        MM_STORE(g+7 *SIMDD, cpx1 * rc01 + b01);
+        MM_STORE(g+10*SIMDD, c0x0 * rc00 + b10);
+        MM_STORE(g+11*SIMDD, c0x1 * rc01 + b11);
+        MM_STORE(g+12*SIMDD, c0x0 * cpx0 + b00);
+        MM_STORE(g+13*SIMDD, c0x1 * cpx1 + b01);
+        MM_STORE(g+30*SIMDD, cpy0 * rc02 + b00);
+        MM_STORE(g+31*SIMDD, cpy1 * rc03 + b01);
+        MM_STORE(g+34*SIMDD, c0y0 * rc02 + b10);
+        MM_STORE(g+35*SIMDD, c0y1 * rc03 + b11);
+        MM_STORE(g+36*SIMDD, c0y0 * cpy0 + b00);
+        MM_STORE(g+37*SIMDD, c0y1 * cpy1 + b01);
+        MM_STORE(g+54*SIMDD,(cpz0 * rc04 + b00)* g48);
+        MM_STORE(g+55*SIMDD,(cpz1 * rc05 + b01)* g49);
+        MM_STORE(g+58*SIMDD,(c0z0 * rc04 + b10)* g48);
+        MM_STORE(g+59*SIMDD,(c0z1 * rc05 + b11)* g49);
+        MM_STORE(g+60*SIMDD,(c0z0 * cpz0 + b00)* g48);
+        MM_STORE(g+61*SIMDD,(c0z1 * cpz1 + b01)* g49);
+        MM_STORE(g+14*SIMDD, cpx0 * MM_LOAD(g+10*SIMDD) + b00 *(rc00 + c0x0));
+        MM_STORE(g+15*SIMDD, cpx1 * MM_LOAD(g+11*SIMDD) + b01 *(rc01 + c0x1));
+        MM_STORE(g+38*SIMDD, cpy0 * MM_LOAD(g+34*SIMDD) + b00 *(rc02 + c0y0));
+        MM_STORE(g+39*SIMDD, cpy1 * MM_LOAD(g+35*SIMDD) + b01 *(rc03 + c0y1));
+        MM_STORE(g+62*SIMDD, cpz0 * MM_LOAD(g+58*SIMDD) + b00 *(MM_LOAD(g+50*SIMDD) + MM_LOAD(g+56*SIMDD)));
+        MM_STORE(g+63*SIMDD, cpz1 * MM_LOAD(g+59*SIMDD) + b01 *(MM_LOAD(g+51*SIMDD) + MM_LOAD(g+57*SIMDD)));
 }
 
 static inline void _g0_2d4d_1200(double *g, Rys2eT *bc, CINTEnvVars *envs)
@@ -2008,20 +2108,21 @@ static inline void _g0_2d4d_1200(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cz = bc->c00z;
         double *b  = bc->b10;
         double *r  = envs->rirj;
-        ALIGNMM double rc[6*SIMDD];
-        _make_rc(rc, cx, cy, cz, r);
-        __MD r0 = MM_LOAD(rc+0*SIMDD);
-        __MD r1 = MM_LOAD(rc+1*SIMDD);
-        __MD r2 = MM_LOAD(rc+2*SIMDD);
-        __MD r3 = MM_LOAD(rc+3*SIMDD);
-        __MD r4 = MM_LOAD(rc+4*SIMDD);
-        __MD r5 = MM_LOAD(rc+5*SIMDD);
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
         __MD cx0 = MM_LOAD(cx+0*SIMDD);
         __MD cx1 = MM_LOAD(cx+1*SIMDD);
         __MD cy0 = MM_LOAD(cy+0*SIMDD);
         __MD cy1 = MM_LOAD(cy+1*SIMDD);
         __MD cz0 = MM_LOAD(cz+0*SIMDD);
         __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD r0 = MM_ADD(rx, cx0);
+        __MD r1 = MM_ADD(rx, cx1);
+        __MD r2 = MM_ADD(ry, cy0);
+        __MD r3 = MM_ADD(ry, cy1);
+        __MD r4 = MM_ADD(rz, cz0);
+        __MD r5 = MM_ADD(rz, cz1);
         __MD b0 = MM_LOAD(b+0*SIMDD);
         __MD b1 = MM_LOAD(b+1*SIMDD);
         __MD i2 = MM_SET1(2.);
@@ -2098,7 +2199,7 @@ static inline void _g0_2d4d_2001(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b0 = bc->b00;
-        double *b1 = bc->b01;
+        double *b1 = bc->b10;
         __MD r0 = MM_LOAD(c0x+0*SIMDD);
         __MD r1 = MM_LOAD(c0x+1*SIMDD);
         __MD r2 = MM_LOAD(c0y+0*SIMDD);
@@ -2158,7 +2259,7 @@ static inline void _g0_2d4d_2010(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b0 = bc->b00;
-        double *b1 = bc->b01;
+        double *b1 = bc->b10;
         __MD r0 = MM_LOAD(c0x+0*SIMDD);
         __MD r1 = MM_LOAD(c0x+1*SIMDD);
         __MD r2 = MM_LOAD(c0y+0*SIMDD);
@@ -2215,26 +2316,27 @@ static inline void _g0_2d4d_2100(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *cy = bc->c00y;
         double *cz = bc->c00z;
         double *b1 = bc->b10;
-        double *rij = envs->rirj;
-        ALIGNMM double rc[6*SIMDD];
-        _make_rc(rc, cx, cy, cz, rij);
-        __MD b10 = MM_LOAD(b1+0*SIMDD);
-        __MD b11 = MM_LOAD(b1+1*SIMDD);
-        __MD g32 = MM_LOAD(g+32*SIMDD);
-        __MD g33 = MM_LOAD(g+33*SIMDD);
-        __MD i2 = MM_SET1(2.);
-        __MD r0 = MM_LOAD(rc+0*SIMDD);
-        __MD r1 = MM_LOAD(rc+1*SIMDD);
-        __MD r2 = MM_LOAD(rc+2*SIMDD);
-        __MD r3 = MM_LOAD(rc+3*SIMDD);
-        __MD r4 = MM_LOAD(rc+4*SIMDD);
-        __MD r5 = MM_LOAD(rc+5*SIMDD);
+        double *r   = envs->rirj;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
         __MD s0 = MM_LOAD(cx+0*SIMDD);
         __MD s1 = MM_LOAD(cx+1*SIMDD);
         __MD s2 = MM_LOAD(cy+0*SIMDD);
         __MD s3 = MM_LOAD(cy+1*SIMDD);
         __MD s4 = MM_LOAD(cz+0*SIMDD);
         __MD s5 = MM_LOAD(cz+1*SIMDD);
+        __MD r0 = MM_ADD(rx, s0);
+        __MD r1 = MM_ADD(rx, s1);
+        __MD r2 = MM_ADD(ry, s2);
+        __MD r3 = MM_ADD(ry, s3);
+        __MD r4 = MM_ADD(rz, s4);
+        __MD r5 = MM_ADD(rz, s5);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD i2 = MM_SET1(2.);
         MM_STORE(g+2 *SIMDD, s0);
         MM_STORE(g+3 *SIMDD, s1);
         MM_STORE(g+8 *SIMDD, r0);
@@ -2353,94 +2455,2990 @@ void CINTg0_2e_2d4d_unrolled(double *g, Rys2eT *bc, CINTEnvVars *envs)
                (int)envs->ll_ceil, (int)envs->lj_ceil);
 }
 
-void CINTg0_2e_lj2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs)
+static inline void _srg0_2d4d_0000(double *g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        CINTg0_2e_2d(g, bc, envs);
-        CINTg0_lj_4d(g, envs);
 }
-void CINTg0_2e_kj2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs)
+
+static inline void _srg0_2d4d_0001(double *g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        CINTg0_2e_2d(g, bc, envs);
-        CINTg0_kj_4d(g, envs);
+        double *cx = bc->c0px;
+        double *cy = bc->c0py;
+        double *cz = bc->c0pz;
+        MM_STORE(g+2*SIMDD, MM_LOAD(cx+0*SIMDD));
+        MM_STORE(g+3*SIMDD, MM_LOAD(cx+1*SIMDD));
+        MM_STORE(g+6*SIMDD, MM_LOAD(cy+0*SIMDD));
+        MM_STORE(g+7*SIMDD, MM_LOAD(cy+1*SIMDD));
+        MM_STORE(g+10*SIMDD, MM_MUL(MM_LOAD(cz+0*SIMDD), MM_LOAD(g+8*SIMDD)));
+        MM_STORE(g+11*SIMDD, MM_MUL(MM_LOAD(cz+1*SIMDD), MM_LOAD(g+9*SIMDD)));
 }
-void CINTg0_2e_ik2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs)
+
+static inline void _srg0_2d4d_0002(double *g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        CINTg0_2e_2d(g, bc, envs);
-        CINTg0_ik_4d(g, envs);
+        double *cx = bc->c0px;
+        double *cy = bc->c0py;
+        double *cz = bc->c0pz;
+        double *b =  bc->b01;
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cx2 = MM_LOAD(cx+2*SIMDD);
+        __MD cx3 = MM_LOAD(cx+3*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cy2 = MM_LOAD(cy+2*SIMDD);
+        __MD cy3 = MM_LOAD(cy+3*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD cz2 = MM_LOAD(cz+2*SIMDD);
+        __MD cz3 = MM_LOAD(cz+3*SIMDD);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD g24 = MM_LOAD(g+24*SIMDD);
+        __MD g25 = MM_LOAD(g+25*SIMDD);
+        __MD g26 = MM_LOAD(g+26*SIMDD);
+        __MD g27 = MM_LOAD(g+27*SIMDD);
+        MM_STORE(g+4 *SIMDD, cx0);
+        MM_STORE(g+5 *SIMDD, cx1);
+        MM_STORE(g+6 *SIMDD, cx2);
+        MM_STORE(g+7 *SIMDD, cx3);
+        MM_STORE(g+16*SIMDD, cy0);
+        MM_STORE(g+17*SIMDD, cy1);
+        MM_STORE(g+18*SIMDD, cy2);
+        MM_STORE(g+19*SIMDD, cy3);
+        MM_STORE(g+28*SIMDD, MM_MUL(cz0, g24));
+        MM_STORE(g+29*SIMDD, MM_MUL(cz1, g25));
+        MM_STORE(g+30*SIMDD, MM_MUL(cz2, g26));
+        MM_STORE(g+31*SIMDD, MM_MUL(cz3, g27));
+        MM_STORE(g+8 *SIMDD, MM_FMA(cx0, cx0, b0));
+        MM_STORE(g+9 *SIMDD, MM_FMA(cx1, cx1, b1));
+        MM_STORE(g+10*SIMDD, MM_FMA(cx2, cx2, b2));
+        MM_STORE(g+11*SIMDD, MM_FMA(cx3, cx3, b3));
+        MM_STORE(g+20*SIMDD, MM_FMA(cy0, cy0, b0));
+        MM_STORE(g+21*SIMDD, MM_FMA(cy1, cy1, b1));
+        MM_STORE(g+22*SIMDD, MM_FMA(cy2, cy2, b2));
+        MM_STORE(g+23*SIMDD, MM_FMA(cy3, cy3, b3));
+        MM_STORE(g+32*SIMDD, MM_MUL(MM_FMA(cz0, cz0, b0), g24));
+        MM_STORE(g+33*SIMDD, MM_MUL(MM_FMA(cz1, cz1, b1), g25));
+        MM_STORE(g+34*SIMDD, MM_MUL(MM_FMA(cz2, cz2, b2), g26));
+        MM_STORE(g+35*SIMDD, MM_MUL(MM_FMA(cz3, cz3, b3), g27));
 }
-void CINTg0_2e_il2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs)
+
+static inline void _srg0_2d4d_0003(double *g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        CINTg0_2e_2d(g, bc, envs);
-        CINTg0_il_4d(g, envs);
+        double *cx = bc->c0px;
+        double *cy = bc->c0py;
+        double *cz = bc->c0pz;
+        double *b =  bc->b01;
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cx2 = MM_LOAD(cx+2*SIMDD);
+        __MD cx3 = MM_LOAD(cx+3*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cy2 = MM_LOAD(cy+2*SIMDD);
+        __MD cy3 = MM_LOAD(cy+3*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD cz2 = MM_LOAD(cz+2*SIMDD);
+        __MD cz3 = MM_LOAD(cz+3*SIMDD);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD g34 = MM_LOAD(g+34*SIMDD);
+        __MD g35 = MM_LOAD(g+35*SIMDD);
+        __MD i3 = MM_SET1(3.);
+        MM_STORE(g+4 *SIMDD, cx0);
+        MM_STORE(g+5 *SIMDD, cx1);
+        MM_STORE(g+6 *SIMDD, cx2);
+        MM_STORE(g+7 *SIMDD, cx3);
+        MM_STORE(g+20*SIMDD, cy0);
+        MM_STORE(g+21*SIMDD, cy1);
+        MM_STORE(g+22*SIMDD, cy2);
+        MM_STORE(g+23*SIMDD, cy3);
+        MM_STORE(g+36*SIMDD, cz0 * g32);
+        MM_STORE(g+37*SIMDD, cz1 * g33);
+        MM_STORE(g+38*SIMDD, cz2 * g34);
+        MM_STORE(g+39*SIMDD, cz3 * g35);
+        MM_STORE(g+8 *SIMDD, cx0 * cx0 + b0);
+        MM_STORE(g+9 *SIMDD, cx1 * cx1 + b1);
+        MM_STORE(g+10*SIMDD, cx2 * cx2 + b2);
+        MM_STORE(g+11*SIMDD, cx3 * cx3 + b3);
+        MM_STORE(g+24*SIMDD, cy0 * cy0 + b0);
+        MM_STORE(g+25*SIMDD, cy1 * cy1 + b1);
+        MM_STORE(g+26*SIMDD, cy2 * cy2 + b2);
+        MM_STORE(g+27*SIMDD, cy3 * cy3 + b3);
+        MM_STORE(g+40*SIMDD,(cz0 * cz0 + b0)* g32);
+        MM_STORE(g+41*SIMDD,(cz1 * cz1 + b1)* g33);
+        MM_STORE(g+42*SIMDD,(cz2 * cz2 + b2)* g34);
+        MM_STORE(g+43*SIMDD,(cz3 * cz3 + b3)* g35);
+        MM_STORE(g+12*SIMDD, cx0 *(cx0 * cx0 + i3 * b0));
+        MM_STORE(g+13*SIMDD, cx1 *(cx1 * cx1 + i3 * b1));
+        MM_STORE(g+14*SIMDD, cx2 *(cx2 * cx2 + i3 * b2));
+        MM_STORE(g+15*SIMDD, cx3 *(cx3 * cx3 + i3 * b3));
+        MM_STORE(g+28*SIMDD, cy0 *(cy0 * cy0 + i3 * b0));
+        MM_STORE(g+29*SIMDD, cy1 *(cy1 * cy1 + i3 * b1));
+        MM_STORE(g+30*SIMDD, cy2 *(cy2 * cy2 + i3 * b2));
+        MM_STORE(g+31*SIMDD, cy3 *(cy3 * cy3 + i3 * b3));
+        MM_STORE(g+44*SIMDD,(cz0 * cz0 + i3 * b0)* MM_LOAD(g+36*SIMDD));
+        MM_STORE(g+45*SIMDD,(cz1 * cz1 + i3 * b1)* MM_LOAD(g+37*SIMDD));
+        MM_STORE(g+46*SIMDD,(cz2 * cz2 + i3 * b2)* MM_LOAD(g+38*SIMDD));
+        MM_STORE(g+47*SIMDD,(cz3 * cz3 + i3 * b3)* MM_LOAD(g+39*SIMDD));
 }
 
-/*
- * g[i,k,l,j] = < ik | lj > = ( i j | k l )
- */
-int CINTg0_2e(double *g, double *cutoff, Rys2eT *bc, CINTEnvVars *envs, int count)
+static inline void _srg0_2d4d_0010(double *g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        ALIGNMM double aij[SIMDD];
-        ALIGNMM double akl[SIMDD];
-        ALIGNMM double a0[SIMDD];
-        ALIGNMM double a1[SIMDD];
-        ALIGNMM double aijkl[SIMDD];
-        ALIGNMM double fac1[SIMDD];
-        ALIGNMM double x[SIMDD];
-        ALIGNMM double rijrkl[SIMDD*3];
-        ALIGNMM double rijrx[SIMDD*3];
-        ALIGNMM double rklrx[SIMDD*3];
-        double *rij = envs->rij;
-        double *rkl = envs->rkl;
-        double *u = bc->u;
-        double *w = bc->w;
-        __MD ra, r0, r1, r2, r3, r4, r5, r6, r7, r8;
-        int nroots = envs->nrys_roots;
-        int i;
+        double *cx = bc->c0px;
+        double *cy = bc->c0py;
+        double *cz = bc->c0pz;
+        MM_STORE(g+2*SIMDD, MM_LOAD(cx+0*SIMDD));
+        MM_STORE(g+3*SIMDD, MM_LOAD(cx+1*SIMDD));
+        MM_STORE(g+6*SIMDD, MM_LOAD(cy+0*SIMDD));
+        MM_STORE(g+7*SIMDD, MM_LOAD(cy+1*SIMDD));
+        MM_STORE(g+10*SIMDD, MM_MUL(MM_LOAD(cz+0*SIMDD), MM_LOAD(g+8*SIMDD)));
+        MM_STORE(g+11*SIMDD, MM_MUL(MM_LOAD(cz+1*SIMDD), MM_LOAD(g+9*SIMDD)));
+}
 
-        //:for (int k = 0; k < count; k++) {
-        //:        aij[k] = envs->ai[k] + envs->aj[k];
-        //:        akl[k] = envs->ak[k] + envs->al[k];
-        //:        aijkl[k] = aij[k] + akl[k];
-        //:        a1[k] = aij[k] * akl[k];
-        //:        a0[k] = a1[k] / aijkl[k];
-        //:        //fac1[k] = sqrt(a0[k] / (a1[k] * a1[k] * a1[k])) * envs->fac[k];
-        //:        fac1[k] = envs->fac[k] / (sqrt(aijakl[k]) * a1[k]);
-        //:}
-        r2 = MM_ADD(MM_LOAD(envs->ai),  MM_LOAD(envs->aj));
-        r3 = MM_ADD(MM_LOAD(envs->ak),  MM_LOAD(envs->al));
-        MM_STORE(aij, r2);
-        MM_STORE(akl, r3);
-        r1 = MM_MUL(r2, r3);
-        MM_STORE(a1, r1);
-        ra = MM_ADD(r2, r3);
-        MM_STORE(aijkl, ra);
-        r0 = MM_DIV(r1, ra);
-        MM_STORE(a0, r0);
+static inline void _srg0_2d4d_0011(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c0px;
+        double *cy = bc->c0py;
+        double *cz = bc->c0pz;
+        double *b  = bc->b01;
+        double *r  = envs->rkrl;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cx2 = MM_LOAD(cx+2*SIMDD);
+        __MD cx3 = MM_LOAD(cx+3*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cy2 = MM_LOAD(cy+2*SIMDD);
+        __MD cy3 = MM_LOAD(cy+3*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD cz2 = MM_LOAD(cz+2*SIMDD);
+        __MD cz3 = MM_LOAD(cz+3*SIMDD);
+        __MD r0 = MM_ADD(rx, cx0);
+        __MD r1 = MM_ADD(rx, cx1);
+        __MD r2 = MM_ADD(rx, cx2);
+        __MD r3 = MM_ADD(rx, cx3);
+        __MD r4 = MM_ADD(ry, cy0);
+        __MD r5 = MM_ADD(ry, cy1);
+        __MD r6 = MM_ADD(ry, cy2);
+        __MD r7 = MM_ADD(ry, cy3);
+        __MD r8 = MM_ADD(rz, cz0);
+        __MD r9 = MM_ADD(rz, cz1);
+        __MD r10= MM_ADD(rz, cz2);
+        __MD r11= MM_ADD(rz, cz3);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+8 *SIMDD, cx0);
+        MM_STORE(g+9 *SIMDD, cx1);
+        MM_STORE(g+10*SIMDD, cx2);
+        MM_STORE(g+11*SIMDD, cx3);
+        MM_STORE(g+28*SIMDD, r4);
+        MM_STORE(g+29*SIMDD, r5);
+        MM_STORE(g+30*SIMDD, r6);
+        MM_STORE(g+31*SIMDD, r7);
+        MM_STORE(g+32*SIMDD, cy0);
+        MM_STORE(g+33*SIMDD, cy1);
+        MM_STORE(g+34*SIMDD, cy2);
+        MM_STORE(g+35*SIMDD, cy3);
+        MM_STORE(g+52*SIMDD, MM_MUL(r8, g48));
+        MM_STORE(g+53*SIMDD, MM_MUL(r9, g49));
+        MM_STORE(g+54*SIMDD, MM_MUL(r10,g50));
+        MM_STORE(g+55*SIMDD, MM_MUL(r11,g51));
+        MM_STORE(g+56*SIMDD, MM_MUL(cz0,g48));
+        MM_STORE(g+57*SIMDD, MM_MUL(cz1,g49));
+        MM_STORE(g+58*SIMDD, MM_MUL(cz2,g50));
+        MM_STORE(g+59*SIMDD, MM_MUL(cz3,g51));
+        MM_STORE(g+12*SIMDD, MM_FMA(r0, cx0, b0));
+        MM_STORE(g+13*SIMDD, MM_FMA(r1, cx1, b1));
+        MM_STORE(g+14*SIMDD, MM_FMA(r2, cx2, b2));
+        MM_STORE(g+15*SIMDD, MM_FMA(r3, cx3, b3));
+        MM_STORE(g+36*SIMDD, MM_FMA(r4, cy0, b0));
+        MM_STORE(g+37*SIMDD, MM_FMA(r5, cy1, b1));
+        MM_STORE(g+38*SIMDD, MM_FMA(r6, cy2, b2));
+        MM_STORE(g+39*SIMDD, MM_FMA(r7, cy3, b3));
+        MM_STORE(g+60*SIMDD, MM_MUL(MM_FMA(r8, cz0, b0), g48));
+        MM_STORE(g+61*SIMDD, MM_MUL(MM_FMA(r9, cz1, b1), g49));
+        MM_STORE(g+62*SIMDD, MM_MUL(MM_FMA(r10,cz2, b2), g50));
+        MM_STORE(g+63*SIMDD, MM_MUL(MM_FMA(r11,cz3, b3), g51));
+}
 
-#ifdef WITH_RANGE_COULOMB
-// Not recommended to mix range-separated Coulomb with regular Coulomb operator.
-// Keep this for backward compatibility to cint2
-        const double omega = envs->env[PTR_RANGE_OMEGA];
-        ALIGNMM double theta[SIMDD];
-        if (omega != 0) {
-                //:theta = omega * omega / (omega * omega + a0);
-                r0 = MM_SET1(omega);
-                r0 = MM_MUL(r0, r0);
-                r1 = MM_LOAD(a0);
-                r0 = MM_DIV(r0, MM_ADD(r0, r1));
-                MM_STORE(theta, r0);
-                if (omega > 0) { // long-range part of range-separated Coulomb operator
-                        //:a0 *= theta;
-                        MM_STORE(a0, MM_MUL(r0, r1));
-                }
+static inline void _srg0_2d4d_0012(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c0px;
+        double *cy = bc->c0py;
+        double *cz = bc->c0pz;
+        double *b =  bc->b01;
+        double *r  = envs->rkrl;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cx2 = MM_LOAD(cx+2*SIMDD);
+        __MD cx3 = MM_LOAD(cx+3*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cy2 = MM_LOAD(cy+2*SIMDD);
+        __MD cy3 = MM_LOAD(cy+3*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD cz2 = MM_LOAD(cz+2*SIMDD);
+        __MD cz3 = MM_LOAD(cz+3*SIMDD);
+        __MD r0 = MM_ADD(rx, cx0);
+        __MD r1 = MM_ADD(rx, cx1);
+        __MD r2 = MM_ADD(rx, cx2);
+        __MD r3 = MM_ADD(rx, cx3);
+        __MD r4 = MM_ADD(ry, cy0);
+        __MD r5 = MM_ADD(ry, cy1);
+        __MD r6 = MM_ADD(ry, cy2);
+        __MD r7 = MM_ADD(ry, cy3);
+        __MD r8 = MM_ADD(rz, cz0);
+        __MD r9 = MM_ADD(rz, cz1);
+        __MD r10= MM_ADD(rz, cz2);
+        __MD r11= MM_ADD(rz, cz3);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD i2 = MM_SET1(2.);
+        __MD g64 = MM_LOAD(g+64*SIMDD);
+        __MD g65 = MM_LOAD(g+65*SIMDD);
+        __MD g66 = MM_LOAD(g+66*SIMDD);
+        __MD g67 = MM_LOAD(g+67*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0 );
+        MM_STORE(g+5 *SIMDD, r1 );
+        MM_STORE(g+6 *SIMDD, r2 );
+        MM_STORE(g+7 *SIMDD, r3 );
+        MM_STORE(g+8 *SIMDD, cx0);
+        MM_STORE(g+9 *SIMDD, cx1);
+        MM_STORE(g+10*SIMDD, cx2);
+        MM_STORE(g+11*SIMDD, cx3);
+        MM_STORE(g+36*SIMDD, r4 );
+        MM_STORE(g+37*SIMDD, r5 );
+        MM_STORE(g+38*SIMDD, r6 );
+        MM_STORE(g+39*SIMDD, r7 );
+        MM_STORE(g+40*SIMDD, cy0);
+        MM_STORE(g+41*SIMDD, cy1);
+        MM_STORE(g+42*SIMDD, cy2);
+        MM_STORE(g+43*SIMDD, cy3);
+        MM_STORE(g+68*SIMDD, MM_MUL(r8 , g64));
+        MM_STORE(g+69*SIMDD, MM_MUL(r9 , g65));
+        MM_STORE(g+70*SIMDD, MM_MUL(r10, g66));
+        MM_STORE(g+71*SIMDD, MM_MUL(r11, g67));
+        MM_STORE(g+72*SIMDD, MM_MUL(cz0, g64));
+        MM_STORE(g+73*SIMDD, MM_MUL(cz1, g65));
+        MM_STORE(g+74*SIMDD, MM_MUL(cz2, g66));
+        MM_STORE(g+75*SIMDD, MM_MUL(cz3, g67));
+        MM_STORE(g+12*SIMDD, MM_FMA(r0 , cx0, b0));
+        MM_STORE(g+13*SIMDD, MM_FMA(r1 , cx1, b1));
+        MM_STORE(g+14*SIMDD, MM_FMA(r2 , cx2, b2));
+        MM_STORE(g+15*SIMDD, MM_FMA(r3 , cx3, b3));
+        MM_STORE(g+16*SIMDD, MM_FMA(cx0, cx0, b0));
+        MM_STORE(g+17*SIMDD, MM_FMA(cx1, cx1, b1));
+        MM_STORE(g+18*SIMDD, MM_FMA(cx2, cx2, b2));
+        MM_STORE(g+19*SIMDD, MM_FMA(cx3, cx3, b3));
+        MM_STORE(g+44*SIMDD, MM_FMA(r4 , cy0, b0));
+        MM_STORE(g+45*SIMDD, MM_FMA(r5 , cy1, b1));
+        MM_STORE(g+46*SIMDD, MM_FMA(r6 , cy2, b2));
+        MM_STORE(g+47*SIMDD, MM_FMA(r7 , cy3, b3));
+        MM_STORE(g+48*SIMDD, MM_FMA(cy0, cy0, b0));
+        MM_STORE(g+49*SIMDD, MM_FMA(cy1, cy1, b1));
+        MM_STORE(g+50*SIMDD, MM_FMA(cy2, cy2, b2));
+        MM_STORE(g+51*SIMDD, MM_FMA(cy3, cy3, b3));
+        MM_STORE(g+76*SIMDD, MM_MUL(MM_FMA(r8 , cz0, b0), g64));
+        MM_STORE(g+77*SIMDD, MM_MUL(MM_FMA(r9 , cz1, b1), g65));
+        MM_STORE(g+78*SIMDD, MM_MUL(MM_FMA(r10, cz2, b2), g66));
+        MM_STORE(g+79*SIMDD, MM_MUL(MM_FMA(r11, cz3, b3), g67));
+        MM_STORE(g+80*SIMDD, MM_MUL(MM_FMA(cz0, cz0, b0), g64));
+        MM_STORE(g+81*SIMDD, MM_MUL(MM_FMA(cz1, cz1, b1), g65));
+        MM_STORE(g+82*SIMDD, MM_MUL(MM_FMA(cz2, cz2, b2), g66));
+        MM_STORE(g+83*SIMDD, MM_MUL(MM_FMA(cz3, cz3, b3), g67));
+        MM_STORE(g+20*SIMDD, r0 * MM_LOAD(g+16*SIMDD) + i2 * b0 * cx0                );
+        MM_STORE(g+21*SIMDD, r1 * MM_LOAD(g+17*SIMDD) + i2 * b1 * cx1                );
+        MM_STORE(g+22*SIMDD, r2 * MM_LOAD(g+18*SIMDD) + i2 * b2 * cx2                );
+        MM_STORE(g+23*SIMDD, r3 * MM_LOAD(g+19*SIMDD) + i2 * b3 * cx3                );
+        MM_STORE(g+52*SIMDD, r4 * MM_LOAD(g+48*SIMDD) + i2 * b0 * cy0                );
+        MM_STORE(g+53*SIMDD, r5 * MM_LOAD(g+49*SIMDD) + i2 * b1 * cy1                );
+        MM_STORE(g+54*SIMDD, r6 * MM_LOAD(g+50*SIMDD) + i2 * b2 * cy2                );
+        MM_STORE(g+55*SIMDD, r7 * MM_LOAD(g+51*SIMDD) + i2 * b3 * cy3                );
+        MM_STORE(g+84*SIMDD, r8 * MM_LOAD(g+80*SIMDD) + i2 * b0 * MM_LOAD(g+72*SIMDD));
+        MM_STORE(g+85*SIMDD, r9 * MM_LOAD(g+81*SIMDD) + i2 * b1 * MM_LOAD(g+73*SIMDD));
+        MM_STORE(g+86*SIMDD, r10* MM_LOAD(g+82*SIMDD) + i2 * b2 * MM_LOAD(g+74*SIMDD));
+        MM_STORE(g+87*SIMDD, r11* MM_LOAD(g+83*SIMDD) + i2 * b3 * MM_LOAD(g+75*SIMDD));
+}
+
+static inline void _srg0_2d4d_0020(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c0px;
+        double *cy = bc->c0py;
+        double *cz = bc->c0pz;
+        double *b  = bc->b01;
+        __MD r0 = MM_LOAD(cx+0*SIMDD);
+        __MD r1 = MM_LOAD(cx+1*SIMDD);
+        __MD r2 = MM_LOAD(cx+2*SIMDD);
+        __MD r3 = MM_LOAD(cx+3*SIMDD);
+        __MD r4 = MM_LOAD(cy+0*SIMDD);
+        __MD r5 = MM_LOAD(cy+1*SIMDD);
+        __MD r6 = MM_LOAD(cy+2*SIMDD);
+        __MD r7 = MM_LOAD(cy+3*SIMDD);
+        __MD r8 = MM_LOAD(cz+0*SIMDD);
+        __MD r9 = MM_LOAD(cz+1*SIMDD);
+        __MD r10 = MM_LOAD(cz+2*SIMDD);
+        __MD r11 = MM_LOAD(cz+3*SIMDD);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD g24 = MM_LOAD(g+24*SIMDD);
+        __MD g25 = MM_LOAD(g+25*SIMDD);
+        __MD g26 = MM_LOAD(g+26*SIMDD);
+        __MD g27 = MM_LOAD(g+27*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+16*SIMDD, r4);
+        MM_STORE(g+17*SIMDD, r5);
+        MM_STORE(g+18*SIMDD, r6);
+        MM_STORE(g+19*SIMDD, r7);
+        MM_STORE(g+28*SIMDD, MM_MUL(r8 , g24));
+        MM_STORE(g+29*SIMDD, MM_MUL(r9 , g25));
+        MM_STORE(g+30*SIMDD, MM_MUL(r10, g26));
+        MM_STORE(g+31*SIMDD, MM_MUL(r11, g27));
+        MM_STORE(g+8 *SIMDD, MM_FMA(r0, r0, b0));
+        MM_STORE(g+9 *SIMDD, MM_FMA(r1, r1, b1));
+        MM_STORE(g+10*SIMDD, MM_FMA(r2, r2, b2));
+        MM_STORE(g+11*SIMDD, MM_FMA(r3, r3, b3));
+        MM_STORE(g+20*SIMDD, MM_FMA(r4, r4, b0));
+        MM_STORE(g+21*SIMDD, MM_FMA(r5, r5, b1));
+        MM_STORE(g+22*SIMDD, MM_FMA(r6, r6, b2));
+        MM_STORE(g+23*SIMDD, MM_FMA(r7, r7, b3));
+        MM_STORE(g+32*SIMDD, MM_MUL(MM_FMA(r8 , r8 , b0), g24));
+        MM_STORE(g+33*SIMDD, MM_MUL(MM_FMA(r9 , r9 , b1), g25));
+        MM_STORE(g+34*SIMDD, MM_MUL(MM_FMA(r10, r10, b2), g26));
+        MM_STORE(g+35*SIMDD, MM_MUL(MM_FMA(r11, r11, b3), g27));
+}
+
+static inline void _srg0_2d4d_0021(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c0px;
+        double *cy = bc->c0py;
+        double *cz = bc->c0pz;
+        double *b1 = bc->b01;
+        double *r = envs->rkrl;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD s0 = MM_LOAD(cx+0*SIMDD);
+        __MD s1 = MM_LOAD(cx+1*SIMDD);
+        __MD s2 = MM_LOAD(cx+2*SIMDD);
+        __MD s3 = MM_LOAD(cx+3*SIMDD);
+        __MD s4 = MM_LOAD(cy+0*SIMDD);
+        __MD s5 = MM_LOAD(cy+1*SIMDD);
+        __MD s6 = MM_LOAD(cy+2*SIMDD);
+        __MD s7 = MM_LOAD(cy+3*SIMDD);
+        __MD s8 = MM_LOAD(cz+0*SIMDD);
+        __MD s9 = MM_LOAD(cz+1*SIMDD);
+        __MD s10= MM_LOAD(cz+2*SIMDD);
+        __MD s11= MM_LOAD(cz+3*SIMDD);
+        __MD r0 = MM_ADD(rx, s0 );
+        __MD r1 = MM_ADD(rx, s1 );
+        __MD r2 = MM_ADD(rx, s2 );
+        __MD r3 = MM_ADD(rx, s3 );
+        __MD r4 = MM_ADD(ry, s4 );
+        __MD r5 = MM_ADD(ry, s5 );
+        __MD r6 = MM_ADD(ry, s6 );
+        __MD r7 = MM_ADD(ry, s7 );
+        __MD r8 = MM_ADD(rz, s8 );
+        __MD r9 = MM_ADD(rz, s9 );
+        __MD r10= MM_ADD(rz, s10);
+        __MD r11= MM_ADD(rz, s11);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g64 = MM_LOAD(g+64*SIMDD);
+        __MD g65 = MM_LOAD(g+65*SIMDD);
+        __MD g66 = MM_LOAD(g+66*SIMDD);
+        __MD g67 = MM_LOAD(g+67*SIMDD);
+        __MD i2 = MM_SET1(2.);
+        MM_STORE(g+4 *SIMDD, s0);
+        MM_STORE(g+5 *SIMDD, s1);
+        MM_STORE(g+6 *SIMDD, s2);
+        MM_STORE(g+7 *SIMDD, s3);
+        MM_STORE(g+16*SIMDD, r0);
+        MM_STORE(g+17*SIMDD, r1);
+        MM_STORE(g+18*SIMDD, r2);
+        MM_STORE(g+19*SIMDD, r3);
+        MM_STORE(g+36*SIMDD, s4);
+        MM_STORE(g+37*SIMDD, s5);
+        MM_STORE(g+38*SIMDD, s6);
+        MM_STORE(g+39*SIMDD, s7);
+        MM_STORE(g+48*SIMDD, r4);
+        MM_STORE(g+49*SIMDD, r5);
+        MM_STORE(g+50*SIMDD, r6);
+        MM_STORE(g+51*SIMDD, r7);
+        MM_STORE(g+68*SIMDD, s8 * g64);
+        MM_STORE(g+69*SIMDD, s9 * g65);
+        MM_STORE(g+70*SIMDD, s10* g66);
+        MM_STORE(g+71*SIMDD, s11* g67);
+        MM_STORE(g+80*SIMDD, r8 * g64);
+        MM_STORE(g+81*SIMDD, r9 * g65);
+        MM_STORE(g+82*SIMDD, r10* g66);
+        MM_STORE(g+83*SIMDD, r11* g67);
+        MM_STORE(g+8 *SIMDD, s0 * s0 + b10);
+        MM_STORE(g+9 *SIMDD, s1 * s1 + b11);
+        MM_STORE(g+10*SIMDD, s2 * s2 + b12);
+        MM_STORE(g+11*SIMDD, s3 * s3 + b13);
+        MM_STORE(g+20*SIMDD, s0 * r0 + b10);
+        MM_STORE(g+21*SIMDD, s1 * r1 + b11);
+        MM_STORE(g+22*SIMDD, s2 * r2 + b12);
+        MM_STORE(g+23*SIMDD, s3 * r3 + b13);
+        MM_STORE(g+40*SIMDD, s4 * s4 + b10);
+        MM_STORE(g+41*SIMDD, s5 * s5 + b11);
+        MM_STORE(g+42*SIMDD, s6 * s6 + b12);
+        MM_STORE(g+43*SIMDD, s7 * s7 + b13);
+        MM_STORE(g+52*SIMDD, s4 * r4 + b10);
+        MM_STORE(g+53*SIMDD, s5 * r5 + b11);
+        MM_STORE(g+54*SIMDD, s6 * r6 + b12);
+        MM_STORE(g+55*SIMDD, s7 * r7 + b13);
+        MM_STORE(g+72*SIMDD,(s8 * s8 + b10) * g64);
+        MM_STORE(g+73*SIMDD,(s9 * s9 + b11) * g65);
+        MM_STORE(g+74*SIMDD,(s10* s10+ b12)* g66);
+        MM_STORE(g+75*SIMDD,(s11* s11+ b13)* g67);
+        MM_STORE(g+84*SIMDD,(s8 * r8 + b10) * g64);
+        MM_STORE(g+85*SIMDD,(s9 * r9 + b11) * g65);
+        MM_STORE(g+86*SIMDD,(s10* r10+ b12)* g66);
+        MM_STORE(g+87*SIMDD,(s11* r11+ b13)* g67);
+        MM_STORE(g+24*SIMDD, r0 * MM_LOAD(g+8 *SIMDD ) + i2 * b10 * s0);
+        MM_STORE(g+25*SIMDD, r1 * MM_LOAD(g+9 *SIMDD ) + i2 * b11 * s1);
+        MM_STORE(g+26*SIMDD, r2 * MM_LOAD(g+10*SIMDD ) + i2 * b12 * s2);
+        MM_STORE(g+27*SIMDD, r3 * MM_LOAD(g+11*SIMDD ) + i2 * b13 * s3);
+        MM_STORE(g+56*SIMDD, r4 * MM_LOAD(g+40*SIMDD ) + i2 * b10 * s4);
+        MM_STORE(g+57*SIMDD, r5 * MM_LOAD(g+41*SIMDD ) + i2 * b11 * s5);
+        MM_STORE(g+58*SIMDD, r6 * MM_LOAD(g+42*SIMDD ) + i2 * b12 * s6);
+        MM_STORE(g+59*SIMDD, r7 * MM_LOAD(g+43*SIMDD ) + i2 * b13 * s7);
+        MM_STORE(g+88*SIMDD, r8 * MM_LOAD(g+72*SIMDD ) + i2 * b10 * MM_LOAD(g+68*SIMDD ));
+        MM_STORE(g+89*SIMDD, r9 * MM_LOAD(g+73*SIMDD ) + i2 * b11 * MM_LOAD(g+69*SIMDD ));
+        MM_STORE(g+90*SIMDD, r10* MM_LOAD(g+74*SIMDD ) + i2 * b12 * MM_LOAD(g+70*SIMDD ));
+        MM_STORE(g+91*SIMDD, r11* MM_LOAD(g+75*SIMDD ) + i2 * b13 * MM_LOAD(g+71*SIMDD ));
+}
+
+static inline void _srg0_2d4d_0030(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c0px;
+        double *cy = bc->c0py;
+        double *cz = bc->c0pz;
+        double *b  = bc->b01;
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD i3 = MM_SET1(3.);
+        __MD r0 = MM_LOAD(cx+0*SIMDD);
+        __MD r1 = MM_LOAD(cx+1*SIMDD);
+        __MD r2 = MM_LOAD(cx+2*SIMDD);
+        __MD r3 = MM_LOAD(cx+3*SIMDD);
+        __MD r4 = MM_LOAD(cy+0*SIMDD);
+        __MD r5 = MM_LOAD(cy+1*SIMDD);
+        __MD r6 = MM_LOAD(cy+2*SIMDD);
+        __MD r7 = MM_LOAD(cy+3*SIMDD);
+        __MD r8 = MM_LOAD(cz+0*SIMDD);
+        __MD r9 = MM_LOAD(cz+1*SIMDD);
+        __MD r10 = MM_LOAD(cz+2*SIMDD);
+        __MD r11 = MM_LOAD(cz+3*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD g34 = MM_LOAD(g+34*SIMDD);
+        __MD g35 = MM_LOAD(g+35*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+20*SIMDD, r4);
+        MM_STORE(g+21*SIMDD, r5);
+        MM_STORE(g+22*SIMDD, r6);
+        MM_STORE(g+23*SIMDD, r7);
+        MM_STORE(g+36*SIMDD, r8 * g32);
+        MM_STORE(g+37*SIMDD, r9 * g33);
+        MM_STORE(g+38*SIMDD, r10* g34);
+        MM_STORE(g+39*SIMDD, r11* g35);
+        MM_STORE(g+8 *SIMDD, r0 * r0 + b0);
+        MM_STORE(g+9 *SIMDD, r1 * r1 + b1);
+        MM_STORE(g+10*SIMDD, r2 * r2 + b2);
+        MM_STORE(g+11*SIMDD, r3 * r3 + b3);
+        MM_STORE(g+24*SIMDD, r4 * r4 + b0);
+        MM_STORE(g+25*SIMDD, r5 * r5 + b1);
+        MM_STORE(g+26*SIMDD, r6 * r6 + b2);
+        MM_STORE(g+27*SIMDD, r7 * r7 + b3);
+        MM_STORE(g+40*SIMDD,(r8 * r8 + b0)* g32);
+        MM_STORE(g+41*SIMDD,(r9 * r9 + b1)* g33);
+        MM_STORE(g+42*SIMDD,(r10* r10+ b2)* g34);
+        MM_STORE(g+43*SIMDD,(r11* r11+ b3)* g35);
+        MM_STORE(g+12*SIMDD, r0 *(r0 * r0 + i3 * b0));
+        MM_STORE(g+13*SIMDD, r1 *(r1 * r1 + i3 * b1));
+        MM_STORE(g+14*SIMDD, r2 *(r2 * r2 + i3 * b2));
+        MM_STORE(g+15*SIMDD, r3 *(r3 * r3 + i3 * b3));
+        MM_STORE(g+28*SIMDD, r4 *(r4 * r4 + i3 * b0));
+        MM_STORE(g+29*SIMDD, r5 *(r5 * r5 + i3 * b1));
+        MM_STORE(g+30*SIMDD, r6 *(r6 * r6 + i3 * b2));
+        MM_STORE(g+31*SIMDD, r7 *(r7 * r7 + i3 * b3));
+        MM_STORE(g+44*SIMDD,(r8 * r8 + i3 * b0) * MM_LOAD(g+36*SIMDD));
+        MM_STORE(g+45*SIMDD,(r9 * r9 + i3 * b1) * MM_LOAD(g+37*SIMDD));
+        MM_STORE(g+46*SIMDD,(r10* r10+ i3 * b2) * MM_LOAD(g+38*SIMDD));
+        MM_STORE(g+47*SIMDD,(r11* r11+ i3 * b3) * MM_LOAD(g+39*SIMDD));
+}
+
+static inline void _srg0_2d4d_0100(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c00x;
+        double *cy = bc->c00y;
+        double *cz = bc->c00z;
+        MM_STORE(g+2*SIMDD, MM_LOAD(cx+0*SIMDD));
+        MM_STORE(g+3*SIMDD, MM_LOAD(cx+1*SIMDD));
+        MM_STORE(g+6*SIMDD, MM_LOAD(cy+0*SIMDD));
+        MM_STORE(g+7*SIMDD, MM_LOAD(cy+1*SIMDD));
+        MM_STORE(g+10*SIMDD, MM_MUL(MM_LOAD(cz+0*SIMDD), MM_LOAD(g+8*SIMDD)));
+        MM_STORE(g+11*SIMDD, MM_MUL(MM_LOAD(cz+1*SIMDD), MM_LOAD(g+9*SIMDD)));
+}
+
+static inline void _srg0_2d4d_0101(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b   = bc->b00;
+        __MD cx0 = MM_LOAD(c0x+0*SIMDD);
+        __MD cx1 = MM_LOAD(c0x+1*SIMDD);
+        __MD cx2 = MM_LOAD(c0x+2*SIMDD);
+        __MD cx3 = MM_LOAD(c0x+3*SIMDD);
+        __MD cy0 = MM_LOAD(c0y+0*SIMDD);
+        __MD cy1 = MM_LOAD(c0y+1*SIMDD);
+        __MD cy2 = MM_LOAD(c0y+2*SIMDD);
+        __MD cy3 = MM_LOAD(c0y+3*SIMDD);
+        __MD cz0 = MM_LOAD(c0z+0*SIMDD);
+        __MD cz1 = MM_LOAD(c0z+1*SIMDD);
+        __MD cz2 = MM_LOAD(c0z+2*SIMDD);
+        __MD cz3 = MM_LOAD(c0z+3*SIMDD);
+        __MD px0 = MM_LOAD(cpx+0*SIMDD);
+        __MD px1 = MM_LOAD(cpx+1*SIMDD);
+        __MD px2 = MM_LOAD(cpx+2*SIMDD);
+        __MD px3 = MM_LOAD(cpx+3*SIMDD);
+        __MD py0 = MM_LOAD(cpy+0*SIMDD);
+        __MD py1 = MM_LOAD(cpy+1*SIMDD);
+        __MD py2 = MM_LOAD(cpy+2*SIMDD);
+        __MD py3 = MM_LOAD(cpy+3*SIMDD);
+        __MD pz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD pz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD pz2 = MM_LOAD(cpz+2*SIMDD);
+        __MD pz3 = MM_LOAD(cpz+3*SIMDD);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD g34 = MM_LOAD(g+34*SIMDD);
+        __MD g35 = MM_LOAD(g+35*SIMDD);
+        MM_STORE(g+4 *SIMDD, px0);
+        MM_STORE(g+5 *SIMDD, px1);
+        MM_STORE(g+6 *SIMDD, px2);
+        MM_STORE(g+7 *SIMDD, px3);
+        MM_STORE(g+8 *SIMDD, cx0);
+        MM_STORE(g+9 *SIMDD, cx1);
+        MM_STORE(g+10*SIMDD, cx2);
+        MM_STORE(g+11*SIMDD, cx3);
+        MM_STORE(g+20*SIMDD, py0);
+        MM_STORE(g+21*SIMDD, py1);
+        MM_STORE(g+22*SIMDD, py2);
+        MM_STORE(g+23*SIMDD, py3);
+        MM_STORE(g+24*SIMDD, cy0);
+        MM_STORE(g+25*SIMDD, cy1);
+        MM_STORE(g+26*SIMDD, cy2);
+        MM_STORE(g+27*SIMDD, cy3);
+        MM_STORE(g+36*SIMDD, MM_MUL(pz0, g32));
+        MM_STORE(g+37*SIMDD, MM_MUL(pz1, g33));
+        MM_STORE(g+38*SIMDD, MM_MUL(pz2, g34));
+        MM_STORE(g+39*SIMDD, MM_MUL(pz3, g35));
+        MM_STORE(g+40*SIMDD, MM_MUL(cz0, g32));
+        MM_STORE(g+41*SIMDD, MM_MUL(cz1, g33));
+        MM_STORE(g+42*SIMDD, MM_MUL(cz2, g34));
+        MM_STORE(g+43*SIMDD, MM_MUL(cz3, g35));
+        MM_STORE(g+12*SIMDD, MM_FMA(px0, cx0, b0));
+        MM_STORE(g+13*SIMDD, MM_FMA(px1, cx1, b1));
+        MM_STORE(g+14*SIMDD, MM_FMA(px2, cx2, b2));
+        MM_STORE(g+15*SIMDD, MM_FMA(px3, cx3, b3));
+        MM_STORE(g+28*SIMDD, MM_FMA(py0, cy0, b0));
+        MM_STORE(g+29*SIMDD, MM_FMA(py1, cy1, b1));
+        MM_STORE(g+30*SIMDD, MM_FMA(py2, cy2, b2));
+        MM_STORE(g+31*SIMDD, MM_FMA(py3, cy3, b3));
+        MM_STORE(g+44*SIMDD, MM_MUL(MM_FMA(pz0, cz0, b0), g32));
+        MM_STORE(g+45*SIMDD, MM_MUL(MM_FMA(pz1, cz1, b1), g33));
+        MM_STORE(g+46*SIMDD, MM_MUL(MM_FMA(pz2, cz2, b2), g34));
+        MM_STORE(g+47*SIMDD, MM_MUL(MM_FMA(pz3, cz3, b3), g35));
+}
+
+static inline void _srg0_2d4d_0102(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0 = bc->b00;
+        double *b1 = bc->b01;
+        __MD r0 = MM_LOAD(c0x+0*SIMDD);
+        __MD r1 = MM_LOAD(c0x+1*SIMDD);
+        __MD r2 = MM_LOAD(c0x+2*SIMDD);
+        __MD r3 = MM_LOAD(c0x+3*SIMDD);
+        __MD r4 = MM_LOAD(c0y+0*SIMDD);
+        __MD r5 = MM_LOAD(c0y+1*SIMDD);
+        __MD r6 = MM_LOAD(c0y+2*SIMDD);
+        __MD r7 = MM_LOAD(c0y+3*SIMDD);
+        __MD r8 = MM_LOAD(c0z+0*SIMDD);
+        __MD r9 = MM_LOAD(c0z+1*SIMDD);
+        __MD r10= MM_LOAD(c0z+2*SIMDD);
+        __MD r11= MM_LOAD(c0z+3*SIMDD);
+        __MD s0 = MM_LOAD(cpx+0*SIMDD);
+        __MD s1 = MM_LOAD(cpx+1*SIMDD);
+        __MD s2 = MM_LOAD(cpx+2*SIMDD);
+        __MD s3 = MM_LOAD(cpx+3*SIMDD);
+        __MD s4 = MM_LOAD(cpy+0*SIMDD);
+        __MD s5 = MM_LOAD(cpy+1*SIMDD);
+        __MD s6 = MM_LOAD(cpy+2*SIMDD);
+        __MD s7 = MM_LOAD(cpy+3*SIMDD);
+        __MD s8 = MM_LOAD(cpz+0*SIMDD);
+        __MD s9 = MM_LOAD(cpz+1*SIMDD);
+        __MD s10= MM_LOAD(cpz+2*SIMDD);
+        __MD s11= MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        MM_STORE(g+4 *SIMDD, s0);
+        MM_STORE(g+5 *SIMDD, s1);
+        MM_STORE(g+6 *SIMDD, s2);
+        MM_STORE(g+7 *SIMDD, s3);
+        MM_STORE(g+12*SIMDD, r0);
+        MM_STORE(g+13*SIMDD, r1);
+        MM_STORE(g+14*SIMDD, r2);
+        MM_STORE(g+15*SIMDD, r3);
+        MM_STORE(g+28*SIMDD, s4);
+        MM_STORE(g+29*SIMDD, s5);
+        MM_STORE(g+30*SIMDD, s6);
+        MM_STORE(g+31*SIMDD, s7);
+        MM_STORE(g+36*SIMDD, r4);
+        MM_STORE(g+37*SIMDD, r5);
+        MM_STORE(g+38*SIMDD, r6);
+        MM_STORE(g+39*SIMDD, r7);
+        MM_STORE(g+52*SIMDD, MM_MUL(s8 , g48));
+        MM_STORE(g+53*SIMDD, MM_MUL(s9 , g49));
+        MM_STORE(g+54*SIMDD, MM_MUL(s10, g50));
+        MM_STORE(g+55*SIMDD, MM_MUL(s11, g51));
+        MM_STORE(g+60*SIMDD, MM_MUL(r8 , g48));
+        MM_STORE(g+61*SIMDD, MM_MUL(r9 , g49));
+        MM_STORE(g+62*SIMDD, MM_MUL(r10, g50));
+        MM_STORE(g+63*SIMDD, MM_MUL(r11, g51));
+        MM_STORE(g+8 *SIMDD, MM_FMA(s0, s0, b10));
+        MM_STORE(g+9 *SIMDD, MM_FMA(s1, s1, b11));
+        MM_STORE(g+10*SIMDD, MM_FMA(s2, s2, b12));
+        MM_STORE(g+11*SIMDD, MM_FMA(s3, s3, b13));
+        MM_STORE(g+16*SIMDD, MM_FMA(r0, s0, b00));
+        MM_STORE(g+17*SIMDD, MM_FMA(r1, s1, b01));
+        MM_STORE(g+18*SIMDD, MM_FMA(r2, s2, b02));
+        MM_STORE(g+19*SIMDD, MM_FMA(r3, s3, b03));
+        MM_STORE(g+32*SIMDD, MM_FMA(s4, s4, b10));
+        MM_STORE(g+33*SIMDD, MM_FMA(s5, s5, b11));
+        MM_STORE(g+34*SIMDD, MM_FMA(s6, s6, b12));
+        MM_STORE(g+35*SIMDD, MM_FMA(s7, s7, b13));
+        MM_STORE(g+40*SIMDD, MM_FMA(r4, s4, b00));
+        MM_STORE(g+41*SIMDD, MM_FMA(r5, s5, b01));
+        MM_STORE(g+42*SIMDD, MM_FMA(r6, s6, b02));
+        MM_STORE(g+43*SIMDD, MM_FMA(r7, s7, b03));
+        MM_STORE(g+56*SIMDD, MM_MUL(MM_FMA(s8 , s8 , b10), g48));
+        MM_STORE(g+57*SIMDD, MM_MUL(MM_FMA(s9 , s9 , b11), g49));
+        MM_STORE(g+58*SIMDD, MM_MUL(MM_FMA(s10, s10, b12), g50));
+        MM_STORE(g+59*SIMDD, MM_MUL(MM_FMA(s11, s11, b13), g51));
+        MM_STORE(g+64*SIMDD, MM_MUL(MM_FMA(r8 , s8 , b00), g48));
+        MM_STORE(g+65*SIMDD, MM_MUL(MM_FMA(r9 , s9 , b01), g49));
+        MM_STORE(g+66*SIMDD, MM_MUL(MM_FMA(r10, s10, b02), g50));
+        MM_STORE(g+67*SIMDD, MM_MUL(MM_FMA(r11, s11, b03), g51));
+        MM_STORE(g+20*SIMDD, s0 *(MM_LOAD(g+16*SIMDD) + b00) + b10 * r0);
+        MM_STORE(g+21*SIMDD, s1 *(MM_LOAD(g+17*SIMDD) + b01) + b11 * r1);
+        MM_STORE(g+22*SIMDD, s2 *(MM_LOAD(g+18*SIMDD) + b02) + b12 * r2);
+        MM_STORE(g+23*SIMDD, s3 *(MM_LOAD(g+19*SIMDD) + b03) + b13 * r3);
+        MM_STORE(g+44*SIMDD, s4 *(MM_LOAD(g+40*SIMDD) + b00) + b10 * r4);
+        MM_STORE(g+45*SIMDD, s5 *(MM_LOAD(g+41*SIMDD) + b01) + b11 * r5);
+        MM_STORE(g+46*SIMDD, s6 *(MM_LOAD(g+42*SIMDD) + b02) + b12 * r6);
+        MM_STORE(g+47*SIMDD, s7 *(MM_LOAD(g+43*SIMDD) + b03) + b13 * r7);
+        MM_STORE(g+68*SIMDD, s8 * MM_LOAD(g+64*SIMDD) + b00 * MM_LOAD(g+52*SIMDD) + b10 * MM_LOAD(g+60*SIMDD));
+        MM_STORE(g+69*SIMDD, s9 * MM_LOAD(g+65*SIMDD) + b01 * MM_LOAD(g+53*SIMDD) + b11 * MM_LOAD(g+61*SIMDD));
+        MM_STORE(g+70*SIMDD, s10* MM_LOAD(g+66*SIMDD) + b02 * MM_LOAD(g+54*SIMDD) + b12 * MM_LOAD(g+62*SIMDD));
+        MM_STORE(g+71*SIMDD, s11* MM_LOAD(g+67*SIMDD) + b03 * MM_LOAD(g+55*SIMDD) + b13 * MM_LOAD(g+63*SIMDD));
+}
+
+static inline void _srg0_2d4d_0110(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b   = bc->b00;
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD r0 = MM_LOAD(c0x+0*SIMDD);
+        __MD r1 = MM_LOAD(c0x+1*SIMDD);
+        __MD r2 = MM_LOAD(c0x+2*SIMDD);
+        __MD r3 = MM_LOAD(c0x+3*SIMDD);
+        __MD r4 = MM_LOAD(c0y+0*SIMDD);
+        __MD r5 = MM_LOAD(c0y+1*SIMDD);
+        __MD r6 = MM_LOAD(c0y+2*SIMDD);
+        __MD r7 = MM_LOAD(c0y+3*SIMDD);
+        __MD r8 = MM_LOAD(c0z+0*SIMDD);
+        __MD r9 = MM_LOAD(c0z+1*SIMDD);
+        __MD r10= MM_LOAD(c0z+2*SIMDD);
+        __MD r11= MM_LOAD(c0z+3*SIMDD);
+        __MD s0 = MM_LOAD(cpx+0*SIMDD);
+        __MD s1 = MM_LOAD(cpx+1*SIMDD);
+        __MD s2 = MM_LOAD(cpx+2*SIMDD);
+        __MD s3 = MM_LOAD(cpx+3*SIMDD);
+        __MD s4 = MM_LOAD(cpy+0*SIMDD);
+        __MD s5 = MM_LOAD(cpy+1*SIMDD);
+        __MD s6 = MM_LOAD(cpy+2*SIMDD);
+        __MD s7 = MM_LOAD(cpy+3*SIMDD);
+        __MD s8 = MM_LOAD(cpz+0*SIMDD);
+        __MD s9 = MM_LOAD(cpz+1*SIMDD);
+        __MD s10= MM_LOAD(cpz+2*SIMDD);
+        __MD s11= MM_LOAD(cpz+3*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD g34 = MM_LOAD(g+34*SIMDD);
+        __MD g35 = MM_LOAD(g+35*SIMDD);
+        MM_STORE(g+4 *SIMDD, s0);
+        MM_STORE(g+5 *SIMDD, s1);
+        MM_STORE(g+6 *SIMDD, s2);
+        MM_STORE(g+7 *SIMDD, s3);
+        MM_STORE(g+20*SIMDD, s4);
+        MM_STORE(g+21*SIMDD, s5);
+        MM_STORE(g+22*SIMDD, s6);
+        MM_STORE(g+23*SIMDD, s7);
+        MM_STORE(g+36*SIMDD, s8 * g32);
+        MM_STORE(g+37*SIMDD, s9 * g33);
+        MM_STORE(g+38*SIMDD, s10* g34);
+        MM_STORE(g+39*SIMDD, s11* g35);
+        MM_STORE(g+8 *SIMDD, r0);
+        MM_STORE(g+9 *SIMDD, r1);
+        MM_STORE(g+10*SIMDD, r2);
+        MM_STORE(g+11*SIMDD, r3);
+        MM_STORE(g+24*SIMDD, r4);
+        MM_STORE(g+25*SIMDD, r5);
+        MM_STORE(g+26*SIMDD, r6);
+        MM_STORE(g+27*SIMDD, r7);
+        MM_STORE(g+40*SIMDD, r8 * g32);
+        MM_STORE(g+41*SIMDD, r9 * g33);
+        MM_STORE(g+42*SIMDD, r10* g34);
+        MM_STORE(g+43*SIMDD, r11* g35);
+        MM_STORE(g+12*SIMDD, s0 * r0 + b0);
+        MM_STORE(g+13*SIMDD, s1 * r1 + b1);
+        MM_STORE(g+14*SIMDD, s2 * r2 + b2);
+        MM_STORE(g+15*SIMDD, s3 * r3 + b3);
+        MM_STORE(g+28*SIMDD, s4 * r4 + b0);
+        MM_STORE(g+29*SIMDD, s5 * r5 + b1);
+        MM_STORE(g+30*SIMDD, s6 * r6 + b2);
+        MM_STORE(g+31*SIMDD, s7 * r7 + b3);
+        MM_STORE(g+44*SIMDD,(s8 * r8 + b0) * g32);
+        MM_STORE(g+45*SIMDD,(s9 * r9 + b1) * g33);
+        MM_STORE(g+46*SIMDD,(s10* r10+ b2) * g34);
+        MM_STORE(g+47*SIMDD,(s11* r11+ b3) * g35);
+}
+
+static inline void _srg0_2d4d_0111(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0  = bc->b00;
+        double *b1  = bc->b01;
+        double *r   = envs->rkrl;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpx2 = MM_LOAD(cpx+2*SIMDD);
+        __MD cpx3 = MM_LOAD(cpx+3*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpy2 = MM_LOAD(cpy+2*SIMDD);
+        __MD cpy3 = MM_LOAD(cpy+3*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD cpz2 = MM_LOAD(cpz+2*SIMDD);
+        __MD cpz3 = MM_LOAD(cpz+3*SIMDD);
+        __MD rcp0 = MM_ADD(rx, cpx0);
+        __MD rcp1 = MM_ADD(rx, cpx1);
+        __MD rcp2 = MM_ADD(rx, cpx2);
+        __MD rcp3 = MM_ADD(rx, cpx3);
+        __MD rcp4 = MM_ADD(ry, cpy0);
+        __MD rcp5 = MM_ADD(ry, cpy1);
+        __MD rcp6 = MM_ADD(ry, cpy2);
+        __MD rcp7 = MM_ADD(ry, cpy3);
+        __MD rcp8 = MM_ADD(rz, cpz0);
+        __MD rcp9 = MM_ADD(rz, cpz1);
+        __MD rcp10= MM_ADD(rz, cpz2);
+        __MD rcp11= MM_ADD(rz, cpz3);
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0x2 = MM_LOAD(c0x+2*SIMDD);
+        __MD c0x3 = MM_LOAD(c0x+3*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0y2 = MM_LOAD(c0y+2*SIMDD);
+        __MD c0y3 = MM_LOAD(c0y+3*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
+        __MD c0z2 = MM_LOAD(c0z+2*SIMDD);
+        __MD c0z3 = MM_LOAD(c0z+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g96 = MM_LOAD(g+96*SIMDD);
+        __MD g97 = MM_LOAD(g+97*SIMDD);
+        __MD g98 = MM_LOAD(g+98*SIMDD);
+        __MD g99 = MM_LOAD(g+99*SIMDD);
+        MM_STORE(g+4 *SIMDD, rcp0);
+        MM_STORE(g+5 *SIMDD, rcp1);
+        MM_STORE(g+6 *SIMDD, rcp2);
+        MM_STORE(g+7 *SIMDD, rcp3);
+        MM_STORE(g+8 *SIMDD, cpx0);
+        MM_STORE(g+9 *SIMDD, cpx1);
+        MM_STORE(g+10*SIMDD, cpx2);
+        MM_STORE(g+11*SIMDD, cpx3);
+        MM_STORE(g+24*SIMDD, c0x0);
+        MM_STORE(g+25*SIMDD, c0x1);
+        MM_STORE(g+26*SIMDD, c0x2);
+        MM_STORE(g+27*SIMDD, c0x3);
+        MM_STORE(g+52*SIMDD, rcp4);
+        MM_STORE(g+53*SIMDD, rcp5);
+        MM_STORE(g+54*SIMDD, rcp6);
+        MM_STORE(g+55*SIMDD, rcp7);
+        MM_STORE(g+56*SIMDD, cpy0);
+        MM_STORE(g+57*SIMDD, cpy1);
+        MM_STORE(g+58*SIMDD, cpy2);
+        MM_STORE(g+59*SIMDD, cpy3);
+        MM_STORE(g+72*SIMDD, c0y0);
+        MM_STORE(g+73*SIMDD, c0y1);
+        MM_STORE(g+74*SIMDD, c0y2);
+        MM_STORE(g+75*SIMDD, c0y3);
+        MM_STORE(g+100*SIMDD, rcp8 * g96);
+        MM_STORE(g+101*SIMDD, rcp9 * g97);
+        MM_STORE(g+102*SIMDD, rcp10* g98);
+        MM_STORE(g+103*SIMDD, rcp11* g99);
+        MM_STORE(g+104*SIMDD, cpz0 * g96);
+        MM_STORE(g+105*SIMDD, cpz1 * g97);
+        MM_STORE(g+106*SIMDD, cpz2 * g98);
+        MM_STORE(g+107*SIMDD, cpz3 * g99);
+        MM_STORE(g+120*SIMDD, c0z0 * g96);
+        MM_STORE(g+121*SIMDD, c0z1 * g97);
+        MM_STORE(g+122*SIMDD, c0z2 * g98);
+        MM_STORE(g+123*SIMDD, c0z3 * g99);
+        MM_STORE(g+28*SIMDD, c0x0 * rcp0 + b00);
+        MM_STORE(g+29*SIMDD, c0x1 * rcp1 + b01);
+        MM_STORE(g+30*SIMDD, c0x2 * rcp2 + b02);
+        MM_STORE(g+31*SIMDD, c0x3 * rcp3 + b03);
+        MM_STORE(g+32*SIMDD, c0x0 * cpx0 + b00);
+        MM_STORE(g+33*SIMDD, c0x1 * cpx1 + b01);
+        MM_STORE(g+34*SIMDD, c0x2 * cpx2 + b02);
+        MM_STORE(g+35*SIMDD, c0x3 * cpx3 + b03);
+        MM_STORE(g+12*SIMDD, cpx0 * rcp0 + b10);
+        MM_STORE(g+13*SIMDD, cpx1 * rcp1 + b11);
+        MM_STORE(g+14*SIMDD, cpx2 * rcp2 + b12);
+        MM_STORE(g+15*SIMDD, cpx3 * rcp3 + b13);
+        MM_STORE(g+60*SIMDD, cpy0 * rcp4 + b10);
+        MM_STORE(g+61*SIMDD, cpy1 * rcp5 + b11);
+        MM_STORE(g+62*SIMDD, cpy2 * rcp6 + b12);
+        MM_STORE(g+63*SIMDD, cpy3 * rcp7 + b13);
+        MM_STORE(g+76*SIMDD, c0y0 * rcp4 + b00);
+        MM_STORE(g+77*SIMDD, c0y1 * rcp5 + b01);
+        MM_STORE(g+78*SIMDD, c0y2 * rcp6 + b02);
+        MM_STORE(g+79*SIMDD, c0y3 * rcp7 + b03);
+        MM_STORE(g+80*SIMDD, c0y0 * cpy0 + b00);
+        MM_STORE(g+81*SIMDD, c0y1 * cpy1 + b01);
+        MM_STORE(g+82*SIMDD, c0y2 * cpy2 + b02);
+        MM_STORE(g+83*SIMDD, c0y3 * cpy3 + b03);
+        MM_STORE(g+108*SIMDD,(cpz0 * rcp8 + b10) * g96);
+        MM_STORE(g+109*SIMDD,(cpz1 * rcp9 + b11) * g97);
+        MM_STORE(g+110*SIMDD,(cpz2 * rcp10+ b12) * g98);
+        MM_STORE(g+111*SIMDD,(cpz3 * rcp11+ b13) * g99);
+        MM_STORE(g+124*SIMDD,(c0z0 * rcp8 + b00) * g96);
+        MM_STORE(g+125*SIMDD,(c0z1 * rcp9 + b01) * g97);
+        MM_STORE(g+126*SIMDD,(c0z2 * rcp10+ b02) * g98);
+        MM_STORE(g+127*SIMDD,(c0z3 * rcp11+ b03) * g99);
+        MM_STORE(g+128*SIMDD,(c0z0 * cpz0 + b00) * g96);
+        MM_STORE(g+129*SIMDD,(c0z1 * cpz1 + b01) * g97);
+        MM_STORE(g+130*SIMDD,(c0z2 * cpz2 + b02) * g98);
+        MM_STORE(g+131*SIMDD,(c0z3 * cpz3 + b03) * g99);
+        MM_STORE(g+36*SIMDD , c0x0 * MM_LOAD(g+12 *SIMDD) + b00 * (rcp0 + cpx0));
+        MM_STORE(g+37*SIMDD , c0x1 * MM_LOAD(g+13 *SIMDD) + b01 * (rcp1 + cpx1));
+        MM_STORE(g+38*SIMDD , c0x2 * MM_LOAD(g+14 *SIMDD) + b02 * (rcp2 + cpx2));
+        MM_STORE(g+39*SIMDD , c0x3 * MM_LOAD(g+15 *SIMDD) + b03 * (rcp3 + cpx3));
+        MM_STORE(g+84*SIMDD , c0y0 * MM_LOAD(g+60 *SIMDD) + b00 * (rcp4 + cpy0));
+        MM_STORE(g+85*SIMDD , c0y1 * MM_LOAD(g+61 *SIMDD) + b01 * (rcp5 + cpy1));
+        MM_STORE(g+86*SIMDD , c0y2 * MM_LOAD(g+62 *SIMDD) + b02 * (rcp6 + cpy2));
+        MM_STORE(g+87*SIMDD , c0y3 * MM_LOAD(g+63 *SIMDD) + b03 * (rcp7 + cpy3));
+        MM_STORE(g+132*SIMDD, c0z0 * MM_LOAD(g+108*SIMDD) + b00 * (MM_LOAD(g+100*SIMDD) + MM_LOAD(g+104*SIMDD)));
+        MM_STORE(g+133*SIMDD, c0z1 * MM_LOAD(g+109*SIMDD) + b01 * (MM_LOAD(g+101*SIMDD) + MM_LOAD(g+105*SIMDD)));
+        MM_STORE(g+134*SIMDD, c0z2 * MM_LOAD(g+110*SIMDD) + b02 * (MM_LOAD(g+102*SIMDD) + MM_LOAD(g+106*SIMDD)));
+        MM_STORE(g+135*SIMDD, c0z3 * MM_LOAD(g+111*SIMDD) + b03 * (MM_LOAD(g+103*SIMDD) + MM_LOAD(g+107*SIMDD)));
+}
+
+static inline void _srg0_2d4d_0120(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0 = bc->b00;
+        double *b1 = bc->b01;
+        __MD r0 = MM_LOAD(c0x+0*SIMDD);
+        __MD r1 = MM_LOAD(c0x+1*SIMDD);
+        __MD r2 = MM_LOAD(c0x+2*SIMDD);
+        __MD r3 = MM_LOAD(c0x+3*SIMDD);
+        __MD r4 = MM_LOAD(c0y+0*SIMDD);
+        __MD r5 = MM_LOAD(c0y+1*SIMDD);
+        __MD r6 = MM_LOAD(c0y+2*SIMDD);
+        __MD r7 = MM_LOAD(c0y+3*SIMDD);
+        __MD r8 = MM_LOAD(c0z+0*SIMDD);
+        __MD r9 = MM_LOAD(c0z+1*SIMDD);
+        __MD r10= MM_LOAD(c0z+2*SIMDD);
+        __MD r11= MM_LOAD(c0z+3*SIMDD);
+        __MD s0 = MM_LOAD(cpx+0*SIMDD);
+        __MD s1 = MM_LOAD(cpx+1*SIMDD);
+        __MD s2 = MM_LOAD(cpx+2*SIMDD);
+        __MD s3 = MM_LOAD(cpx+3*SIMDD);
+        __MD s4 = MM_LOAD(cpy+0*SIMDD);
+        __MD s5 = MM_LOAD(cpy+1*SIMDD);
+        __MD s6 = MM_LOAD(cpy+2*SIMDD);
+        __MD s7 = MM_LOAD(cpy+3*SIMDD);
+        __MD s8 = MM_LOAD(cpz+0*SIMDD);
+        __MD s9 = MM_LOAD(cpz+1*SIMDD);
+        __MD s10= MM_LOAD(cpz+2*SIMDD);
+        __MD s11= MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        MM_STORE(g+4 *SIMDD, s0);
+        MM_STORE(g+5 *SIMDD, s1);
+        MM_STORE(g+6 *SIMDD, s2);
+        MM_STORE(g+7 *SIMDD, s3);
+        MM_STORE(g+12*SIMDD, r0);
+        MM_STORE(g+13*SIMDD, r1);
+        MM_STORE(g+14*SIMDD, r2);
+        MM_STORE(g+15*SIMDD, r3);
+        MM_STORE(g+28*SIMDD, s4);
+        MM_STORE(g+29*SIMDD, s5);
+        MM_STORE(g+30*SIMDD, s6);
+        MM_STORE(g+31*SIMDD, s7);
+        MM_STORE(g+36*SIMDD, r4);
+        MM_STORE(g+37*SIMDD, r5);
+        MM_STORE(g+38*SIMDD, r6);
+        MM_STORE(g+39*SIMDD, r7);
+        MM_STORE(g+52*SIMDD, MM_MUL(s8, g48));
+        MM_STORE(g+53*SIMDD, MM_MUL(s9, g49));
+        MM_STORE(g+54*SIMDD, MM_MUL(s10,g50));
+        MM_STORE(g+55*SIMDD, MM_MUL(s11,g51));
+        MM_STORE(g+60*SIMDD, MM_MUL(r8, g48));
+        MM_STORE(g+61*SIMDD, MM_MUL(r9, g49));
+        MM_STORE(g+62*SIMDD, MM_MUL(r10,g50));
+        MM_STORE(g+63*SIMDD, MM_MUL(r11,g51));
+        MM_STORE(g+8 *SIMDD, MM_FMA(s0, s0, b10));
+        MM_STORE(g+9 *SIMDD, MM_FMA(s1, s1, b11));
+        MM_STORE(g+10*SIMDD, MM_FMA(s2, s2, b12));
+        MM_STORE(g+11*SIMDD, MM_FMA(s3, s3, b13));
+        MM_STORE(g+16*SIMDD, MM_FMA(r0, s0, b00));
+        MM_STORE(g+17*SIMDD, MM_FMA(r1, s1, b01));
+        MM_STORE(g+18*SIMDD, MM_FMA(r2, s2, b02));
+        MM_STORE(g+19*SIMDD, MM_FMA(r3, s3, b03));
+        MM_STORE(g+32*SIMDD, MM_FMA(s4, s4, b10));
+        MM_STORE(g+33*SIMDD, MM_FMA(s5, s5, b11));
+        MM_STORE(g+34*SIMDD, MM_FMA(s6, s6, b12));
+        MM_STORE(g+35*SIMDD, MM_FMA(s7, s7, b13));
+        MM_STORE(g+40*SIMDD, MM_FMA(r4, s4, b00));
+        MM_STORE(g+41*SIMDD, MM_FMA(r5, s5, b01));
+        MM_STORE(g+42*SIMDD, MM_FMA(r6, s6, b02));
+        MM_STORE(g+43*SIMDD, MM_FMA(r7, s7, b03));
+        MM_STORE(g+56*SIMDD, MM_MUL(MM_FMA(s8 , s8 , b10), g48));
+        MM_STORE(g+57*SIMDD, MM_MUL(MM_FMA(s9 , s9 , b11), g49));
+        MM_STORE(g+58*SIMDD, MM_MUL(MM_FMA(s10, s10, b12), g50));
+        MM_STORE(g+59*SIMDD, MM_MUL(MM_FMA(s11, s11, b13), g51));
+        MM_STORE(g+64*SIMDD, MM_MUL(MM_FMA(r8 , s8 , b00), g48));
+        MM_STORE(g+65*SIMDD, MM_MUL(MM_FMA(r9 , s9 , b01), g49));
+        MM_STORE(g+66*SIMDD, MM_MUL(MM_FMA(r10, s10, b02), g50));
+        MM_STORE(g+67*SIMDD, MM_MUL(MM_FMA(r11, s11, b03), g51));
+        MM_STORE(g+20*SIMDD, s0 *(MM_LOAD(g+16*SIMDD) + b00) + b10 * r0);
+        MM_STORE(g+21*SIMDD, s1 *(MM_LOAD(g+17*SIMDD) + b01) + b11 * r1);
+        MM_STORE(g+22*SIMDD, s2 *(MM_LOAD(g+18*SIMDD) + b02) + b12 * r2);
+        MM_STORE(g+23*SIMDD, s3 *(MM_LOAD(g+19*SIMDD) + b03) + b13 * r3);
+        MM_STORE(g+44*SIMDD, s4 *(MM_LOAD(g+40*SIMDD) + b00) + b10 * r4);
+        MM_STORE(g+45*SIMDD, s5 *(MM_LOAD(g+41*SIMDD) + b01) + b11 * r5);
+        MM_STORE(g+46*SIMDD, s6 *(MM_LOAD(g+42*SIMDD) + b02) + b12 * r6);
+        MM_STORE(g+47*SIMDD, s7 *(MM_LOAD(g+43*SIMDD) + b03) + b13 * r7);
+        MM_STORE(g+68*SIMDD, s8 * MM_LOAD(g+64*SIMDD) + b00 * MM_LOAD(g+52*SIMDD) + b10 * MM_LOAD(g+60*SIMDD));
+        MM_STORE(g+69*SIMDD, s9 * MM_LOAD(g+65*SIMDD) + b01 * MM_LOAD(g+53*SIMDD) + b11 * MM_LOAD(g+61*SIMDD));
+        MM_STORE(g+70*SIMDD, s10* MM_LOAD(g+66*SIMDD) + b02 * MM_LOAD(g+54*SIMDD) + b12 * MM_LOAD(g+62*SIMDD));
+        MM_STORE(g+71*SIMDD, s11* MM_LOAD(g+67*SIMDD) + b03 * MM_LOAD(g+55*SIMDD) + b13 * MM_LOAD(g+63*SIMDD));
+}
+
+static inline void _srg0_2d4d_0200(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c00x;
+        double *cy = bc->c00y;
+        double *cz = bc->c00z;
+        double *b =  bc->b10;
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cx2 = MM_LOAD(cx+2*SIMDD);
+        __MD cx3 = MM_LOAD(cx+3*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cy2 = MM_LOAD(cy+2*SIMDD);
+        __MD cy3 = MM_LOAD(cy+3*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD cz2 = MM_LOAD(cz+2*SIMDD);
+        __MD cz3 = MM_LOAD(cz+3*SIMDD);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD g24 = MM_LOAD(g+24*SIMDD);
+        __MD g25 = MM_LOAD(g+25*SIMDD);
+        __MD g26 = MM_LOAD(g+26*SIMDD);
+        __MD g27 = MM_LOAD(g+27*SIMDD);
+        MM_STORE(g+4 *SIMDD, cx0);
+        MM_STORE(g+5 *SIMDD, cx1);
+        MM_STORE(g+6 *SIMDD, cx2);
+        MM_STORE(g+7 *SIMDD, cx3);
+        MM_STORE(g+16*SIMDD, cy0);
+        MM_STORE(g+17*SIMDD, cy1);
+        MM_STORE(g+18*SIMDD, cy2);
+        MM_STORE(g+19*SIMDD, cy3);
+        MM_STORE(g+28*SIMDD, MM_MUL(cz0, g24));
+        MM_STORE(g+29*SIMDD, MM_MUL(cz1, g25));
+        MM_STORE(g+30*SIMDD, MM_MUL(cz2, g26));
+        MM_STORE(g+31*SIMDD, MM_MUL(cz3, g27));
+        MM_STORE(g+8 *SIMDD, MM_FMA(cx0, cx0, b0));
+        MM_STORE(g+9 *SIMDD, MM_FMA(cx1, cx1, b1));
+        MM_STORE(g+10*SIMDD, MM_FMA(cx2, cx2, b2));
+        MM_STORE(g+11*SIMDD, MM_FMA(cx3, cx3, b3));
+        MM_STORE(g+20*SIMDD, MM_FMA(cy0, cy0, b0));
+        MM_STORE(g+21*SIMDD, MM_FMA(cy1, cy1, b1));
+        MM_STORE(g+22*SIMDD, MM_FMA(cy2, cy2, b2));
+        MM_STORE(g+23*SIMDD, MM_FMA(cy3, cy3, b3));
+        MM_STORE(g+32*SIMDD, MM_MUL(MM_FMA(cz0, cz0, b0), g24));
+        MM_STORE(g+33*SIMDD, MM_MUL(MM_FMA(cz1, cz1, b1), g25));
+        MM_STORE(g+34*SIMDD, MM_MUL(MM_FMA(cz2, cz2, b2), g26));
+        MM_STORE(g+35*SIMDD, MM_MUL(MM_FMA(cz3, cz3, b3), g27));
+}
+
+static inline void _srg0_2d4d_0201(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0  = bc->b00;
+        double *b1  = bc->b10;
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0x2 = MM_LOAD(c0x+2*SIMDD);
+        __MD c0x3 = MM_LOAD(c0x+3*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0y2 = MM_LOAD(c0y+2*SIMDD);
+        __MD c0y3 = MM_LOAD(c0y+3*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
+        __MD c0z2 = MM_LOAD(c0z+2*SIMDD);
+        __MD c0z3 = MM_LOAD(c0z+3*SIMDD);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpx2 = MM_LOAD(cpx+2*SIMDD);
+        __MD cpx3 = MM_LOAD(cpx+3*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpy2 = MM_LOAD(cpy+2*SIMDD);
+        __MD cpy3 = MM_LOAD(cpy+3*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD cpz2 = MM_LOAD(cpz+2*SIMDD);
+        __MD cpz3 = MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        __MD i2 = MM_SET1(2.);
+        MM_STORE(g+4 *SIMDD, cpx0);
+        MM_STORE(g+5 *SIMDD, cpx1);
+        MM_STORE(g+6 *SIMDD, cpx2);
+        MM_STORE(g+7 *SIMDD, cpx3);
+        MM_STORE(g+8 *SIMDD, c0x0);
+        MM_STORE(g+9 *SIMDD, c0x1);
+        MM_STORE(g+10*SIMDD, c0x2);
+        MM_STORE(g+11*SIMDD, c0x3);
+        MM_STORE(g+28*SIMDD, cpy0);
+        MM_STORE(g+29*SIMDD, cpy1);
+        MM_STORE(g+30*SIMDD, cpy2);
+        MM_STORE(g+31*SIMDD, cpy3);
+        MM_STORE(g+32*SIMDD, c0y0);
+        MM_STORE(g+33*SIMDD, c0y1);
+        MM_STORE(g+34*SIMDD, c0y2);
+        MM_STORE(g+35*SIMDD, c0y3);
+        MM_STORE(g+52*SIMDD, cpz0 * g48);
+        MM_STORE(g+53*SIMDD, cpz1 * g49);
+        MM_STORE(g+54*SIMDD, cpz2 * g50);
+        MM_STORE(g+55*SIMDD, cpz3 * g51);
+        MM_STORE(g+56*SIMDD, c0z0 * g48);
+        MM_STORE(g+57*SIMDD, c0z1 * g49);
+        MM_STORE(g+58*SIMDD, c0z2 * g50);
+        MM_STORE(g+59*SIMDD, c0z3 * g51);
+        MM_STORE(g+12*SIMDD, cpx0 * c0x0 + b00);
+        MM_STORE(g+13*SIMDD, cpx1 * c0x1 + b01);
+        MM_STORE(g+14*SIMDD, cpx2 * c0x2 + b02);
+        MM_STORE(g+15*SIMDD, cpx3 * c0x3 + b03);
+        MM_STORE(g+16*SIMDD, c0x0 * c0x0 + b10);
+        MM_STORE(g+17*SIMDD, c0x1 * c0x1 + b11);
+        MM_STORE(g+18*SIMDD, c0x2 * c0x2 + b12);
+        MM_STORE(g+19*SIMDD, c0x3 * c0x3 + b13);
+        MM_STORE(g+36*SIMDD, cpy0 * c0y0 + b00);
+        MM_STORE(g+37*SIMDD, cpy1 * c0y1 + b01);
+        MM_STORE(g+38*SIMDD, cpy2 * c0y2 + b02);
+        MM_STORE(g+39*SIMDD, cpy3 * c0y3 + b03);
+        MM_STORE(g+40*SIMDD, c0y0 * c0y0 + b10);
+        MM_STORE(g+41*SIMDD, c0y1 * c0y1 + b11);
+        MM_STORE(g+42*SIMDD, c0y2 * c0y2 + b12);
+        MM_STORE(g+43*SIMDD, c0y3 * c0y3 + b13);
+        MM_STORE(g+60*SIMDD,(cpz0 * c0z0 + b00) * g48);
+        MM_STORE(g+61*SIMDD,(cpz1 * c0z1 + b01) * g49);
+        MM_STORE(g+62*SIMDD,(cpz2 * c0z2 + b02) * g50);
+        MM_STORE(g+63*SIMDD,(cpz3 * c0z3 + b03) * g51);
+        MM_STORE(g+64*SIMDD,(c0z0 * c0z0 + b10) * g48);
+        MM_STORE(g+65*SIMDD,(c0z1 * c0z1 + b11) * g49);
+        MM_STORE(g+66*SIMDD,(c0z2 * c0z2 + b12) * g50);
+        MM_STORE(g+67*SIMDD,(c0z3 * c0z3 + b13) * g51);
+        MM_STORE(g+20*SIMDD, cpx0 * MM_LOAD(g+16*SIMDD) + i2 * b00 * c0x0);
+        MM_STORE(g+21*SIMDD, cpx1 * MM_LOAD(g+17*SIMDD) + i2 * b01 * c0x1);
+        MM_STORE(g+22*SIMDD, cpx2 * MM_LOAD(g+18*SIMDD) + i2 * b02 * c0x2);
+        MM_STORE(g+23*SIMDD, cpx3 * MM_LOAD(g+19*SIMDD) + i2 * b03 * c0x3);
+        MM_STORE(g+44*SIMDD, cpy0 * MM_LOAD(g+40*SIMDD) + i2 * b00 * c0y0);
+        MM_STORE(g+45*SIMDD, cpy1 * MM_LOAD(g+41*SIMDD) + i2 * b01 * c0y1);
+        MM_STORE(g+46*SIMDD, cpy2 * MM_LOAD(g+42*SIMDD) + i2 * b02 * c0y2);
+        MM_STORE(g+47*SIMDD, cpy3 * MM_LOAD(g+43*SIMDD) + i2 * b03 * c0y3);
+        MM_STORE(g+68*SIMDD, cpz0 * MM_LOAD(g+64*SIMDD) + i2 * b00 * MM_LOAD(g+56*SIMDD));
+        MM_STORE(g+69*SIMDD, cpz1 * MM_LOAD(g+65*SIMDD) + i2 * b01 * MM_LOAD(g+57*SIMDD));
+        MM_STORE(g+70*SIMDD, cpz2 * MM_LOAD(g+66*SIMDD) + i2 * b02 * MM_LOAD(g+58*SIMDD));
+        MM_STORE(g+71*SIMDD, cpz3 * MM_LOAD(g+67*SIMDD) + i2 * b03 * MM_LOAD(g+59*SIMDD));
+}
+
+static inline void _srg0_2d4d_0210(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0 = bc->b00;
+        double *b1 = bc->b10;
+        __MD r0 = MM_LOAD(c0x+0*SIMDD);
+        __MD r1 = MM_LOAD(c0x+1*SIMDD);
+        __MD r2 = MM_LOAD(c0x+2*SIMDD);
+        __MD r3 = MM_LOAD(c0x+3*SIMDD);
+        __MD r4 = MM_LOAD(c0y+0*SIMDD);
+        __MD r5 = MM_LOAD(c0y+1*SIMDD);
+        __MD r6 = MM_LOAD(c0y+2*SIMDD);
+        __MD r7 = MM_LOAD(c0y+3*SIMDD);
+        __MD r8 = MM_LOAD(c0z+0*SIMDD);
+        __MD r9 = MM_LOAD(c0z+1*SIMDD);
+        __MD r10= MM_LOAD(c0z+2*SIMDD);
+        __MD r11= MM_LOAD(c0z+3*SIMDD);
+        __MD s0 = MM_LOAD(cpx+0*SIMDD);
+        __MD s1 = MM_LOAD(cpx+1*SIMDD);
+        __MD s2 = MM_LOAD(cpx+2*SIMDD);
+        __MD s3 = MM_LOAD(cpx+3*SIMDD);
+        __MD s4 = MM_LOAD(cpy+0*SIMDD);
+        __MD s5 = MM_LOAD(cpy+1*SIMDD);
+        __MD s6 = MM_LOAD(cpy+2*SIMDD);
+        __MD s7 = MM_LOAD(cpy+3*SIMDD);
+        __MD s8 = MM_LOAD(cpz+0*SIMDD);
+        __MD s9 = MM_LOAD(cpz+1*SIMDD);
+        __MD s10= MM_LOAD(cpz+2*SIMDD);
+        __MD s11= MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        MM_STORE(g+4 *SIMDD, s0);
+        MM_STORE(g+5 *SIMDD, s1);
+        MM_STORE(g+6 *SIMDD, s2);
+        MM_STORE(g+7 *SIMDD, s3);
+        MM_STORE(g+8 *SIMDD, r0);
+        MM_STORE(g+9 *SIMDD, r1);
+        MM_STORE(g+10*SIMDD, r2);
+        MM_STORE(g+11*SIMDD, r3);
+        MM_STORE(g+28*SIMDD, s4);
+        MM_STORE(g+29*SIMDD, s5);
+        MM_STORE(g+30*SIMDD, s6);
+        MM_STORE(g+31*SIMDD, s7);
+        MM_STORE(g+32*SIMDD, r4);
+        MM_STORE(g+33*SIMDD, r5);
+        MM_STORE(g+34*SIMDD, r6);
+        MM_STORE(g+35*SIMDD, r7);
+        MM_STORE(g+52*SIMDD, MM_MUL(s8, g48));
+        MM_STORE(g+53*SIMDD, MM_MUL(s9, g49));
+        MM_STORE(g+54*SIMDD, MM_MUL(s10,g50));
+        MM_STORE(g+55*SIMDD, MM_MUL(s11,g51));
+        MM_STORE(g+56*SIMDD, MM_MUL(r8, g48));
+        MM_STORE(g+57*SIMDD, MM_MUL(r9, g49));
+        MM_STORE(g+58*SIMDD, MM_MUL(r10,g50));
+        MM_STORE(g+59*SIMDD, MM_MUL(r11,g51));
+        MM_STORE(g+12*SIMDD, MM_FMA(s0, r0, b00));
+        MM_STORE(g+13*SIMDD, MM_FMA(s1, r1, b01));
+        MM_STORE(g+14*SIMDD, MM_FMA(s2, r2, b02));
+        MM_STORE(g+15*SIMDD, MM_FMA(s3, r3, b03));
+        MM_STORE(g+16*SIMDD, MM_FMA(r0, r0, b10));
+        MM_STORE(g+17*SIMDD, MM_FMA(r1, r1, b11));
+        MM_STORE(g+18*SIMDD, MM_FMA(r2, r2, b12));
+        MM_STORE(g+19*SIMDD, MM_FMA(r3, r3, b13));
+        MM_STORE(g+36*SIMDD, MM_FMA(s4, r4, b00));
+        MM_STORE(g+37*SIMDD, MM_FMA(s5, r5, b01));
+        MM_STORE(g+38*SIMDD, MM_FMA(s6, r6, b02));
+        MM_STORE(g+39*SIMDD, MM_FMA(s7, r7, b03));
+        MM_STORE(g+40*SIMDD, MM_FMA(r4, r4, b10));
+        MM_STORE(g+41*SIMDD, MM_FMA(r5, r5, b11));
+        MM_STORE(g+42*SIMDD, MM_FMA(r6, r6, b12));
+        MM_STORE(g+43*SIMDD, MM_FMA(r7, r7, b13));
+        MM_STORE(g+60*SIMDD, MM_MUL(MM_FMA(s8, r8, b00), g48));
+        MM_STORE(g+61*SIMDD, MM_MUL(MM_FMA(s9, r9, b01), g49));
+        MM_STORE(g+62*SIMDD, MM_MUL(MM_FMA(s10,r10,b02), g50));
+        MM_STORE(g+63*SIMDD, MM_MUL(MM_FMA(s11,r11,b03), g51));
+        MM_STORE(g+64*SIMDD, MM_MUL(MM_FMA(r8, r8, b10), g48));
+        MM_STORE(g+65*SIMDD, MM_MUL(MM_FMA(r9, r9, b11), g49));
+        MM_STORE(g+66*SIMDD, MM_MUL(MM_FMA(r10,r10,b12), g50));
+        MM_STORE(g+67*SIMDD, MM_MUL(MM_FMA(r11,r11,b13), g51));
+        MM_STORE(g+20*SIMDD, r0 *(MM_LOAD(g+12*SIMDD) + b00) + b10 * s0);
+        MM_STORE(g+21*SIMDD, r1 *(MM_LOAD(g+13*SIMDD) + b01) + b11 * s1);
+        MM_STORE(g+22*SIMDD, r2 *(MM_LOAD(g+14*SIMDD) + b02) + b12 * s2);
+        MM_STORE(g+23*SIMDD, r3 *(MM_LOAD(g+15*SIMDD) + b03) + b13 * s3);
+        MM_STORE(g+44*SIMDD, r4 *(MM_LOAD(g+36*SIMDD) + b00) + b10 * s4);
+        MM_STORE(g+45*SIMDD, r5 *(MM_LOAD(g+37*SIMDD) + b01) + b11 * s5);
+        MM_STORE(g+46*SIMDD, r6 *(MM_LOAD(g+38*SIMDD) + b02) + b12 * s6);
+        MM_STORE(g+47*SIMDD, r7 *(MM_LOAD(g+39*SIMDD) + b03) + b13 * s7);
+        MM_STORE(g+68*SIMDD, r8 * MM_LOAD(g+60*SIMDD) + b10 * MM_LOAD(g+52*SIMDD) + b00 * MM_LOAD(g+56*SIMDD));
+        MM_STORE(g+69*SIMDD, r9 * MM_LOAD(g+61*SIMDD) + b11 * MM_LOAD(g+53*SIMDD) + b01 * MM_LOAD(g+57*SIMDD));
+        MM_STORE(g+70*SIMDD, r10* MM_LOAD(g+62*SIMDD) + b12 * MM_LOAD(g+54*SIMDD) + b02 * MM_LOAD(g+58*SIMDD));
+        MM_STORE(g+71*SIMDD, r11* MM_LOAD(g+63*SIMDD) + b13 * MM_LOAD(g+55*SIMDD) + b03 * MM_LOAD(g+59*SIMDD));
+}
+
+static inline void _srg0_2d4d_0300(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c00x;
+        double *cy = bc->c00y;
+        double *cz = bc->c00z;
+        double *b =  bc->b10;
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cx2 = MM_LOAD(cx+2*SIMDD);
+        __MD cx3 = MM_LOAD(cx+3*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cy2 = MM_LOAD(cy+2*SIMDD);
+        __MD cy3 = MM_LOAD(cy+3*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD cz2 = MM_LOAD(cz+2*SIMDD);
+        __MD cz3 = MM_LOAD(cz+3*SIMDD);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD g34 = MM_LOAD(g+34*SIMDD);
+        __MD g35 = MM_LOAD(g+35*SIMDD);
+        __MD i3 = MM_SET1(3.);
+        MM_STORE(g+4 *SIMDD, cx0);
+        MM_STORE(g+5 *SIMDD, cx1);
+        MM_STORE(g+6 *SIMDD, cx2);
+        MM_STORE(g+7 *SIMDD, cx3);
+        MM_STORE(g+20*SIMDD, cy0);
+        MM_STORE(g+21*SIMDD, cy1);
+        MM_STORE(g+22*SIMDD, cy2);
+        MM_STORE(g+23*SIMDD, cy3);
+        MM_STORE(g+36*SIMDD, cz0 * g32);
+        MM_STORE(g+37*SIMDD, cz1 * g33);
+        MM_STORE(g+38*SIMDD, cz2 * g34);
+        MM_STORE(g+39*SIMDD, cz3 * g35);
+        MM_STORE(g+8 *SIMDD, cx0 * cx0 + b0);
+        MM_STORE(g+9 *SIMDD, cx1 * cx1 + b1);
+        MM_STORE(g+10*SIMDD, cx2 * cx2 + b2);
+        MM_STORE(g+11*SIMDD, cx3 * cx3 + b3);
+        MM_STORE(g+24*SIMDD, cy0 * cy0 + b0);
+        MM_STORE(g+25*SIMDD, cy1 * cy1 + b1);
+        MM_STORE(g+26*SIMDD, cy2 * cy2 + b2);
+        MM_STORE(g+27*SIMDD, cy3 * cy3 + b3);
+        MM_STORE(g+40*SIMDD,(cz0 * cz0 + b0)* g32);
+        MM_STORE(g+41*SIMDD,(cz1 * cz1 + b1)* g33);
+        MM_STORE(g+42*SIMDD,(cz2 * cz2 + b2)* g34);
+        MM_STORE(g+43*SIMDD,(cz3 * cz3 + b3)* g35);
+        MM_STORE(g+12*SIMDD, cx0 *(cx0 * cx0 + i3 * b0));
+        MM_STORE(g+13*SIMDD, cx1 *(cx1 * cx1 + i3 * b1));
+        MM_STORE(g+14*SIMDD, cx2 *(cx2 * cx2 + i3 * b2));
+        MM_STORE(g+15*SIMDD, cx3 *(cx3 * cx3 + i3 * b3));
+        MM_STORE(g+28*SIMDD, cy0 *(cy0 * cy0 + i3 * b0));
+        MM_STORE(g+29*SIMDD, cy1 *(cy1 * cy1 + i3 * b1));
+        MM_STORE(g+30*SIMDD, cy2 *(cy2 * cy2 + i3 * b2));
+        MM_STORE(g+31*SIMDD, cy3 *(cy3 * cy3 + i3 * b3));
+        MM_STORE(g+44*SIMDD,(cz0 * cz0 + i3 * b0)* MM_LOAD(g+36*SIMDD));
+        MM_STORE(g+45*SIMDD,(cz1 * cz1 + i3 * b1)* MM_LOAD(g+37*SIMDD));
+        MM_STORE(g+46*SIMDD,(cz2 * cz2 + i3 * b2)* MM_LOAD(g+38*SIMDD));
+        MM_STORE(g+47*SIMDD,(cz3 * cz3 + i3 * b3)* MM_LOAD(g+39*SIMDD));
+}
+
+static inline void _srg0_2d4d_1000(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c00x;
+        double *cy = bc->c00y;
+        double *cz = bc->c00z;
+        MM_STORE(g+2*SIMDD, MM_LOAD(cx+0*SIMDD));
+        MM_STORE(g+3*SIMDD, MM_LOAD(cx+1*SIMDD));
+        MM_STORE(g+6*SIMDD, MM_LOAD(cy+0*SIMDD));
+        MM_STORE(g+7*SIMDD, MM_LOAD(cy+1*SIMDD));
+        MM_STORE(g+10*SIMDD, MM_MUL(MM_LOAD(cz+0*SIMDD), MM_LOAD(g+8*SIMDD)));
+        MM_STORE(g+11*SIMDD, MM_MUL(MM_LOAD(cz+1*SIMDD), MM_LOAD(g+9*SIMDD)));
+}
+
+static inline void _srg0_2d4d_1001(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b   = bc->b00;
+        __MD cx0 = MM_LOAD(c0x+0*SIMDD);
+        __MD cx1 = MM_LOAD(c0x+1*SIMDD);
+        __MD cx2 = MM_LOAD(c0x+2*SIMDD);
+        __MD cx3 = MM_LOAD(c0x+3*SIMDD);
+        __MD cy0 = MM_LOAD(c0y+0*SIMDD);
+        __MD cy1 = MM_LOAD(c0y+1*SIMDD);
+        __MD cy2 = MM_LOAD(c0y+2*SIMDD);
+        __MD cy3 = MM_LOAD(c0y+3*SIMDD);
+        __MD cz0 = MM_LOAD(c0z+0*SIMDD);
+        __MD cz1 = MM_LOAD(c0z+1*SIMDD);
+        __MD cz2 = MM_LOAD(c0z+2*SIMDD);
+        __MD cz3 = MM_LOAD(c0z+3*SIMDD);
+        __MD px0 = MM_LOAD(cpx+0*SIMDD);
+        __MD px1 = MM_LOAD(cpx+1*SIMDD);
+        __MD px2 = MM_LOAD(cpx+2*SIMDD);
+        __MD px3 = MM_LOAD(cpx+3*SIMDD);
+        __MD py0 = MM_LOAD(cpy+0*SIMDD);
+        __MD py1 = MM_LOAD(cpy+1*SIMDD);
+        __MD py2 = MM_LOAD(cpy+2*SIMDD);
+        __MD py3 = MM_LOAD(cpy+3*SIMDD);
+        __MD pz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD pz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD pz2 = MM_LOAD(cpz+2*SIMDD);
+        __MD pz3 = MM_LOAD(cpz+3*SIMDD);
+        __MD b0  = MM_LOAD(b+0*SIMDD);
+        __MD b1  = MM_LOAD(b+1*SIMDD);
+        __MD b2  = MM_LOAD(b+2*SIMDD);
+        __MD b3  = MM_LOAD(b+3*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD g34 = MM_LOAD(g+34*SIMDD);
+        __MD g35 = MM_LOAD(g+35*SIMDD);
+        MM_STORE(g+4 *SIMDD, cx0);
+        MM_STORE(g+5 *SIMDD, cx1);
+        MM_STORE(g+6 *SIMDD, cx2);
+        MM_STORE(g+7 *SIMDD, cx3);
+        MM_STORE(g+8 *SIMDD, px0);
+        MM_STORE(g+9 *SIMDD, px1);
+        MM_STORE(g+10*SIMDD, px2);
+        MM_STORE(g+11*SIMDD, px3);
+        MM_STORE(g+20*SIMDD, cy0);
+        MM_STORE(g+21*SIMDD, cy1);
+        MM_STORE(g+22*SIMDD, cy2);
+        MM_STORE(g+23*SIMDD, cy3);
+        MM_STORE(g+24*SIMDD, py0);
+        MM_STORE(g+25*SIMDD, py1);
+        MM_STORE(g+26*SIMDD, py2);
+        MM_STORE(g+27*SIMDD, py3);
+        MM_STORE(g+36*SIMDD, MM_MUL(cz0, g32));
+        MM_STORE(g+37*SIMDD, MM_MUL(cz1, g33));
+        MM_STORE(g+38*SIMDD, MM_MUL(cz2, g34));
+        MM_STORE(g+39*SIMDD, MM_MUL(cz3, g35));
+        MM_STORE(g+40*SIMDD, MM_MUL(pz0, g32));
+        MM_STORE(g+41*SIMDD, MM_MUL(pz1, g33));
+        MM_STORE(g+42*SIMDD, MM_MUL(pz2, g34));
+        MM_STORE(g+43*SIMDD, MM_MUL(pz3, g35));
+        MM_STORE(g+12*SIMDD, MM_FMA(px0, cx0, b0));
+        MM_STORE(g+13*SIMDD, MM_FMA(px1, cx1, b1));
+        MM_STORE(g+14*SIMDD, MM_FMA(px2, cx2, b2));
+        MM_STORE(g+15*SIMDD, MM_FMA(px3, cx3, b3));
+        MM_STORE(g+28*SIMDD, MM_FMA(py0, cy0, b0));
+        MM_STORE(g+29*SIMDD, MM_FMA(py1, cy1, b1));
+        MM_STORE(g+30*SIMDD, MM_FMA(py2, cy2, b2));
+        MM_STORE(g+31*SIMDD, MM_FMA(py3, cy3, b3));
+        MM_STORE(g+44*SIMDD, MM_MUL(MM_FMA(pz0, cz0, b0), g32));
+        MM_STORE(g+45*SIMDD, MM_MUL(MM_FMA(pz1, cz1, b1), g33));
+        MM_STORE(g+46*SIMDD, MM_MUL(MM_FMA(pz2, cz2, b2), g34));
+        MM_STORE(g+47*SIMDD, MM_MUL(MM_FMA(pz3, cz3, b3), g35));
+}
+
+static inline void _srg0_2d4d_1002(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0 = bc->b00;
+        double *b1 = bc->b01;
+        __MD r0 = MM_LOAD(c0x+0*SIMDD);
+        __MD r1 = MM_LOAD(c0x+1*SIMDD);
+        __MD r2 = MM_LOAD(c0x+2*SIMDD);
+        __MD r3 = MM_LOAD(c0x+3*SIMDD);
+        __MD r4 = MM_LOAD(c0y+0*SIMDD);
+        __MD r5 = MM_LOAD(c0y+1*SIMDD);
+        __MD r6 = MM_LOAD(c0y+2*SIMDD);
+        __MD r7 = MM_LOAD(c0y+3*SIMDD);
+        __MD r8 = MM_LOAD(c0z+0*SIMDD);
+        __MD r9 = MM_LOAD(c0z+1*SIMDD);
+        __MD r10= MM_LOAD(c0z+2*SIMDD);
+        __MD r11= MM_LOAD(c0z+3*SIMDD);
+        __MD s0 = MM_LOAD(cpx+0*SIMDD);
+        __MD s1 = MM_LOAD(cpx+1*SIMDD);
+        __MD s2 = MM_LOAD(cpx+2*SIMDD);
+        __MD s3 = MM_LOAD(cpx+3*SIMDD);
+        __MD s4 = MM_LOAD(cpy+0*SIMDD);
+        __MD s5 = MM_LOAD(cpy+1*SIMDD);
+        __MD s6 = MM_LOAD(cpy+2*SIMDD);
+        __MD s7 = MM_LOAD(cpy+3*SIMDD);
+        __MD s8 = MM_LOAD(cpz+0*SIMDD);
+        __MD s9 = MM_LOAD(cpz+1*SIMDD);
+        __MD s10= MM_LOAD(cpz+2*SIMDD);
+        __MD s11= MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+8 *SIMDD, s0);
+        MM_STORE(g+9 *SIMDD, s1);
+        MM_STORE(g+10*SIMDD, s2);
+        MM_STORE(g+11*SIMDD, s3);
+        MM_STORE(g+28*SIMDD, r4);
+        MM_STORE(g+29*SIMDD, r5);
+        MM_STORE(g+30*SIMDD, r6);
+        MM_STORE(g+31*SIMDD, r7);
+        MM_STORE(g+32*SIMDD, s4);
+        MM_STORE(g+33*SIMDD, s5);
+        MM_STORE(g+34*SIMDD, s6);
+        MM_STORE(g+35*SIMDD, s7);
+        MM_STORE(g+52*SIMDD, MM_MUL(r8, g48));
+        MM_STORE(g+53*SIMDD, MM_MUL(r9, g49));
+        MM_STORE(g+54*SIMDD, MM_MUL(r10,g50));
+        MM_STORE(g+55*SIMDD, MM_MUL(r11,g51));
+        MM_STORE(g+56*SIMDD, MM_MUL(s8, g48));
+        MM_STORE(g+57*SIMDD, MM_MUL(s9, g49));
+        MM_STORE(g+58*SIMDD, MM_MUL(s10,g50));
+        MM_STORE(g+59*SIMDD, MM_MUL(s11,g51));
+        MM_STORE(g+12*SIMDD, MM_FMA(r0, s0, b00));
+        MM_STORE(g+13*SIMDD, MM_FMA(r1, s1, b01));
+        MM_STORE(g+14*SIMDD, MM_FMA(r2, s2, b02));
+        MM_STORE(g+15*SIMDD, MM_FMA(r3, s3, b03));
+        MM_STORE(g+16*SIMDD, MM_FMA(s0, s0, b10));
+        MM_STORE(g+17*SIMDD, MM_FMA(s1, s1, b11));
+        MM_STORE(g+18*SIMDD, MM_FMA(s2, s2, b12));
+        MM_STORE(g+19*SIMDD, MM_FMA(s3, s3, b13));
+        MM_STORE(g+36*SIMDD, MM_FMA(r4, s4, b00));
+        MM_STORE(g+37*SIMDD, MM_FMA(r5, s5, b01));
+        MM_STORE(g+38*SIMDD, MM_FMA(r6, s6, b02));
+        MM_STORE(g+39*SIMDD, MM_FMA(r7, s7, b03));
+        MM_STORE(g+40*SIMDD, MM_FMA(s4, s4, b10));
+        MM_STORE(g+41*SIMDD, MM_FMA(s5, s5, b11));
+        MM_STORE(g+42*SIMDD, MM_FMA(s6, s6, b12));
+        MM_STORE(g+43*SIMDD, MM_FMA(s7, s7, b13));
+        MM_STORE(g+60*SIMDD, MM_MUL(MM_FMA(r8, s8, b00), g48));
+        MM_STORE(g+61*SIMDD, MM_MUL(MM_FMA(r9, s9, b01), g49));
+        MM_STORE(g+62*SIMDD, MM_MUL(MM_FMA(r10,s10,b02), g50));
+        MM_STORE(g+63*SIMDD, MM_MUL(MM_FMA(r11,s11,b03), g51));
+        MM_STORE(g+64*SIMDD, MM_MUL(MM_FMA(s8, s8, b10), g48));
+        MM_STORE(g+65*SIMDD, MM_MUL(MM_FMA(s9, s9, b11), g49));
+        MM_STORE(g+66*SIMDD, MM_MUL(MM_FMA(s10,s10,b12), g50));
+        MM_STORE(g+67*SIMDD, MM_MUL(MM_FMA(s11,s11,b13), g51));
+        MM_STORE(g+20*SIMDD, s0 *(MM_LOAD(g+12*SIMDD) + b00) + b10 * r0);
+        MM_STORE(g+21*SIMDD, s1 *(MM_LOAD(g+13*SIMDD) + b01) + b11 * r1);
+        MM_STORE(g+22*SIMDD, s2 *(MM_LOAD(g+14*SIMDD) + b02) + b12 * r2);
+        MM_STORE(g+23*SIMDD, s3 *(MM_LOAD(g+15*SIMDD) + b03) + b13 * r3);
+        MM_STORE(g+44*SIMDD, s4 *(MM_LOAD(g+36*SIMDD) + b00) + b10 * r4);
+        MM_STORE(g+45*SIMDD, s5 *(MM_LOAD(g+37*SIMDD) + b01) + b11 * r5);
+        MM_STORE(g+46*SIMDD, s6 *(MM_LOAD(g+38*SIMDD) + b02) + b12 * r6);
+        MM_STORE(g+47*SIMDD, s7 *(MM_LOAD(g+39*SIMDD) + b03) + b13 * r7);
+        MM_STORE(g+68*SIMDD, s8 * MM_LOAD(g+60*SIMDD) + b10 * MM_LOAD(g+52*SIMDD) + b00 * MM_LOAD(g+56*SIMDD));
+        MM_STORE(g+69*SIMDD, s9 * MM_LOAD(g+61*SIMDD) + b11 * MM_LOAD(g+53*SIMDD) + b01 * MM_LOAD(g+57*SIMDD));
+        MM_STORE(g+70*SIMDD, s10* MM_LOAD(g+62*SIMDD) + b12 * MM_LOAD(g+54*SIMDD) + b02 * MM_LOAD(g+58*SIMDD));
+        MM_STORE(g+71*SIMDD, s11* MM_LOAD(g+63*SIMDD) + b13 * MM_LOAD(g+55*SIMDD) + b03 * MM_LOAD(g+59*SIMDD));
+}
+
+static inline void _srg0_2d4d_1010(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b   = bc->b00;
+        __MD cx0 = MM_LOAD(c0x+0*SIMDD);
+        __MD cx1 = MM_LOAD(c0x+1*SIMDD);
+        __MD cx2 = MM_LOAD(c0x+2*SIMDD);
+        __MD cx3 = MM_LOAD(c0x+3*SIMDD);
+        __MD cy0 = MM_LOAD(c0y+0*SIMDD);
+        __MD cy1 = MM_LOAD(c0y+1*SIMDD);
+        __MD cy2 = MM_LOAD(c0y+2*SIMDD);
+        __MD cy3 = MM_LOAD(c0y+3*SIMDD);
+        __MD cz0 = MM_LOAD(c0z+0*SIMDD);
+        __MD cz1 = MM_LOAD(c0z+1*SIMDD);
+        __MD cz2 = MM_LOAD(c0z+2*SIMDD);
+        __MD cz3 = MM_LOAD(c0z+3*SIMDD);
+        __MD px0 = MM_LOAD(cpx+0*SIMDD);
+        __MD px1 = MM_LOAD(cpx+1*SIMDD);
+        __MD px2 = MM_LOAD(cpx+2*SIMDD);
+        __MD px3 = MM_LOAD(cpx+3*SIMDD);
+        __MD py0 = MM_LOAD(cpy+0*SIMDD);
+        __MD py1 = MM_LOAD(cpy+1*SIMDD);
+        __MD py2 = MM_LOAD(cpy+2*SIMDD);
+        __MD py3 = MM_LOAD(cpy+3*SIMDD);
+        __MD pz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD pz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD pz2 = MM_LOAD(cpz+2*SIMDD);
+        __MD pz3 = MM_LOAD(cpz+3*SIMDD);
+        __MD b0  = MM_LOAD(b+0*SIMDD);
+        __MD b1  = MM_LOAD(b+1*SIMDD);
+        __MD b2  = MM_LOAD(b+2*SIMDD);
+        __MD b3  = MM_LOAD(b+3*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD g34 = MM_LOAD(g+34*SIMDD);
+        __MD g35 = MM_LOAD(g+35*SIMDD);
+        MM_STORE(g+4 *SIMDD, cx0);
+        MM_STORE(g+5 *SIMDD, cx1);
+        MM_STORE(g+6 *SIMDD, cx2);
+        MM_STORE(g+7 *SIMDD, cx3);
+        MM_STORE(g+8 *SIMDD, px0);
+        MM_STORE(g+9 *SIMDD, px1);
+        MM_STORE(g+10*SIMDD, px2);
+        MM_STORE(g+11*SIMDD, px3);
+        MM_STORE(g+20*SIMDD, cy0);
+        MM_STORE(g+21*SIMDD, cy1);
+        MM_STORE(g+22*SIMDD, cy2);
+        MM_STORE(g+23*SIMDD, cy3);
+        MM_STORE(g+24*SIMDD, py0);
+        MM_STORE(g+25*SIMDD, py1);
+        MM_STORE(g+26*SIMDD, py2);
+        MM_STORE(g+27*SIMDD, py3);
+        MM_STORE(g+36*SIMDD, MM_MUL(cz0, g32));
+        MM_STORE(g+37*SIMDD, MM_MUL(cz1, g33));
+        MM_STORE(g+38*SIMDD, MM_MUL(cz2, g34));
+        MM_STORE(g+39*SIMDD, MM_MUL(cz3, g35));
+        MM_STORE(g+40*SIMDD, MM_MUL(pz0, g32));
+        MM_STORE(g+41*SIMDD, MM_MUL(pz1, g33));
+        MM_STORE(g+42*SIMDD, MM_MUL(pz2, g34));
+        MM_STORE(g+43*SIMDD, MM_MUL(pz3, g35));
+        MM_STORE(g+12*SIMDD, MM_FMA(px0, cx0, b0));
+        MM_STORE(g+13*SIMDD, MM_FMA(px1, cx1, b1));
+        MM_STORE(g+14*SIMDD, MM_FMA(px2, cx2, b2));
+        MM_STORE(g+15*SIMDD, MM_FMA(px3, cx3, b3));
+        MM_STORE(g+28*SIMDD, MM_FMA(py0, cy0, b0));
+        MM_STORE(g+29*SIMDD, MM_FMA(py1, cy1, b1));
+        MM_STORE(g+30*SIMDD, MM_FMA(py2, cy2, b2));
+        MM_STORE(g+31*SIMDD, MM_FMA(py3, cy3, b3));
+        MM_STORE(g+44*SIMDD, MM_MUL(MM_FMA(pz0, cz0, b0), g32));
+        MM_STORE(g+45*SIMDD, MM_MUL(MM_FMA(pz1, cz1, b1), g33));
+        MM_STORE(g+46*SIMDD, MM_MUL(MM_FMA(pz2, cz2, b2), g34));
+        MM_STORE(g+47*SIMDD, MM_MUL(MM_FMA(pz3, cz3, b3), g35));
+}
+
+static inline void _srg0_2d4d_1011(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0  = bc->b00;
+        double *b1  = bc->b01;
+        double *r   = envs->rkrl;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpx2 = MM_LOAD(cpx+2*SIMDD);
+        __MD cpx3 = MM_LOAD(cpx+3*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpy2 = MM_LOAD(cpy+2*SIMDD);
+        __MD cpy3 = MM_LOAD(cpy+3*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD cpz2 = MM_LOAD(cpz+2*SIMDD);
+        __MD cpz3 = MM_LOAD(cpz+3*SIMDD);
+        __MD rcp0 = MM_ADD(rx, cpx0);
+        __MD rcp1 = MM_ADD(rx, cpx1);
+        __MD rcp2 = MM_ADD(rx, cpx2);
+        __MD rcp3 = MM_ADD(rx, cpx3);
+        __MD rcp4 = MM_ADD(ry, cpy0);
+        __MD rcp5 = MM_ADD(ry, cpy1);
+        __MD rcp6 = MM_ADD(ry, cpy2);
+        __MD rcp7 = MM_ADD(ry, cpy3);
+        __MD rcp8 = MM_ADD(rz, cpz0);
+        __MD rcp9 = MM_ADD(rz, cpz1);
+        __MD rcp10= MM_ADD(rz, cpz2);
+        __MD rcp11= MM_ADD(rz, cpz3);
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0x2 = MM_LOAD(c0x+2*SIMDD);
+        __MD c0x3 = MM_LOAD(c0x+3*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0y2 = MM_LOAD(c0y+2*SIMDD);
+        __MD c0y3 = MM_LOAD(c0y+3*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
+        __MD c0z2 = MM_LOAD(c0z+2*SIMDD);
+        __MD c0z3 = MM_LOAD(c0z+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g96 = MM_LOAD(g+96*SIMDD);
+        __MD g97 = MM_LOAD(g+97*SIMDD);
+        __MD g98 = MM_LOAD(g+98*SIMDD);
+        __MD g99 = MM_LOAD(g+99*SIMDD);
+        MM_STORE(g+4 *SIMDD, c0x0);
+        MM_STORE(g+5 *SIMDD, c0x1);
+        MM_STORE(g+6 *SIMDD, c0x2);
+        MM_STORE(g+7 *SIMDD, c0x3);
+        MM_STORE(g+8 *SIMDD, rcp0);
+        MM_STORE(g+9 *SIMDD, rcp1);
+        MM_STORE(g+10*SIMDD, rcp2);
+        MM_STORE(g+11*SIMDD, rcp3);
+        MM_STORE(g+16*SIMDD, cpx0);
+        MM_STORE(g+17*SIMDD, cpx1);
+        MM_STORE(g+18*SIMDD, cpx2);
+        MM_STORE(g+19*SIMDD, cpx3);
+        MM_STORE(g+52*SIMDD, c0y0);
+        MM_STORE(g+53*SIMDD, c0y1);
+        MM_STORE(g+54*SIMDD, c0y2);
+        MM_STORE(g+55*SIMDD, c0y3);
+        MM_STORE(g+56*SIMDD, rcp4);
+        MM_STORE(g+57*SIMDD, rcp5);
+        MM_STORE(g+58*SIMDD, rcp6);
+        MM_STORE(g+59*SIMDD, rcp7);
+        MM_STORE(g+64*SIMDD, cpy0);
+        MM_STORE(g+65*SIMDD, cpy1);
+        MM_STORE(g+66*SIMDD, cpy2);
+        MM_STORE(g+67*SIMDD, cpy3);
+        MM_STORE(g+100*SIMDD, c0z0 * g96);
+        MM_STORE(g+101*SIMDD, c0z1 * g97);
+        MM_STORE(g+102*SIMDD, c0z2 * g98);
+        MM_STORE(g+103*SIMDD, c0z3 * g99);
+        MM_STORE(g+104*SIMDD, rcp8 * g96);
+        MM_STORE(g+105*SIMDD, rcp9 * g97);
+        MM_STORE(g+106*SIMDD, rcp10* g98);
+        MM_STORE(g+107*SIMDD, rcp11* g99);
+        MM_STORE(g+112*SIMDD, cpz0 * g96);
+        MM_STORE(g+113*SIMDD, cpz1 * g97);
+        MM_STORE(g+114*SIMDD, cpz2 * g98);
+        MM_STORE(g+115*SIMDD, cpz3 * g99);
+        MM_STORE(g+12*SIMDD, rcp0 * c0x0 + b00);
+        MM_STORE(g+13*SIMDD, rcp1 * c0x1 + b01);
+        MM_STORE(g+14*SIMDD, rcp2 * c0x2 + b02);
+        MM_STORE(g+15*SIMDD, rcp3 * c0x3 + b03);
+        MM_STORE(g+20*SIMDD, cpx0 * c0x0 + b00);
+        MM_STORE(g+21*SIMDD, cpx1 * c0x1 + b01);
+        MM_STORE(g+22*SIMDD, cpx2 * c0x2 + b02);
+        MM_STORE(g+23*SIMDD, cpx3 * c0x3 + b03);
+        MM_STORE(g+24*SIMDD, cpx0 * rcp0 + b10);
+        MM_STORE(g+25*SIMDD, cpx1 * rcp1 + b11);
+        MM_STORE(g+26*SIMDD, cpx2 * rcp2 + b12);
+        MM_STORE(g+27*SIMDD, cpx3 * rcp3 + b13);
+        MM_STORE(g+60*SIMDD, rcp4 * c0y0 + b00);
+        MM_STORE(g+61*SIMDD, rcp5 * c0y1 + b01);
+        MM_STORE(g+62*SIMDD, rcp6 * c0y2 + b02);
+        MM_STORE(g+63*SIMDD, rcp7 * c0y3 + b03);
+        MM_STORE(g+68*SIMDD, cpy0 * c0y0 + b00);
+        MM_STORE(g+69*SIMDD, cpy1 * c0y1 + b01);
+        MM_STORE(g+70*SIMDD, cpy2 * c0y2 + b02);
+        MM_STORE(g+71*SIMDD, cpy3 * c0y3 + b03);
+        MM_STORE(g+72*SIMDD, cpy0 * rcp4 + b10);
+        MM_STORE(g+73*SIMDD, cpy1 * rcp5 + b11);
+        MM_STORE(g+74*SIMDD, cpy2 * rcp6 + b12);
+        MM_STORE(g+75*SIMDD, cpy3 * rcp7 + b13);
+        MM_STORE(g+108*SIMDD,(rcp8 * c0z0 + b00)* g96);
+        MM_STORE(g+109*SIMDD,(rcp9 * c0z1 + b01)* g97);
+        MM_STORE(g+110*SIMDD,(rcp10* c0z2 + b02)* g98);
+        MM_STORE(g+111*SIMDD,(rcp11* c0z3 + b03)* g99);
+        MM_STORE(g+116*SIMDD,(cpz0 * c0z0 + b00)* g96);
+        MM_STORE(g+117*SIMDD,(cpz1 * c0z1 + b01)* g97);
+        MM_STORE(g+118*SIMDD,(cpz2 * c0z2 + b02)* g98);
+        MM_STORE(g+119*SIMDD,(cpz3 * c0z3 + b03)* g99);
+        MM_STORE(g+120*SIMDD,(cpz0 * rcp8 + b10)* g96);
+        MM_STORE(g+121*SIMDD,(cpz1 * rcp9 + b11)* g97);
+        MM_STORE(g+122*SIMDD,(cpz2 * rcp10+ b12)* g98);
+        MM_STORE(g+123*SIMDD,(cpz3 * rcp11+ b13)* g99);
+        MM_STORE(g+28*SIMDD , rcp0 * MM_LOAD(g+20 *SIMDD) + b00 * cpx0 + b10 * c0x0);
+        MM_STORE(g+29*SIMDD , rcp1 * MM_LOAD(g+21 *SIMDD) + b01 * cpx1 + b11 * c0x1);
+        MM_STORE(g+30*SIMDD , rcp2 * MM_LOAD(g+22 *SIMDD) + b02 * cpx2 + b12 * c0x2);
+        MM_STORE(g+31*SIMDD , rcp3 * MM_LOAD(g+23 *SIMDD) + b03 * cpx3 + b13 * c0x3);
+        MM_STORE(g+76*SIMDD , rcp4 * MM_LOAD(g+68 *SIMDD) + b00 * cpy0 + b10 * c0y0);
+        MM_STORE(g+77*SIMDD , rcp5 * MM_LOAD(g+69 *SIMDD) + b01 * cpy1 + b11 * c0y1);
+        MM_STORE(g+78*SIMDD , rcp6 * MM_LOAD(g+70 *SIMDD) + b02 * cpy2 + b12 * c0y2);
+        MM_STORE(g+79*SIMDD , rcp7 * MM_LOAD(g+71 *SIMDD) + b03 * cpy3 + b13 * c0y3);
+        MM_STORE(g+124*SIMDD, rcp8 * MM_LOAD(g+116*SIMDD) + b00 * MM_LOAD(g+112*SIMDD) + b10 * MM_LOAD(g+100*SIMDD));
+        MM_STORE(g+125*SIMDD, rcp9 * MM_LOAD(g+117*SIMDD) + b01 * MM_LOAD(g+113*SIMDD) + b11 * MM_LOAD(g+101*SIMDD));
+        MM_STORE(g+126*SIMDD, rcp10* MM_LOAD(g+118*SIMDD) + b02 * MM_LOAD(g+114*SIMDD) + b12 * MM_LOAD(g+102*SIMDD));
+        MM_STORE(g+127*SIMDD, rcp11* MM_LOAD(g+119*SIMDD) + b03 * MM_LOAD(g+115*SIMDD) + b13 * MM_LOAD(g+103*SIMDD));
+}
+
+static inline void _srg0_2d4d_1020(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0 = bc->b00;
+        double *b1 = bc->b01;
+        __MD r0 = MM_LOAD(c0x+0*SIMDD);
+        __MD r1 = MM_LOAD(c0x+1*SIMDD);
+        __MD r2 = MM_LOAD(c0x+2*SIMDD);
+        __MD r3 = MM_LOAD(c0x+3*SIMDD);
+        __MD r4 = MM_LOAD(c0y+0*SIMDD);
+        __MD r5 = MM_LOAD(c0y+1*SIMDD);
+        __MD r6 = MM_LOAD(c0y+2*SIMDD);
+        __MD r7 = MM_LOAD(c0y+3*SIMDD);
+        __MD r8 = MM_LOAD(c0z+0*SIMDD);
+        __MD r9 = MM_LOAD(c0z+1*SIMDD);
+        __MD r10= MM_LOAD(c0z+2*SIMDD);
+        __MD r11= MM_LOAD(c0z+3*SIMDD);
+        __MD s0 = MM_LOAD(cpx+0*SIMDD);
+        __MD s1 = MM_LOAD(cpx+1*SIMDD);
+        __MD s2 = MM_LOAD(cpx+2*SIMDD);
+        __MD s3 = MM_LOAD(cpx+3*SIMDD);
+        __MD s4 = MM_LOAD(cpy+0*SIMDD);
+        __MD s5 = MM_LOAD(cpy+1*SIMDD);
+        __MD s6 = MM_LOAD(cpy+2*SIMDD);
+        __MD s7 = MM_LOAD(cpy+3*SIMDD);
+        __MD s8 = MM_LOAD(cpz+0*SIMDD);
+        __MD s9 = MM_LOAD(cpz+1*SIMDD);
+        __MD s10= MM_LOAD(cpz+2*SIMDD);
+        __MD s11= MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+8 *SIMDD, s0);
+        MM_STORE(g+9 *SIMDD, s1);
+        MM_STORE(g+10*SIMDD, s2);
+        MM_STORE(g+11*SIMDD, s3);
+        MM_STORE(g+28*SIMDD, r4);
+        MM_STORE(g+29*SIMDD, r5);
+        MM_STORE(g+30*SIMDD, r6);
+        MM_STORE(g+31*SIMDD, r7);
+        MM_STORE(g+32*SIMDD, s4);
+        MM_STORE(g+33*SIMDD, s5);
+        MM_STORE(g+34*SIMDD, s6);
+        MM_STORE(g+35*SIMDD, s7);
+        MM_STORE(g+52*SIMDD, MM_MUL(r8, g48));
+        MM_STORE(g+53*SIMDD, MM_MUL(r9, g49));
+        MM_STORE(g+54*SIMDD, MM_MUL(r10,g50));
+        MM_STORE(g+55*SIMDD, MM_MUL(r11,g51));
+        MM_STORE(g+56*SIMDD, MM_MUL(s8, g48));
+        MM_STORE(g+57*SIMDD, MM_MUL(s9, g49));
+        MM_STORE(g+58*SIMDD, MM_MUL(s10,g50));
+        MM_STORE(g+59*SIMDD, MM_MUL(s11,g51));
+        MM_STORE(g+12*SIMDD, MM_FMA(r0, s0, b00));
+        MM_STORE(g+13*SIMDD, MM_FMA(r1, s1, b01));
+        MM_STORE(g+14*SIMDD, MM_FMA(r2, s2, b02));
+        MM_STORE(g+15*SIMDD, MM_FMA(r3, s3, b03));
+        MM_STORE(g+16*SIMDD, MM_FMA(s0, s0, b10));
+        MM_STORE(g+17*SIMDD, MM_FMA(s1, s1, b11));
+        MM_STORE(g+18*SIMDD, MM_FMA(s2, s2, b12));
+        MM_STORE(g+19*SIMDD, MM_FMA(s3, s3, b13));
+        MM_STORE(g+36*SIMDD, MM_FMA(r4, s4, b00));
+        MM_STORE(g+37*SIMDD, MM_FMA(r5, s5, b01));
+        MM_STORE(g+38*SIMDD, MM_FMA(r6, s6, b02));
+        MM_STORE(g+39*SIMDD, MM_FMA(r7, s7, b03));
+        MM_STORE(g+40*SIMDD, MM_FMA(s4, s4, b10));
+        MM_STORE(g+41*SIMDD, MM_FMA(s5, s5, b11));
+        MM_STORE(g+42*SIMDD, MM_FMA(s6, s6, b12));
+        MM_STORE(g+43*SIMDD, MM_FMA(s7, s7, b13));
+        MM_STORE(g+60*SIMDD, MM_MUL(MM_FMA(r8, s8, b00), g48));
+        MM_STORE(g+61*SIMDD, MM_MUL(MM_FMA(r9, s9, b01), g49));
+        MM_STORE(g+62*SIMDD, MM_MUL(MM_FMA(r10,s10,b02), g50));
+        MM_STORE(g+63*SIMDD, MM_MUL(MM_FMA(r11,s11,b03), g51));
+        MM_STORE(g+64*SIMDD, MM_MUL(MM_FMA(s8, s8, b10), g48));
+        MM_STORE(g+65*SIMDD, MM_MUL(MM_FMA(s9, s9, b11), g49));
+        MM_STORE(g+66*SIMDD, MM_MUL(MM_FMA(s10,s10,b12), g50));
+        MM_STORE(g+67*SIMDD, MM_MUL(MM_FMA(s11,s11,b13), g51));
+        MM_STORE(g+20*SIMDD, s0 *(MM_LOAD(g+12*SIMDD) + b00) + b10 * r0);
+        MM_STORE(g+21*SIMDD, s1 *(MM_LOAD(g+13*SIMDD) + b01) + b11 * r1);
+        MM_STORE(g+22*SIMDD, s2 *(MM_LOAD(g+14*SIMDD) + b02) + b12 * r2);
+        MM_STORE(g+23*SIMDD, s3 *(MM_LOAD(g+15*SIMDD) + b03) + b13 * r3);
+        MM_STORE(g+44*SIMDD, s4 *(MM_LOAD(g+36*SIMDD) + b00) + b10 * r4);
+        MM_STORE(g+45*SIMDD, s5 *(MM_LOAD(g+37*SIMDD) + b01) + b11 * r5);
+        MM_STORE(g+46*SIMDD, s6 *(MM_LOAD(g+38*SIMDD) + b02) + b12 * r6);
+        MM_STORE(g+47*SIMDD, s7 *(MM_LOAD(g+39*SIMDD) + b03) + b13 * r7);
+        MM_STORE(g+68*SIMDD, s8 * MM_LOAD(g+60*SIMDD) + b10 * MM_LOAD(g+52*SIMDD) + b00 * MM_LOAD(g+56*SIMDD));
+        MM_STORE(g+69*SIMDD, s9 * MM_LOAD(g+61*SIMDD) + b11 * MM_LOAD(g+53*SIMDD) + b01 * MM_LOAD(g+57*SIMDD));
+        MM_STORE(g+70*SIMDD, s10* MM_LOAD(g+62*SIMDD) + b12 * MM_LOAD(g+54*SIMDD) + b02 * MM_LOAD(g+58*SIMDD));
+        MM_STORE(g+71*SIMDD, s11* MM_LOAD(g+63*SIMDD) + b13 * MM_LOAD(g+55*SIMDD) + b03 * MM_LOAD(g+59*SIMDD));
+}
+
+static inline void _srg0_2d4d_1100(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c00x;
+        double *cy = bc->c00y;
+        double *cz = bc->c00z;
+        double *b  = bc->b10;
+        double *r  = envs->rirj;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cx2 = MM_LOAD(cx+2*SIMDD);
+        __MD cx3 = MM_LOAD(cx+3*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cy2 = MM_LOAD(cy+2*SIMDD);
+        __MD cy3 = MM_LOAD(cy+3*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD cz2 = MM_LOAD(cz+2*SIMDD);
+        __MD cz3 = MM_LOAD(cz+3*SIMDD);
+        __MD r0 = MM_ADD(rx, cx0);
+        __MD r1 = MM_ADD(rx, cx1);
+        __MD r2 = MM_ADD(rx, cx2);
+        __MD r3 = MM_ADD(rx, cx3);
+        __MD r4 = MM_ADD(ry, cy0);
+        __MD r5 = MM_ADD(ry, cy1);
+        __MD r6 = MM_ADD(ry, cy2);
+        __MD r7 = MM_ADD(ry, cy3);
+        __MD r8 = MM_ADD(rz, cz0);
+        __MD r9 = MM_ADD(rz, cz1);
+        __MD r10= MM_ADD(rz, cz2);
+        __MD r11= MM_ADD(rz, cz3);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+8 *SIMDD, cx0);
+        MM_STORE(g+9 *SIMDD, cx1);
+        MM_STORE(g+10*SIMDD, cx2);
+        MM_STORE(g+11*SIMDD, cx3);
+        MM_STORE(g+28*SIMDD, r4);
+        MM_STORE(g+29*SIMDD, r5);
+        MM_STORE(g+30*SIMDD, r6);
+        MM_STORE(g+31*SIMDD, r7);
+        MM_STORE(g+32*SIMDD, cy0);
+        MM_STORE(g+33*SIMDD, cy1);
+        MM_STORE(g+34*SIMDD, cy2);
+        MM_STORE(g+35*SIMDD, cy3);
+        MM_STORE(g+52*SIMDD, MM_MUL(r8 , g48));
+        MM_STORE(g+53*SIMDD, MM_MUL(r9 , g49));
+        MM_STORE(g+54*SIMDD, MM_MUL(r10, g50));
+        MM_STORE(g+55*SIMDD, MM_MUL(r11, g51));
+        MM_STORE(g+56*SIMDD, MM_MUL(cz0,g48));
+        MM_STORE(g+57*SIMDD, MM_MUL(cz1,g49));
+        MM_STORE(g+58*SIMDD, MM_MUL(cz2,g50));
+        MM_STORE(g+59*SIMDD, MM_MUL(cz3,g51));
+        MM_STORE(g+12*SIMDD, MM_FMA(r0, cx0, b0));
+        MM_STORE(g+13*SIMDD, MM_FMA(r1, cx1, b1));
+        MM_STORE(g+14*SIMDD, MM_FMA(r2, cx2, b2));
+        MM_STORE(g+15*SIMDD, MM_FMA(r3, cx3, b3));
+        MM_STORE(g+36*SIMDD, MM_FMA(r4, cy0, b0));
+        MM_STORE(g+37*SIMDD, MM_FMA(r5, cy1, b1));
+        MM_STORE(g+38*SIMDD, MM_FMA(r6, cy2, b2));
+        MM_STORE(g+39*SIMDD, MM_FMA(r7, cy3, b3));
+        MM_STORE(g+60*SIMDD, MM_MUL(MM_FMA(r8 , cz0, b0), g48));
+        MM_STORE(g+61*SIMDD, MM_MUL(MM_FMA(r9 , cz1, b1), g49));
+        MM_STORE(g+62*SIMDD, MM_MUL(MM_FMA(r10, cz2, b2), g50));
+        MM_STORE(g+63*SIMDD, MM_MUL(MM_FMA(r11, cz3, b3), g51));
+}
+
+static inline void _srg0_2d4d_1101(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0  = bc->b00;
+        double *b1  = bc->b10;
+        double *r   = envs->rirj;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0x2 = MM_LOAD(c0x+2*SIMDD);
+        __MD c0x3 = MM_LOAD(c0x+3*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0y2 = MM_LOAD(c0y+2*SIMDD);
+        __MD c0y3 = MM_LOAD(c0y+3*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
+        __MD c0z2 = MM_LOAD(c0z+2*SIMDD);
+        __MD c0z3 = MM_LOAD(c0z+3*SIMDD);
+        __MD rc00 = MM_ADD(rx, c0x0);
+        __MD rc01 = MM_ADD(rx, c0x1);
+        __MD rc02 = MM_ADD(rx, c0x2);
+        __MD rc03 = MM_ADD(rx, c0x3);
+        __MD rc04 = MM_ADD(ry, c0y0);
+        __MD rc05 = MM_ADD(ry, c0y1);
+        __MD rc06 = MM_ADD(ry, c0y2);
+        __MD rc07 = MM_ADD(ry, c0y3);
+        __MD rc08 = MM_ADD(rz, c0z0);
+        __MD rc09 = MM_ADD(rz, c0z1);
+        __MD rc10 = MM_ADD(rz, c0z2);
+        __MD rc11 = MM_ADD(rz, c0z3);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpx2 = MM_LOAD(cpx+2*SIMDD);
+        __MD cpx3 = MM_LOAD(cpx+3*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpy2 = MM_LOAD(cpy+2*SIMDD);
+        __MD cpy3 = MM_LOAD(cpy+3*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD cpz2 = MM_LOAD(cpz+2*SIMDD);
+        __MD cpz3 = MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g96 = MM_LOAD(g+96*SIMDD);
+        __MD g97 = MM_LOAD(g+97*SIMDD);
+        __MD g98 = MM_LOAD(g+98*SIMDD);
+        __MD g99 = MM_LOAD(g+99*SIMDD);
+        MM_STORE(g+4 *SIMDD, rc00);
+        MM_STORE(g+5 *SIMDD, rc01);
+        MM_STORE(g+6 *SIMDD, rc02);
+        MM_STORE(g+7 *SIMDD, rc03);
+        MM_STORE(g+8 *SIMDD, cpx0);
+        MM_STORE(g+9 *SIMDD, cpx1);
+        MM_STORE(g+10*SIMDD, cpx2);
+        MM_STORE(g+11*SIMDD, cpx3);
+        MM_STORE(g+16*SIMDD, c0x0);
+        MM_STORE(g+17*SIMDD, c0x1);
+        MM_STORE(g+18*SIMDD, c0x2);
+        MM_STORE(g+19*SIMDD, c0x3);
+        MM_STORE(g+52*SIMDD, rc04);
+        MM_STORE(g+53*SIMDD, rc05);
+        MM_STORE(g+54*SIMDD, rc06);
+        MM_STORE(g+55*SIMDD, rc07);
+        MM_STORE(g+56*SIMDD, cpy0);
+        MM_STORE(g+57*SIMDD, cpy1);
+        MM_STORE(g+58*SIMDD, cpy2);
+        MM_STORE(g+59*SIMDD, cpy3);
+        MM_STORE(g+64*SIMDD, c0y0);
+        MM_STORE(g+65*SIMDD, c0y1);
+        MM_STORE(g+66*SIMDD, c0y2);
+        MM_STORE(g+67*SIMDD, c0y3);
+        MM_STORE(g+100*SIMDD, rc08 * g96);
+        MM_STORE(g+101*SIMDD, rc09 * g97);
+        MM_STORE(g+102*SIMDD, rc10 * g98);
+        MM_STORE(g+103*SIMDD, rc11 * g99);
+        MM_STORE(g+104*SIMDD, cpz0 * g96);
+        MM_STORE(g+105*SIMDD, cpz1 * g97);
+        MM_STORE(g+106*SIMDD, cpz2 * g98);
+        MM_STORE(g+107*SIMDD, cpz3 * g99);
+        MM_STORE(g+112*SIMDD, c0z0 * g96);
+        MM_STORE(g+113*SIMDD, c0z1 * g97);
+        MM_STORE(g+114*SIMDD, c0z2 * g98);
+        MM_STORE(g+115*SIMDD, c0z3 * g99);
+        MM_STORE(g+12*SIMDD, cpx0 * rc00 + b00);
+        MM_STORE(g+13*SIMDD, cpx1 * rc01 + b01);
+        MM_STORE(g+14*SIMDD, cpx2 * rc02 + b02);
+        MM_STORE(g+15*SIMDD, cpx3 * rc03 + b03);
+        MM_STORE(g+20*SIMDD, c0x0 * rc00 + b10);
+        MM_STORE(g+21*SIMDD, c0x1 * rc01 + b11);
+        MM_STORE(g+22*SIMDD, c0x2 * rc02 + b12);
+        MM_STORE(g+23*SIMDD, c0x3 * rc03 + b13);
+        MM_STORE(g+24*SIMDD, c0x0 * cpx0 + b00);
+        MM_STORE(g+25*SIMDD, c0x1 * cpx1 + b01);
+        MM_STORE(g+26*SIMDD, c0x2 * cpx2 + b02);
+        MM_STORE(g+27*SIMDD, c0x3 * cpx3 + b03);
+        MM_STORE(g+60*SIMDD, cpy0 * rc04 + b00);
+        MM_STORE(g+61*SIMDD, cpy1 * rc05 + b01);
+        MM_STORE(g+62*SIMDD, cpy2 * rc06 + b02);
+        MM_STORE(g+63*SIMDD, cpy3 * rc07 + b03);
+        MM_STORE(g+68*SIMDD, c0y0 * rc04 + b10);
+        MM_STORE(g+69*SIMDD, c0y1 * rc05 + b11);
+        MM_STORE(g+70*SIMDD, c0y2 * rc06 + b12);
+        MM_STORE(g+71*SIMDD, c0y3 * rc07 + b13);
+        MM_STORE(g+72*SIMDD, c0y0 * cpy0 + b00);
+        MM_STORE(g+73*SIMDD, c0y1 * cpy1 + b01);
+        MM_STORE(g+74*SIMDD, c0y2 * cpy2 + b02);
+        MM_STORE(g+75*SIMDD, c0y3 * cpy3 + b03);
+        MM_STORE(g+108*SIMDD,(cpz0 * rc08 + b00)* g96);
+        MM_STORE(g+109*SIMDD,(cpz1 * rc09 + b01)* g97);
+        MM_STORE(g+110*SIMDD,(cpz2 * rc10 + b02)* g98);
+        MM_STORE(g+111*SIMDD,(cpz3 * rc11 + b03)* g99);
+        MM_STORE(g+116*SIMDD,(c0z0 * rc08 + b10)* g96);
+        MM_STORE(g+117*SIMDD,(c0z1 * rc09 + b11)* g97);
+        MM_STORE(g+118*SIMDD,(c0z2 * rc10 + b12)* g98);
+        MM_STORE(g+119*SIMDD,(c0z3 * rc11 + b13)* g99);
+        MM_STORE(g+120*SIMDD,(c0z0 * cpz0 + b00)* g96);
+        MM_STORE(g+121*SIMDD,(c0z1 * cpz1 + b01)* g97);
+        MM_STORE(g+122*SIMDD,(c0z2 * cpz2 + b02)* g98);
+        MM_STORE(g+123*SIMDD,(c0z3 * cpz3 + b03)* g99);
+        MM_STORE(g+28 *SIMDD, cpx0 * MM_LOAD(g+20 *SIMDD) + b00 *(rc00 + c0x0));
+        MM_STORE(g+29 *SIMDD, cpx1 * MM_LOAD(g+21 *SIMDD) + b01 *(rc01 + c0x1));
+        MM_STORE(g+30 *SIMDD, cpx2 * MM_LOAD(g+22 *SIMDD) + b02 *(rc02 + c0x2));
+        MM_STORE(g+31 *SIMDD, cpx3 * MM_LOAD(g+23 *SIMDD) + b03 *(rc03 + c0x3));
+        MM_STORE(g+76 *SIMDD, cpy0 * MM_LOAD(g+68 *SIMDD) + b00 *(rc04 + c0y0));
+        MM_STORE(g+77 *SIMDD, cpy1 * MM_LOAD(g+69 *SIMDD) + b01 *(rc05 + c0y1));
+        MM_STORE(g+78 *SIMDD, cpy2 * MM_LOAD(g+70 *SIMDD) + b02 *(rc06 + c0y2));
+        MM_STORE(g+79 *SIMDD, cpy3 * MM_LOAD(g+71 *SIMDD) + b03 *(rc07 + c0y3));
+        MM_STORE(g+124*SIMDD, cpz0 * MM_LOAD(g+116*SIMDD) + b00 *(MM_LOAD(g+100*SIMDD) + MM_LOAD(g+112*SIMDD)));
+        MM_STORE(g+125*SIMDD, cpz1 * MM_LOAD(g+117*SIMDD) + b01 *(MM_LOAD(g+101*SIMDD) + MM_LOAD(g+113*SIMDD)));
+        MM_STORE(g+126*SIMDD, cpz2 * MM_LOAD(g+118*SIMDD) + b02 *(MM_LOAD(g+102*SIMDD) + MM_LOAD(g+114*SIMDD)));
+        MM_STORE(g+127*SIMDD, cpz3 * MM_LOAD(g+119*SIMDD) + b03 *(MM_LOAD(g+103*SIMDD) + MM_LOAD(g+115*SIMDD)));
+}
+
+static inline void _srg0_2d4d_1110(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0  = bc->b00;
+        double *b1  = bc->b10;
+        double *r   = envs->rirj;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD c0x0 = MM_LOAD(c0x+0*SIMDD);
+        __MD c0x1 = MM_LOAD(c0x+1*SIMDD);
+        __MD c0x2 = MM_LOAD(c0x+2*SIMDD);
+        __MD c0x3 = MM_LOAD(c0x+3*SIMDD);
+        __MD c0y0 = MM_LOAD(c0y+0*SIMDD);
+        __MD c0y1 = MM_LOAD(c0y+1*SIMDD);
+        __MD c0y2 = MM_LOAD(c0y+2*SIMDD);
+        __MD c0y3 = MM_LOAD(c0y+3*SIMDD);
+        __MD c0z0 = MM_LOAD(c0z+0*SIMDD);
+        __MD c0z1 = MM_LOAD(c0z+1*SIMDD);
+        __MD c0z2 = MM_LOAD(c0z+2*SIMDD);
+        __MD c0z3 = MM_LOAD(c0z+3*SIMDD);
+        __MD rc00 = MM_ADD(rx, c0x0);
+        __MD rc01 = MM_ADD(rx, c0x1);
+        __MD rc02 = MM_ADD(rx, c0x2);
+        __MD rc03 = MM_ADD(rx, c0x3);
+        __MD rc04 = MM_ADD(ry, c0y0);
+        __MD rc05 = MM_ADD(ry, c0y1);
+        __MD rc06 = MM_ADD(ry, c0y2);
+        __MD rc07 = MM_ADD(ry, c0y3);
+        __MD rc08 = MM_ADD(rz, c0z0);
+        __MD rc09 = MM_ADD(rz, c0z1);
+        __MD rc10 = MM_ADD(rz, c0z2);
+        __MD rc11 = MM_ADD(rz, c0z3);
+        __MD cpx0 = MM_LOAD(cpx+0*SIMDD);
+        __MD cpx1 = MM_LOAD(cpx+1*SIMDD);
+        __MD cpx2 = MM_LOAD(cpx+2*SIMDD);
+        __MD cpx3 = MM_LOAD(cpx+3*SIMDD);
+        __MD cpy0 = MM_LOAD(cpy+0*SIMDD);
+        __MD cpy1 = MM_LOAD(cpy+1*SIMDD);
+        __MD cpy2 = MM_LOAD(cpy+2*SIMDD);
+        __MD cpy3 = MM_LOAD(cpy+3*SIMDD);
+        __MD cpz0 = MM_LOAD(cpz+0*SIMDD);
+        __MD cpz1 = MM_LOAD(cpz+1*SIMDD);
+        __MD cpz2 = MM_LOAD(cpz+2*SIMDD);
+        __MD cpz3 = MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g96 = MM_LOAD(g+96*SIMDD);
+        __MD g97 = MM_LOAD(g+97*SIMDD);
+        __MD g98 = MM_LOAD(g+98*SIMDD);
+        __MD g99 = MM_LOAD(g+99*SIMDD);
+        MM_STORE(g+4 *SIMDD, rc00);
+        MM_STORE(g+5 *SIMDD, rc01);
+        MM_STORE(g+6 *SIMDD, rc02);
+        MM_STORE(g+7 *SIMDD, rc03);
+        MM_STORE(g+8 *SIMDD, cpx0);
+        MM_STORE(g+9 *SIMDD, cpx1);
+        MM_STORE(g+10*SIMDD, cpx2);
+        MM_STORE(g+11*SIMDD, cpx3);
+        MM_STORE(g+16*SIMDD, c0x0);
+        MM_STORE(g+17*SIMDD, c0x1);
+        MM_STORE(g+18*SIMDD, c0x2);
+        MM_STORE(g+19*SIMDD, c0x3);
+        MM_STORE(g+52*SIMDD, rc04);
+        MM_STORE(g+53*SIMDD, rc05);
+        MM_STORE(g+54*SIMDD, rc06);
+        MM_STORE(g+55*SIMDD, rc07);
+        MM_STORE(g+56*SIMDD, cpy0);
+        MM_STORE(g+57*SIMDD, cpy1);
+        MM_STORE(g+58*SIMDD, cpy2);
+        MM_STORE(g+59*SIMDD, cpy3);
+        MM_STORE(g+64*SIMDD, c0y0);
+        MM_STORE(g+65*SIMDD, c0y1);
+        MM_STORE(g+66*SIMDD, c0y2);
+        MM_STORE(g+67*SIMDD, c0y3);
+        MM_STORE(g+100*SIMDD, rc08 * g96);
+        MM_STORE(g+101*SIMDD, rc09 * g97);
+        MM_STORE(g+102*SIMDD, rc10 * g98);
+        MM_STORE(g+103*SIMDD, rc11 * g99);
+        MM_STORE(g+104*SIMDD, cpz0 * g96);
+        MM_STORE(g+105*SIMDD, cpz1 * g97);
+        MM_STORE(g+106*SIMDD, cpz2 * g98);
+        MM_STORE(g+107*SIMDD, cpz3 * g99);
+        MM_STORE(g+112*SIMDD, c0z0 * g96);
+        MM_STORE(g+113*SIMDD, c0z1 * g97);
+        MM_STORE(g+114*SIMDD, c0z2 * g98);
+        MM_STORE(g+115*SIMDD, c0z3 * g99);
+        MM_STORE(g+12*SIMDD, cpx0 * rc00 + b00);
+        MM_STORE(g+13*SIMDD, cpx1 * rc01 + b01);
+        MM_STORE(g+14*SIMDD, cpx2 * rc02 + b02);
+        MM_STORE(g+15*SIMDD, cpx3 * rc03 + b03);
+        MM_STORE(g+20*SIMDD, c0x0 * rc00 + b10);
+        MM_STORE(g+21*SIMDD, c0x1 * rc01 + b11);
+        MM_STORE(g+22*SIMDD, c0x2 * rc02 + b12);
+        MM_STORE(g+23*SIMDD, c0x3 * rc03 + b13);
+        MM_STORE(g+24*SIMDD, c0x0 * cpx0 + b00);
+        MM_STORE(g+25*SIMDD, c0x1 * cpx1 + b01);
+        MM_STORE(g+26*SIMDD, c0x2 * cpx2 + b02);
+        MM_STORE(g+27*SIMDD, c0x3 * cpx3 + b03);
+        MM_STORE(g+60*SIMDD, cpy0 * rc04 + b00);
+        MM_STORE(g+61*SIMDD, cpy1 * rc05 + b01);
+        MM_STORE(g+62*SIMDD, cpy2 * rc06 + b02);
+        MM_STORE(g+63*SIMDD, cpy3 * rc07 + b03);
+        MM_STORE(g+68*SIMDD, c0y0 * rc04 + b10);
+        MM_STORE(g+69*SIMDD, c0y1 * rc05 + b11);
+        MM_STORE(g+70*SIMDD, c0y2 * rc06 + b12);
+        MM_STORE(g+71*SIMDD, c0y3 * rc07 + b13);
+        MM_STORE(g+72*SIMDD, c0y0 * cpy0 + b00);
+        MM_STORE(g+73*SIMDD, c0y1 * cpy1 + b01);
+        MM_STORE(g+74*SIMDD, c0y2 * cpy2 + b02);
+        MM_STORE(g+75*SIMDD, c0y3 * cpy3 + b03);
+        MM_STORE(g+108*SIMDD,(cpz0 * rc08 + b00)* g96);
+        MM_STORE(g+109*SIMDD,(cpz1 * rc09 + b01)* g97);
+        MM_STORE(g+110*SIMDD,(cpz2 * rc10 + b02)* g98);
+        MM_STORE(g+111*SIMDD,(cpz3 * rc11 + b03)* g99);
+        MM_STORE(g+116*SIMDD,(c0z0 * rc08 + b10)* g96);
+        MM_STORE(g+117*SIMDD,(c0z1 * rc09 + b11)* g97);
+        MM_STORE(g+118*SIMDD,(c0z2 * rc10 + b12)* g98);
+        MM_STORE(g+119*SIMDD,(c0z3 * rc11 + b13)* g99);
+        MM_STORE(g+120*SIMDD,(c0z0 * cpz0 + b00)* g96);
+        MM_STORE(g+121*SIMDD,(c0z1 * cpz1 + b01)* g97);
+        MM_STORE(g+122*SIMDD,(c0z2 * cpz2 + b02)* g98);
+        MM_STORE(g+123*SIMDD,(c0z3 * cpz3 + b03)* g99);
+        MM_STORE(g+28 *SIMDD, cpx0 * MM_LOAD(g+20 *SIMDD) + b00 *(rc00 + c0x0));
+        MM_STORE(g+29 *SIMDD, cpx1 * MM_LOAD(g+21 *SIMDD) + b01 *(rc01 + c0x1));
+        MM_STORE(g+30 *SIMDD, cpx2 * MM_LOAD(g+22 *SIMDD) + b02 *(rc02 + c0x2));
+        MM_STORE(g+31 *SIMDD, cpx3 * MM_LOAD(g+23 *SIMDD) + b03 *(rc03 + c0x3));
+        MM_STORE(g+76 *SIMDD, cpy0 * MM_LOAD(g+68 *SIMDD) + b00 *(rc04 + c0y0));
+        MM_STORE(g+77 *SIMDD, cpy1 * MM_LOAD(g+69 *SIMDD) + b01 *(rc05 + c0y1));
+        MM_STORE(g+78 *SIMDD, cpy2 * MM_LOAD(g+70 *SIMDD) + b02 *(rc06 + c0y2));
+        MM_STORE(g+79 *SIMDD, cpy3 * MM_LOAD(g+71 *SIMDD) + b03 *(rc07 + c0y3));
+        MM_STORE(g+124*SIMDD, cpz0 * MM_LOAD(g+116*SIMDD) + b00 *(MM_LOAD(g+100*SIMDD) + MM_LOAD(g+112*SIMDD)));
+        MM_STORE(g+125*SIMDD, cpz1 * MM_LOAD(g+117*SIMDD) + b01 *(MM_LOAD(g+101*SIMDD) + MM_LOAD(g+113*SIMDD)));
+        MM_STORE(g+126*SIMDD, cpz2 * MM_LOAD(g+118*SIMDD) + b02 *(MM_LOAD(g+102*SIMDD) + MM_LOAD(g+114*SIMDD)));
+        MM_STORE(g+127*SIMDD, cpz3 * MM_LOAD(g+119*SIMDD) + b03 *(MM_LOAD(g+103*SIMDD) + MM_LOAD(g+115*SIMDD)));
+}
+
+static inline void _srg0_2d4d_1200(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c00x;
+        double *cy = bc->c00y;
+        double *cz = bc->c00z;
+        double *b  = bc->b10;
+        double *r  = envs->rirj;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD cx0 = MM_LOAD(cx+0*SIMDD);
+        __MD cx1 = MM_LOAD(cx+1*SIMDD);
+        __MD cx2 = MM_LOAD(cx+2*SIMDD);
+        __MD cx3 = MM_LOAD(cx+3*SIMDD);
+        __MD cy0 = MM_LOAD(cy+0*SIMDD);
+        __MD cy1 = MM_LOAD(cy+1*SIMDD);
+        __MD cy2 = MM_LOAD(cy+2*SIMDD);
+        __MD cy3 = MM_LOAD(cy+3*SIMDD);
+        __MD cz0 = MM_LOAD(cz+0*SIMDD);
+        __MD cz1 = MM_LOAD(cz+1*SIMDD);
+        __MD cz2 = MM_LOAD(cz+2*SIMDD);
+        __MD cz3 = MM_LOAD(cz+3*SIMDD);
+        __MD r0 = MM_ADD(rx, cx0);
+        __MD r1 = MM_ADD(rx, cx1);
+        __MD r2 = MM_ADD(rx, cx2);
+        __MD r3 = MM_ADD(rx, cx3);
+        __MD r4 = MM_ADD(ry, cy0);
+        __MD r5 = MM_ADD(ry, cy1);
+        __MD r6 = MM_ADD(ry, cy2);
+        __MD r7 = MM_ADD(ry, cy3);
+        __MD r8 = MM_ADD(rz, cz0);
+        __MD r9 = MM_ADD(rz, cz1);
+        __MD r10= MM_ADD(rz, cz2);
+        __MD r11= MM_ADD(rz, cz3);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD i2 = MM_SET1(2.);
+        __MD g64 = MM_LOAD(g+64*SIMDD);
+        __MD g65 = MM_LOAD(g+65*SIMDD);
+        __MD g66 = MM_LOAD(g+66*SIMDD);
+        __MD g67 = MM_LOAD(g+67*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0 );
+        MM_STORE(g+5 *SIMDD, r1 );
+        MM_STORE(g+6 *SIMDD, r2 );
+        MM_STORE(g+7 *SIMDD, r3 );
+        MM_STORE(g+8 *SIMDD, cx0);
+        MM_STORE(g+9 *SIMDD, cx1);
+        MM_STORE(g+10*SIMDD, cx2);
+        MM_STORE(g+11*SIMDD, cx3);
+        MM_STORE(g+36*SIMDD, r4 );
+        MM_STORE(g+37*SIMDD, r5 );
+        MM_STORE(g+38*SIMDD, r6 );
+        MM_STORE(g+39*SIMDD, r7 );
+        MM_STORE(g+40*SIMDD, cy0);
+        MM_STORE(g+41*SIMDD, cy1);
+        MM_STORE(g+42*SIMDD, cy2);
+        MM_STORE(g+43*SIMDD, cy3);
+        MM_STORE(g+68*SIMDD, MM_MUL(r8 , g64));
+        MM_STORE(g+69*SIMDD, MM_MUL(r9 , g65));
+        MM_STORE(g+70*SIMDD, MM_MUL(r10, g66));
+        MM_STORE(g+71*SIMDD, MM_MUL(r11, g67));
+        MM_STORE(g+72*SIMDD, MM_MUL(cz0, g64));
+        MM_STORE(g+73*SIMDD, MM_MUL(cz1, g65));
+        MM_STORE(g+74*SIMDD, MM_MUL(cz2, g66));
+        MM_STORE(g+75*SIMDD, MM_MUL(cz3, g67));
+        MM_STORE(g+12*SIMDD, MM_FMA(r0 , cx0, b0));
+        MM_STORE(g+13*SIMDD, MM_FMA(r1 , cx1, b1));
+        MM_STORE(g+14*SIMDD, MM_FMA(r2 , cx2, b2));
+        MM_STORE(g+15*SIMDD, MM_FMA(r3 , cx3, b3));
+        MM_STORE(g+16*SIMDD, MM_FMA(cx0, cx0, b0));
+        MM_STORE(g+17*SIMDD, MM_FMA(cx1, cx1, b1));
+        MM_STORE(g+18*SIMDD, MM_FMA(cx2, cx2, b2));
+        MM_STORE(g+19*SIMDD, MM_FMA(cx3, cx3, b3));
+        MM_STORE(g+44*SIMDD, MM_FMA(r4 , cy0, b0));
+        MM_STORE(g+45*SIMDD, MM_FMA(r5 , cy1, b1));
+        MM_STORE(g+46*SIMDD, MM_FMA(r6 , cy2, b2));
+        MM_STORE(g+47*SIMDD, MM_FMA(r7 , cy3, b3));
+        MM_STORE(g+48*SIMDD, MM_FMA(cy0, cy0, b0));
+        MM_STORE(g+49*SIMDD, MM_FMA(cy1, cy1, b1));
+        MM_STORE(g+50*SIMDD, MM_FMA(cy2, cy2, b2));
+        MM_STORE(g+51*SIMDD, MM_FMA(cy3, cy3, b3));
+        MM_STORE(g+76*SIMDD, MM_MUL(MM_FMA(r8 , cz0, b0), g64));
+        MM_STORE(g+77*SIMDD, MM_MUL(MM_FMA(r9 , cz1, b1), g65));
+        MM_STORE(g+78*SIMDD, MM_MUL(MM_FMA(r10, cz2, b2), g66));
+        MM_STORE(g+79*SIMDD, MM_MUL(MM_FMA(r11, cz3, b3), g67));
+        MM_STORE(g+80*SIMDD, MM_MUL(MM_FMA(cz0, cz0, b0), g64));
+        MM_STORE(g+81*SIMDD, MM_MUL(MM_FMA(cz1, cz1, b1), g65));
+        MM_STORE(g+82*SIMDD, MM_MUL(MM_FMA(cz2, cz2, b2), g66));
+        MM_STORE(g+83*SIMDD, MM_MUL(MM_FMA(cz3, cz3, b3), g67));
+        MM_STORE(g+20*SIMDD, r0 * MM_LOAD(g+16*SIMDD) + i2 * b0 * cx0);
+        MM_STORE(g+21*SIMDD, r1 * MM_LOAD(g+17*SIMDD) + i2 * b1 * cx1);
+        MM_STORE(g+22*SIMDD, r2 * MM_LOAD(g+18*SIMDD) + i2 * b2 * cx2);
+        MM_STORE(g+23*SIMDD, r3 * MM_LOAD(g+19*SIMDD) + i2 * b3 * cx3);
+        MM_STORE(g+52*SIMDD, r4 * MM_LOAD(g+48*SIMDD) + i2 * b0 * cy0);
+        MM_STORE(g+53*SIMDD, r5 * MM_LOAD(g+49*SIMDD) + i2 * b1 * cy1);
+        MM_STORE(g+54*SIMDD, r6 * MM_LOAD(g+50*SIMDD) + i2 * b2 * cy2);
+        MM_STORE(g+55*SIMDD, r7 * MM_LOAD(g+51*SIMDD) + i2 * b3 * cy3);
+        MM_STORE(g+84*SIMDD, r8 * MM_LOAD(g+80*SIMDD) + i2 * b0 * MM_LOAD(g+72*SIMDD));
+        MM_STORE(g+85*SIMDD, r9 * MM_LOAD(g+81*SIMDD) + i2 * b1 * MM_LOAD(g+73*SIMDD));
+        MM_STORE(g+86*SIMDD, r10* MM_LOAD(g+82*SIMDD) + i2 * b2 * MM_LOAD(g+74*SIMDD));
+        MM_STORE(g+87*SIMDD, r11* MM_LOAD(g+83*SIMDD) + i2 * b3 * MM_LOAD(g+75*SIMDD));
+}
+
+static inline void _srg0_2d4d_2000(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c00x;
+        double *cy = bc->c00y;
+        double *cz = bc->c00z;
+        double *b  = bc->b10;
+        __MD r0 = MM_LOAD(cx+0*SIMDD);
+        __MD r1 = MM_LOAD(cx+1*SIMDD);
+        __MD r2 = MM_LOAD(cx+2*SIMDD);
+        __MD r3 = MM_LOAD(cx+3*SIMDD);
+        __MD r4 = MM_LOAD(cy+0*SIMDD);
+        __MD r5 = MM_LOAD(cy+1*SIMDD);
+        __MD r6 = MM_LOAD(cy+2*SIMDD);
+        __MD r7 = MM_LOAD(cy+3*SIMDD);
+        __MD r8 = MM_LOAD(cz+0*SIMDD);
+        __MD r9 = MM_LOAD(cz+1*SIMDD);
+        __MD r10 = MM_LOAD(cz+2*SIMDD);
+        __MD r11 = MM_LOAD(cz+3*SIMDD);
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD g24 = MM_LOAD(g+24*SIMDD);
+        __MD g25 = MM_LOAD(g+25*SIMDD);
+        __MD g26 = MM_LOAD(g+26*SIMDD);
+        __MD g27 = MM_LOAD(g+27*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+16*SIMDD, r4);
+        MM_STORE(g+17*SIMDD, r5);
+        MM_STORE(g+18*SIMDD, r6);
+        MM_STORE(g+19*SIMDD, r7);
+        MM_STORE(g+28*SIMDD, MM_MUL(r8, g24));
+        MM_STORE(g+29*SIMDD, MM_MUL(r9, g25));
+        MM_STORE(g+30*SIMDD, MM_MUL(r10,g26));
+        MM_STORE(g+31*SIMDD, MM_MUL(r11,g27));
+        MM_STORE(g+8 *SIMDD, MM_FMA(r0, r0, b0));
+        MM_STORE(g+9 *SIMDD, MM_FMA(r1, r1, b1));
+        MM_STORE(g+10*SIMDD, MM_FMA(r2, r2, b2));
+        MM_STORE(g+11*SIMDD, MM_FMA(r3, r3, b3));
+        MM_STORE(g+20*SIMDD, MM_FMA(r4, r4, b0));
+        MM_STORE(g+21*SIMDD, MM_FMA(r5, r5, b1));
+        MM_STORE(g+22*SIMDD, MM_FMA(r6, r6, b2));
+        MM_STORE(g+23*SIMDD, MM_FMA(r7, r7, b3));
+        MM_STORE(g+32*SIMDD, MM_MUL(MM_FMA(r8 , r8 , b0), g24));
+        MM_STORE(g+33*SIMDD, MM_MUL(MM_FMA(r9 , r9 , b1), g25));
+        MM_STORE(g+34*SIMDD, MM_MUL(MM_FMA(r10, r10, b2), g26));
+        MM_STORE(g+35*SIMDD, MM_MUL(MM_FMA(r11, r11, b3), g27));
+}
+
+static inline void _srg0_2d4d_2001(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0 = bc->b00;
+        double *b1 = bc->b10;
+        __MD r0 = MM_LOAD(c0x+0*SIMDD);
+        __MD r1 = MM_LOAD(c0x+1*SIMDD);
+        __MD r2 = MM_LOAD(c0x+2*SIMDD);
+        __MD r3 = MM_LOAD(c0x+3*SIMDD);
+        __MD r4 = MM_LOAD(c0y+0*SIMDD);
+        __MD r5 = MM_LOAD(c0y+1*SIMDD);
+        __MD r6 = MM_LOAD(c0y+2*SIMDD);
+        __MD r7 = MM_LOAD(c0y+3*SIMDD);
+        __MD r8 = MM_LOAD(c0z+0*SIMDD);
+        __MD r9 = MM_LOAD(c0z+1*SIMDD);
+        __MD r10= MM_LOAD(c0z+2*SIMDD);
+        __MD r11= MM_LOAD(c0z+3*SIMDD);
+        __MD s0 = MM_LOAD(cpx+0*SIMDD);
+        __MD s1 = MM_LOAD(cpx+1*SIMDD);
+        __MD s2 = MM_LOAD(cpx+2*SIMDD);
+        __MD s3 = MM_LOAD(cpx+3*SIMDD);
+        __MD s4 = MM_LOAD(cpy+0*SIMDD);
+        __MD s5 = MM_LOAD(cpy+1*SIMDD);
+        __MD s6 = MM_LOAD(cpy+2*SIMDD);
+        __MD s7 = MM_LOAD(cpy+3*SIMDD);
+        __MD s8 = MM_LOAD(cpz+0*SIMDD);
+        __MD s9 = MM_LOAD(cpz+1*SIMDD);
+        __MD s10= MM_LOAD(cpz+2*SIMDD);
+        __MD s11= MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+12*SIMDD, s0);
+        MM_STORE(g+13*SIMDD, s1);
+        MM_STORE(g+14*SIMDD, s2);
+        MM_STORE(g+15*SIMDD, s3);
+        MM_STORE(g+28*SIMDD, r4);
+        MM_STORE(g+29*SIMDD, r5);
+        MM_STORE(g+30*SIMDD, r6);
+        MM_STORE(g+31*SIMDD, r7);
+        MM_STORE(g+36*SIMDD, s4);
+        MM_STORE(g+37*SIMDD, s5);
+        MM_STORE(g+38*SIMDD, s6);
+        MM_STORE(g+39*SIMDD, s7);
+        MM_STORE(g+52*SIMDD, MM_MUL(r8 , g48));
+        MM_STORE(g+53*SIMDD, MM_MUL(r9 , g49));
+        MM_STORE(g+54*SIMDD, MM_MUL(r10, g50));
+        MM_STORE(g+55*SIMDD, MM_MUL(r11, g51));
+        MM_STORE(g+60*SIMDD, MM_MUL(s8 , g48));
+        MM_STORE(g+61*SIMDD, MM_MUL(s9 , g49));
+        MM_STORE(g+62*SIMDD, MM_MUL(s10, g50));
+        MM_STORE(g+63*SIMDD, MM_MUL(s11, g51));
+        MM_STORE(g+8 *SIMDD, MM_FMA(r0, r0, b10));
+        MM_STORE(g+9 *SIMDD, MM_FMA(r1, r1, b11));
+        MM_STORE(g+10*SIMDD, MM_FMA(r2, r2, b12));
+        MM_STORE(g+11*SIMDD, MM_FMA(r3, r3, b13));
+        MM_STORE(g+16*SIMDD, MM_FMA(s0, r0, b00));
+        MM_STORE(g+17*SIMDD, MM_FMA(s1, r1, b01));
+        MM_STORE(g+18*SIMDD, MM_FMA(s2, r2, b02));
+        MM_STORE(g+19*SIMDD, MM_FMA(s3, r3, b03));
+        MM_STORE(g+32*SIMDD, MM_FMA(r4, r4, b10));
+        MM_STORE(g+33*SIMDD, MM_FMA(r5, r5, b11));
+        MM_STORE(g+34*SIMDD, MM_FMA(r6, r6, b12));
+        MM_STORE(g+35*SIMDD, MM_FMA(r7, r7, b13));
+        MM_STORE(g+40*SIMDD, MM_FMA(s4, r4, b00));
+        MM_STORE(g+41*SIMDD, MM_FMA(s5, r5, b01));
+        MM_STORE(g+42*SIMDD, MM_FMA(s6, r6, b02));
+        MM_STORE(g+43*SIMDD, MM_FMA(s7, r7, b03));
+        MM_STORE(g+56*SIMDD, MM_MUL(MM_FMA(r8 , r8 , b10), g48));
+        MM_STORE(g+57*SIMDD, MM_MUL(MM_FMA(r9 , r9 , b11), g49));
+        MM_STORE(g+58*SIMDD, MM_MUL(MM_FMA(r10, r10, b12), g50));
+        MM_STORE(g+59*SIMDD, MM_MUL(MM_FMA(r11, r11, b13), g51));
+        MM_STORE(g+64*SIMDD, MM_MUL(MM_FMA(s8 , r8 , b00), g48));
+        MM_STORE(g+65*SIMDD, MM_MUL(MM_FMA(s9 , r9 , b01), g49));
+        MM_STORE(g+66*SIMDD, MM_MUL(MM_FMA(s10, r10, b02), g50));
+        MM_STORE(g+67*SIMDD, MM_MUL(MM_FMA(s11, r11, b03), g51));
+        MM_STORE(g+20*SIMDD, r0 *(MM_LOAD(g+16*SIMDD) + b00) + b10 * s0);
+        MM_STORE(g+21*SIMDD, r1 *(MM_LOAD(g+17*SIMDD) + b01) + b11 * s1);
+        MM_STORE(g+22*SIMDD, r2 *(MM_LOAD(g+18*SIMDD) + b02) + b12 * s2);
+        MM_STORE(g+23*SIMDD, r3 *(MM_LOAD(g+19*SIMDD) + b03) + b13 * s3);
+        MM_STORE(g+44*SIMDD, r4 *(MM_LOAD(g+40*SIMDD) + b00) + b10 * s4);
+        MM_STORE(g+45*SIMDD, r5 *(MM_LOAD(g+41*SIMDD) + b01) + b11 * s5);
+        MM_STORE(g+46*SIMDD, r6 *(MM_LOAD(g+42*SIMDD) + b02) + b12 * s6);
+        MM_STORE(g+47*SIMDD, r7 *(MM_LOAD(g+43*SIMDD) + b03) + b13 * s7);
+        MM_STORE(g+68*SIMDD, r8 * MM_LOAD(g+64*SIMDD) + b00 * MM_LOAD(g+52*SIMDD) + b10 * MM_LOAD(g+60*SIMDD));
+        MM_STORE(g+69*SIMDD, r9 * MM_LOAD(g+65*SIMDD) + b01 * MM_LOAD(g+53*SIMDD) + b11 * MM_LOAD(g+61*SIMDD));
+        MM_STORE(g+70*SIMDD, r10* MM_LOAD(g+66*SIMDD) + b02 * MM_LOAD(g+54*SIMDD) + b12 * MM_LOAD(g+62*SIMDD));
+        MM_STORE(g+71*SIMDD, r11* MM_LOAD(g+67*SIMDD) + b03 * MM_LOAD(g+55*SIMDD) + b13 * MM_LOAD(g+63*SIMDD));
+}
+
+static inline void _srg0_2d4d_2010(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b0 = bc->b00;
+        double *b1 = bc->b10;
+        __MD r0 = MM_LOAD(c0x+0*SIMDD);
+        __MD r1 = MM_LOAD(c0x+1*SIMDD);
+        __MD r2 = MM_LOAD(c0x+2*SIMDD);
+        __MD r3 = MM_LOAD(c0x+3*SIMDD);
+        __MD r4 = MM_LOAD(c0y+0*SIMDD);
+        __MD r5 = MM_LOAD(c0y+1*SIMDD);
+        __MD r6 = MM_LOAD(c0y+2*SIMDD);
+        __MD r7 = MM_LOAD(c0y+3*SIMDD);
+        __MD r8 = MM_LOAD(c0z+0*SIMDD);
+        __MD r9 = MM_LOAD(c0z+1*SIMDD);
+        __MD r10= MM_LOAD(c0z+2*SIMDD);
+        __MD r11= MM_LOAD(c0z+3*SIMDD);
+        __MD s0 = MM_LOAD(cpx+0*SIMDD);
+        __MD s1 = MM_LOAD(cpx+1*SIMDD);
+        __MD s2 = MM_LOAD(cpx+2*SIMDD);
+        __MD s3 = MM_LOAD(cpx+3*SIMDD);
+        __MD s4 = MM_LOAD(cpy+0*SIMDD);
+        __MD s5 = MM_LOAD(cpy+1*SIMDD);
+        __MD s6 = MM_LOAD(cpy+2*SIMDD);
+        __MD s7 = MM_LOAD(cpy+3*SIMDD);
+        __MD s8 = MM_LOAD(cpz+0*SIMDD);
+        __MD s9 = MM_LOAD(cpz+1*SIMDD);
+        __MD s10= MM_LOAD(cpz+2*SIMDD);
+        __MD s11= MM_LOAD(cpz+3*SIMDD);
+        __MD b00 = MM_LOAD(b0+0*SIMDD);
+        __MD b01 = MM_LOAD(b0+1*SIMDD);
+        __MD b02 = MM_LOAD(b0+2*SIMDD);
+        __MD b03 = MM_LOAD(b0+3*SIMDD);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g48 = MM_LOAD(g+48*SIMDD);
+        __MD g49 = MM_LOAD(g+49*SIMDD);
+        __MD g50 = MM_LOAD(g+50*SIMDD);
+        __MD g51 = MM_LOAD(g+51*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+12*SIMDD, s0);
+        MM_STORE(g+13*SIMDD, s1);
+        MM_STORE(g+14*SIMDD, s2);
+        MM_STORE(g+15*SIMDD, s3);
+        MM_STORE(g+28*SIMDD, r4);
+        MM_STORE(g+29*SIMDD, r5);
+        MM_STORE(g+30*SIMDD, r6);
+        MM_STORE(g+31*SIMDD, r7);
+        MM_STORE(g+36*SIMDD, s4);
+        MM_STORE(g+37*SIMDD, s5);
+        MM_STORE(g+38*SIMDD, s6);
+        MM_STORE(g+39*SIMDD, s7);
+        MM_STORE(g+52*SIMDD, MM_MUL(r8 , g48));
+        MM_STORE(g+53*SIMDD, MM_MUL(r9 , g49));
+        MM_STORE(g+54*SIMDD, MM_MUL(r10, g50));
+        MM_STORE(g+55*SIMDD, MM_MUL(r11, g51));
+        MM_STORE(g+60*SIMDD, MM_MUL(s8 , g48));
+        MM_STORE(g+61*SIMDD, MM_MUL(s9 , g49));
+        MM_STORE(g+62*SIMDD, MM_MUL(s10, g50));
+        MM_STORE(g+63*SIMDD, MM_MUL(s11, g51));
+        MM_STORE(g+8 *SIMDD, MM_FMA(r0, r0, b10));
+        MM_STORE(g+9 *SIMDD, MM_FMA(r1, r1, b11));
+        MM_STORE(g+10*SIMDD, MM_FMA(r2, r2, b12));
+        MM_STORE(g+11*SIMDD, MM_FMA(r3, r3, b13));
+        MM_STORE(g+16*SIMDD, MM_FMA(s0, r0, b00));
+        MM_STORE(g+17*SIMDD, MM_FMA(s1, r1, b01));
+        MM_STORE(g+18*SIMDD, MM_FMA(s2, r2, b02));
+        MM_STORE(g+19*SIMDD, MM_FMA(s3, r3, b03));
+        MM_STORE(g+32*SIMDD, MM_FMA(r4, r4, b10));
+        MM_STORE(g+33*SIMDD, MM_FMA(r5, r5, b11));
+        MM_STORE(g+34*SIMDD, MM_FMA(r6, r6, b12));
+        MM_STORE(g+35*SIMDD, MM_FMA(r7, r7, b13));
+        MM_STORE(g+40*SIMDD, MM_FMA(s4, r4, b00));
+        MM_STORE(g+41*SIMDD, MM_FMA(s5, r5, b01));
+        MM_STORE(g+42*SIMDD, MM_FMA(s6, r6, b02));
+        MM_STORE(g+43*SIMDD, MM_FMA(s7, r7, b03));
+        MM_STORE(g+56*SIMDD, MM_MUL(MM_FMA(r8 , r8 , b10), g48));
+        MM_STORE(g+57*SIMDD, MM_MUL(MM_FMA(r9 , r9 , b11), g49));
+        MM_STORE(g+58*SIMDD, MM_MUL(MM_FMA(r10, r10, b12), g50));
+        MM_STORE(g+59*SIMDD, MM_MUL(MM_FMA(r11, r11, b13), g51));
+        MM_STORE(g+64*SIMDD, MM_MUL(MM_FMA(s8 , r8 , b00), g48));
+        MM_STORE(g+65*SIMDD, MM_MUL(MM_FMA(s9 , r9 , b01), g49));
+        MM_STORE(g+66*SIMDD, MM_MUL(MM_FMA(s10, r10, b02), g50));
+        MM_STORE(g+67*SIMDD, MM_MUL(MM_FMA(s11, r11, b03), g51));
+        MM_STORE(g+20*SIMDD, r0 *(MM_LOAD(g+16*SIMDD) + b00) + b10 * s0);
+        MM_STORE(g+21*SIMDD, r1 *(MM_LOAD(g+17*SIMDD) + b01) + b11 * s1);
+        MM_STORE(g+22*SIMDD, r2 *(MM_LOAD(g+18*SIMDD) + b02) + b12 * s2);
+        MM_STORE(g+23*SIMDD, r3 *(MM_LOAD(g+19*SIMDD) + b03) + b13 * s3);
+        MM_STORE(g+44*SIMDD, r4 *(MM_LOAD(g+40*SIMDD) + b00) + b10 * s4);
+        MM_STORE(g+45*SIMDD, r5 *(MM_LOAD(g+41*SIMDD) + b01) + b11 * s5);
+        MM_STORE(g+46*SIMDD, r6 *(MM_LOAD(g+42*SIMDD) + b02) + b12 * s6);
+        MM_STORE(g+47*SIMDD, r7 *(MM_LOAD(g+43*SIMDD) + b03) + b13 * s7);
+        MM_STORE(g+68*SIMDD, r8 * MM_LOAD(g+64*SIMDD) + b00 * MM_LOAD(g+52*SIMDD) + b10 * MM_LOAD(g+60*SIMDD));
+        MM_STORE(g+69*SIMDD, r9 * MM_LOAD(g+65*SIMDD) + b01 * MM_LOAD(g+53*SIMDD) + b11 * MM_LOAD(g+61*SIMDD));
+        MM_STORE(g+70*SIMDD, r10* MM_LOAD(g+66*SIMDD) + b02 * MM_LOAD(g+54*SIMDD) + b12 * MM_LOAD(g+62*SIMDD));
+        MM_STORE(g+71*SIMDD, r11* MM_LOAD(g+67*SIMDD) + b03 * MM_LOAD(g+55*SIMDD) + b13 * MM_LOAD(g+63*SIMDD));
+}
+
+static inline void _srg0_2d4d_2100(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c00x;
+        double *cy = bc->c00y;
+        double *cz = bc->c00z;
+        double *b1 = bc->b10;
+        double *r  = envs->rirj;
+        __MD rx = MM_SET1(r[0]);
+        __MD ry = MM_SET1(r[1]);
+        __MD rz = MM_SET1(r[2]);
+        __MD s0 = MM_LOAD(cx+0*SIMDD);
+        __MD s1 = MM_LOAD(cx+1*SIMDD);
+        __MD s2 = MM_LOAD(cx+2*SIMDD);
+        __MD s3 = MM_LOAD(cx+3*SIMDD);
+        __MD s4 = MM_LOAD(cy+0*SIMDD);
+        __MD s5 = MM_LOAD(cy+1*SIMDD);
+        __MD s6 = MM_LOAD(cy+2*SIMDD);
+        __MD s7 = MM_LOAD(cy+3*SIMDD);
+        __MD s8 = MM_LOAD(cz+0*SIMDD);
+        __MD s9 = MM_LOAD(cz+1*SIMDD);
+        __MD s10= MM_LOAD(cz+2*SIMDD);
+        __MD s11= MM_LOAD(cz+3*SIMDD);
+        __MD r0  = MM_ADD(rx, s0 );
+        __MD r1  = MM_ADD(rx, s1 );
+        __MD r2  = MM_ADD(rx, s2 );
+        __MD r3  = MM_ADD(rx, s3 );
+        __MD r4  = MM_ADD(ry, s4 );
+        __MD r5  = MM_ADD(ry, s5 );
+        __MD r6  = MM_ADD(ry, s6 );
+        __MD r7  = MM_ADD(ry, s7 );
+        __MD r8  = MM_ADD(rz, s8 );
+        __MD r9  = MM_ADD(rz, s9 );
+        __MD r10 = MM_ADD(rz, s10);
+        __MD r11 = MM_ADD(rz, s11);
+        __MD b10 = MM_LOAD(b1+0*SIMDD);
+        __MD b11 = MM_LOAD(b1+1*SIMDD);
+        __MD b12 = MM_LOAD(b1+2*SIMDD);
+        __MD b13 = MM_LOAD(b1+3*SIMDD);
+        __MD g64 = MM_LOAD(g+64*SIMDD);
+        __MD g65 = MM_LOAD(g+65*SIMDD);
+        __MD g66 = MM_LOAD(g+66*SIMDD);
+        __MD g67 = MM_LOAD(g+67*SIMDD);
+        __MD i2 = MM_SET1(2.);
+        MM_STORE(g+4 *SIMDD, s0);
+        MM_STORE(g+5 *SIMDD, s1);
+        MM_STORE(g+6 *SIMDD, s2);
+        MM_STORE(g+7 *SIMDD, s3);
+        MM_STORE(g+16*SIMDD, r0);
+        MM_STORE(g+17*SIMDD, r1);
+        MM_STORE(g+18*SIMDD, r2);
+        MM_STORE(g+19*SIMDD, r3);
+        MM_STORE(g+36*SIMDD, s4);
+        MM_STORE(g+37*SIMDD, s5);
+        MM_STORE(g+38*SIMDD, s6);
+        MM_STORE(g+39*SIMDD, s7);
+        MM_STORE(g+48*SIMDD, r4);
+        MM_STORE(g+49*SIMDD, r5);
+        MM_STORE(g+50*SIMDD, r6);
+        MM_STORE(g+51*SIMDD, r7);
+        MM_STORE(g+68*SIMDD, s8 * g64);
+        MM_STORE(g+69*SIMDD, s9 * g65);
+        MM_STORE(g+70*SIMDD, s10* g66);
+        MM_STORE(g+71*SIMDD, s11* g67);
+        MM_STORE(g+80*SIMDD, r8 * g64);
+        MM_STORE(g+81*SIMDD, r9 * g65);
+        MM_STORE(g+82*SIMDD, r10* g66);
+        MM_STORE(g+83*SIMDD, r11* g67);
+        MM_STORE(g+8 *SIMDD, s0 * s0 + b10);
+        MM_STORE(g+9 *SIMDD, s1 * s1 + b11);
+        MM_STORE(g+10*SIMDD, s2 * s2 + b12);
+        MM_STORE(g+11*SIMDD, s3 * s3 + b13);
+        MM_STORE(g+20*SIMDD, s0 * r0 + b10);
+        MM_STORE(g+21*SIMDD, s1 * r1 + b11);
+        MM_STORE(g+22*SIMDD, s2 * r2 + b12);
+        MM_STORE(g+23*SIMDD, s3 * r3 + b13);
+        MM_STORE(g+40*SIMDD, s4 * s4 + b10);
+        MM_STORE(g+41*SIMDD, s5 * s5 + b11);
+        MM_STORE(g+42*SIMDD, s6 * s6 + b12);
+        MM_STORE(g+43*SIMDD, s7 * s7 + b13);
+        MM_STORE(g+52*SIMDD, s4 * r4 + b10);
+        MM_STORE(g+53*SIMDD, s5 * r5 + b11);
+        MM_STORE(g+54*SIMDD, s6 * r6 + b12);
+        MM_STORE(g+55*SIMDD, s7 * r7 + b13);
+        MM_STORE(g+72*SIMDD,(s8 * s8 + b10) * g64);
+        MM_STORE(g+73*SIMDD,(s9 * s9 + b11) * g65);
+        MM_STORE(g+74*SIMDD,(s10* s10+ b12) * g66);
+        MM_STORE(g+75*SIMDD,(s11* s11+ b13) * g67);
+        MM_STORE(g+84*SIMDD,(s8 * r8 + b10) * g64);
+        MM_STORE(g+85*SIMDD,(s9 * r9 + b11) * g65);
+        MM_STORE(g+86*SIMDD,(s10* r10+ b12) * g66);
+        MM_STORE(g+87*SIMDD,(s11* r11+ b13) * g67);
+        MM_STORE(g+24*SIMDD, r0 * MM_LOAD(g+8 *SIMDD) + i2 * b10 * s0);
+        MM_STORE(g+25*SIMDD, r1 * MM_LOAD(g+9 *SIMDD) + i2 * b11 * s1);
+        MM_STORE(g+26*SIMDD, r2 * MM_LOAD(g+10*SIMDD) + i2 * b12 * s2);
+        MM_STORE(g+27*SIMDD, r3 * MM_LOAD(g+11*SIMDD) + i2 * b13 * s3);
+        MM_STORE(g+56*SIMDD, r4 * MM_LOAD(g+40*SIMDD) + i2 * b10 * s4);
+        MM_STORE(g+57*SIMDD, r5 * MM_LOAD(g+41*SIMDD) + i2 * b11 * s5);
+        MM_STORE(g+58*SIMDD, r6 * MM_LOAD(g+42*SIMDD) + i2 * b12 * s6);
+        MM_STORE(g+59*SIMDD, r7 * MM_LOAD(g+43*SIMDD) + i2 * b13 * s7);
+        MM_STORE(g+88*SIMDD, r8 * MM_LOAD(g+72*SIMDD) + i2 * b10 * MM_LOAD(g+68*SIMDD));
+        MM_STORE(g+89*SIMDD, r9 * MM_LOAD(g+73*SIMDD) + i2 * b11 * MM_LOAD(g+69*SIMDD));
+        MM_STORE(g+90*SIMDD, r10* MM_LOAD(g+74*SIMDD) + i2 * b12 * MM_LOAD(g+70*SIMDD));
+        MM_STORE(g+91*SIMDD, r11* MM_LOAD(g+75*SIMDD) + i2 * b13 * MM_LOAD(g+71*SIMDD));
+}
+
+static inline void _srg0_2d4d_3000(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cx = bc->c00x;
+        double *cy = bc->c00y;
+        double *cz = bc->c00z;
+        double *b  = bc->b10;
+        __MD b0 = MM_LOAD(b+0*SIMDD);
+        __MD b1 = MM_LOAD(b+1*SIMDD);
+        __MD b2 = MM_LOAD(b+2*SIMDD);
+        __MD b3 = MM_LOAD(b+3*SIMDD);
+        __MD i3 = MM_SET1(3.);
+        __MD r0 = MM_LOAD(cx+0*SIMDD);
+        __MD r1 = MM_LOAD(cx+1*SIMDD);
+        __MD r2 = MM_LOAD(cx+2*SIMDD);
+        __MD r3 = MM_LOAD(cx+3*SIMDD);
+        __MD r4 = MM_LOAD(cy+0*SIMDD);
+        __MD r5 = MM_LOAD(cy+1*SIMDD);
+        __MD r6 = MM_LOAD(cy+2*SIMDD);
+        __MD r7 = MM_LOAD(cy+3*SIMDD);
+        __MD r8 = MM_LOAD(cz+0*SIMDD);
+        __MD r9 = MM_LOAD(cz+1*SIMDD);
+        __MD r10= MM_LOAD(cz+2*SIMDD);
+        __MD r11= MM_LOAD(cz+3*SIMDD);
+        __MD g32 = MM_LOAD(g+32*SIMDD);
+        __MD g33 = MM_LOAD(g+33*SIMDD);
+        __MD g34 = MM_LOAD(g+34*SIMDD);
+        __MD g35 = MM_LOAD(g+35*SIMDD);
+        MM_STORE(g+4 *SIMDD, r0);
+        MM_STORE(g+5 *SIMDD, r1);
+        MM_STORE(g+6 *SIMDD, r2);
+        MM_STORE(g+7 *SIMDD, r3);
+        MM_STORE(g+20*SIMDD, r4);
+        MM_STORE(g+21*SIMDD, r5);
+        MM_STORE(g+22*SIMDD, r6);
+        MM_STORE(g+23*SIMDD, r7);
+        MM_STORE(g+36*SIMDD, r8 * g32);
+        MM_STORE(g+37*SIMDD, r9 * g33);
+        MM_STORE(g+38*SIMDD, r10* g34);
+        MM_STORE(g+39*SIMDD, r11* g35);
+        MM_STORE(g+8 *SIMDD, r0 * r0 + b0);
+        MM_STORE(g+9 *SIMDD, r1 * r1 + b1);
+        MM_STORE(g+10*SIMDD, r2 * r2 + b2);
+        MM_STORE(g+11*SIMDD, r3 * r3 + b3);
+        MM_STORE(g+24*SIMDD, r4 * r4 + b0);
+        MM_STORE(g+25*SIMDD, r5 * r5 + b1);
+        MM_STORE(g+26*SIMDD, r6 * r6 + b2);
+        MM_STORE(g+27*SIMDD, r7 * r7 + b3);
+        MM_STORE(g+40*SIMDD,(r8 * r8 + b0)* g32);
+        MM_STORE(g+41*SIMDD,(r9 * r9 + b1)* g33);
+        MM_STORE(g+42*SIMDD,(r10* r10+ b2)* g34);
+        MM_STORE(g+43*SIMDD,(r11* r11+ b3)* g35);
+        MM_STORE(g+12*SIMDD, r0 *(r0 * r0 + i3 * b0));
+        MM_STORE(g+13*SIMDD, r1 *(r1 * r1 + i3 * b1));
+        MM_STORE(g+14*SIMDD, r2 *(r2 * r2 + i3 * b2));
+        MM_STORE(g+15*SIMDD, r3 *(r3 * r3 + i3 * b3));
+        MM_STORE(g+28*SIMDD, r4 *(r4 * r4 + i3 * b0));
+        MM_STORE(g+29*SIMDD, r5 *(r5 * r5 + i3 * b1));
+        MM_STORE(g+30*SIMDD, r6 *(r6 * r6 + i3 * b2));
+        MM_STORE(g+31*SIMDD, r7 *(r7 * r7 + i3 * b3));
+        MM_STORE(g+44*SIMDD,(r8 * r8 + i3 * b0) * MM_LOAD(g+36*SIMDD));
+        MM_STORE(g+45*SIMDD,(r9 * r9 + i3 * b1) * MM_LOAD(g+37*SIMDD));
+        MM_STORE(g+46*SIMDD,(r10*r10 + i3 * b2) * MM_LOAD(g+38*SIMDD));
+        MM_STORE(g+47*SIMDD,(r11*r11 + i3 * b3) * MM_LOAD(g+39*SIMDD));
+}
+
+void CINTsrg0_2e_2d4d_unrolled(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        int type_ijkl = ((envs->li_ceil << 6) | (envs->lj_ceil << 4) |
+                         (envs->lk_ceil << 2) | (envs->ll_ceil));
+        switch (type_ijkl) {
+                case 0b00000000: _srg0_2d4d_0000(g, bc, envs); return;
+                case 0b00000001: _srg0_2d4d_0001(g, bc, envs); return;
+                case 0b00000010: _srg0_2d4d_0002(g, bc, envs); return;
+                case 0b00000011: _srg0_2d4d_0003(g, bc, envs); return;
+                case 0b00000100: _srg0_2d4d_0010(g, bc, envs); return;
+                case 0b00000101: _srg0_2d4d_0011(g, bc, envs); return;
+                case 0b00000110: _srg0_2d4d_0012(g, bc, envs); return;
+                case 0b00001000: _srg0_2d4d_0020(g, bc, envs); return;
+                case 0b00001001: _srg0_2d4d_0021(g, bc, envs); return;
+                case 0b00001100: _srg0_2d4d_0030(g, bc, envs); return;
+                case 0b00010000: _srg0_2d4d_0100(g, bc, envs); return;
+                case 0b00010001: _srg0_2d4d_0101(g, bc, envs); return;
+                case 0b00010010: _srg0_2d4d_0102(g, bc, envs); return;
+                case 0b00010100: _srg0_2d4d_0110(g, bc, envs); return;
+                case 0b00010101: _srg0_2d4d_0111(g, bc, envs); return;
+                case 0b00011000: _srg0_2d4d_0120(g, bc, envs); return;
+                case 0b00100000: _srg0_2d4d_0200(g, bc, envs); return;
+                case 0b00100001: _srg0_2d4d_0201(g, bc, envs); return;
+                case 0b00100100: _srg0_2d4d_0210(g, bc, envs); return;
+                case 0b00110000: _srg0_2d4d_0300(g, bc, envs); return;
+                case 0b01000000: _srg0_2d4d_1000(g, bc, envs); return;
+                case 0b01000001: _srg0_2d4d_1001(g, bc, envs); return;
+                case 0b01000010: _srg0_2d4d_1002(g, bc, envs); return;
+                case 0b01000100: _srg0_2d4d_1010(g, bc, envs); return;
+                case 0b01000101: _srg0_2d4d_1011(g, bc, envs); return;
+                case 0b01001000: _srg0_2d4d_1020(g, bc, envs); return;
+                case 0b01010000: _srg0_2d4d_1100(g, bc, envs); return;
+                case 0b01010001: _srg0_2d4d_1101(g, bc, envs); return;
+                case 0b01010100: _srg0_2d4d_1110(g, bc, envs); return;
+                case 0b01100000: _srg0_2d4d_1200(g, bc, envs); return;
+                case 0b10000000: _srg0_2d4d_2000(g, bc, envs); return;
+                case 0b10000001: _srg0_2d4d_2001(g, bc, envs); return;
+                case 0b10000100: _srg0_2d4d_2010(g, bc, envs); return;
+                case 0b10010000: _srg0_2d4d_2100(g, bc, envs); return;
+                case 0b11000000: _srg0_2d4d_3000(g, bc, envs); return;
         }
+        fprintf(stderr, "Dimension error for CINTg0_2e_lj2d4d: iklj = %d %d %d %d",
+               (int)envs->li_ceil, (int)envs->lk_ceil,
+               (int)envs->ll_ceil, (int)envs->lj_ceil);
+}
 
-        r1 = MM_LOAD(a1);
-        r0 = MM_DIV(MM_LOAD(a0), MM_MUL(r1, MM_MUL(r1, r1)));
+void CINTg0_2e_lj2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        CINTg0_2e_2d(g, bc, envs);
+        CINTg0_lj_4d(g, envs);
+}
+void CINTg0_2e_kj2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        CINTg0_2e_2d(g, bc, envs);
+        CINTg0_kj_4d(g, envs);
+}
+void CINTg0_2e_ik2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        CINTg0_2e_2d(g, bc, envs);
+        CINTg0_ik_4d(g, envs);
+}
+void CINTg0_2e_il2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        CINTg0_2e_2d(g, bc, envs);
+        CINTg0_il_4d(g, envs);
+}
+
+/*
+ * g[i,k,l,j] = < ik | lj > = ( i j | k l )
+ */
+int CINTg0_2e(double *g, double *cutoff, Rys2eT *bc, CINTEnvVars *envs, int count)
+{
+        ALIGNMM double a0[SIMDD];
+        ALIGNMM double a1[SIMDD];
+        ALIGNMM double fac1[SIMDD];
+        ALIGNMM double x[SIMDD];
+        ALIGNMM double rijrkl[SIMDD*3];
+        ALIGNMM double rijrx[SIMDD*3];
+        ALIGNMM double rklrx[SIMDD*3];
+        ALIGNMM double u[MXRYSROOTS*SIMDD];
+        DEF_GXYZ(double, g, gx, gy, gz);
+        double *rij = envs->rij;
+        double *rkl = envs->rkl;
+        double *w = gz;
+        __MD ra, r0, r1, r2, r3, r4, r5, r6, r7, r8;
+        int nroots = envs->nrys_roots;
+        int i;
+
+        //:for (int k = 0; k < count; k++) {
+        //:        aij[k] = envs->ai[k] + envs->aj[k];
+        //:        akl[k] = envs->ak[k] + envs->al[k];
+        //:        aijkl[k] = aij[k] + akl[k];
+        //:        a1[k] = aij[k] * akl[k];
+        //:        a0[k] = a1[k] / aijkl[k];
+        //:        //fac1[k] = sqrt(a0[k] / (a1[k] * a1[k] * a1[k])) * envs->fac[k];
+        //:        fac1[k] = envs->fac[k] / (sqrt(aijakl[k]) * a1[k]);
+        //:}
+        __MD aij = MM_ADD(MM_LOAD(envs->ai),  MM_LOAD(envs->aj));
+        __MD akl = MM_ADD(MM_LOAD(envs->ak),  MM_LOAD(envs->al));
+        r1 = MM_MUL(aij, akl);
+        MM_STORE(a1, r1);
+        ra = MM_ADD(aij, akl);
+        r0 = MM_DIV(r1, ra);
+        MM_STORE(a0, r0);
+
+        r0 = MM_DIV(r0, MM_MUL(r1, MM_MUL(r1, r1)));
         MM_STORE(fac1, MM_MUL(MM_SQRT(r0), MM_LOAD(envs->fac)));
-#else
-        MM_STORE(fac1, MM_DIV(MM_LOAD(envs->fac), MM_MUL(MM_SQRT(ra), r1)));
-#endif
 
         r0 = MM_LOAD(rij+0*SIMDD);
         r1 = MM_LOAD(rij+1*SIMDD);
@@ -2475,39 +5473,91 @@ int CINTg0_2e(double *g, double *cutoff, Rys2eT *bc, CINTEnvVars *envs, int coun
         MM_STORE(rklrx+1*SIMDD, MM_SUB(r4, MM_SET1(envs->rx_in_rklrx[1])));
         MM_STORE(rklrx+2*SIMDD, MM_SUB(r5, MM_SET1(envs->rx_in_rklrx[2])));
 
-//ABORT        ALIGNMM double erfx[SIMDD];
-//ABORT        if (envs->g_size == 1) {
-//ABORT                // gz = erf(sqrt(rr*aij*akl/(aij+akl)))/sqrt(rr)
-//ABORT                MM_STORE(erfx, MM_SQRT(MM_LOAD(x)));
-//ABORT                for (k = 0; k < count; k++) {
-//ABORT                        erfx[k] = erf(erfx[k]);
-//ABORT                }
-//ABORT                MM_STORE(gz, MM_DIV(MM_LOAD(erfx), MM_SQRT(MM_LOAD(x))));
-//ABORT                return;
-//ABORT        }
-#ifdef WITH_RANGE_COULOMB
-        if (omega < 0) {
-                int all_negligible = _CINTsr_rys_roots_batch(
-                        envs, x, theta, u, w, cutoff, count);
-                if (all_negligible) {
-                        // g still has to be evaluated since iempty (which
-                        // indicates whether g is zero) in cint2e is determined
-                        // before calling the g0_2e function.
-                        for (i = 0; i < envs->g_size * SIMDD * 3; i++) {
-                                g[i] = 0;
+// Not recommended to mix range-separated Coulomb with regular Coulomb operator.
+// Keep this for backward compatibility to cint2
+        const double omega = envs->env[PTR_RANGE_OMEGA];
+        ALIGNMM double theta[SIMDD];
+        if (omega == 0) {
+                _CINTrys_roots_batch(nroots, x, u, w, count);
+        } else if (omega < 0) {
+                //:theta = omega * omega / (omega * omega + a0);
+                r0 = MM_SET1(omega);
+                r0 = MM_MUL(r0, r0);
+                r1 = MM_LOAD(a0);
+                r0 = MM_DIV(r0, MM_ADD(r0, r1));
+                MM_STORE(theta, r0);
+                int rorder = envs->rys_order;
+                int all_negligible = 1;
+                if (rorder == nroots) {
+                        all_negligible = _CINTsr_rys_roots_batch(
+                                envs, x, theta, u, w, cutoff, count);
+                        if (all_negligible) {
+                                return 0;
+                        }
+                } else {
+                        r0 = MM_SET1(0.);
+                        for (i = 0; i < nroots; i++) {
+                                MM_STORE(u+i*SIMDD, r0);
+                                MM_STORE(w+i*SIMDD, r0);
+                        }
+                        ALIGNMM double xt[SIMDD];
+                        __MD rtheta = MM_LOAD(theta);
+                        __MD rtheta1 = -MM_SQRT(rtheta);
+                        MM_STORE(xt, MM_MUL(rtheta, MM_LOAD(x)));
+                        double rt[MXRYSROOTS * 2];
+                        double *wt = rt + nroots;
+                        int k;
+                        for (i = 0; i < count; i++) {
+                                if (xt[i] < cutoff[i] && xt[i] < EXPCUTOFF_SR) {
+                                        CINTrys_roots(rorder, x[i], rt, wt);
+                                        CINTrys_roots(rorder, xt[i], rt+rorder, wt+rorder);
+                                        for (k = 0; k < nroots; k++) {
+                                                u[k*SIMDD+i] = rt[k];
+                                                w[k*SIMDD+i] = wt[k];
+                                        }
+                                        all_negligible = 0;
+                                }
+                        }
+                        if (all_negligible) {
+                                return 0;
+                        }
+
+                        if (envs->g_size == 2) {
+                                r0 = MM_LOAD(fac1);
+                                r1 = MM_SET1(1.);
+                                MM_STORE(g+0*SIMDD, r1);
+                                MM_STORE(g+1*SIMDD, r1);
+                                MM_STORE(g+2*SIMDD, r1);
+                                MM_STORE(g+3*SIMDD, r1);
+                                MM_STORE(g+4*SIMDD, MM_LOAD(w      ) * r0);
+                                MM_STORE(g+5*SIMDD, MM_LOAD(w+SIMDD) * r0 * rtheta1);
+                                return 1;
+                        }
+                        r1 = MM_SET1(1.);
+                        for (i = rorder; i < nroots; i++) {
+                                r0 = MM_LOAD(u+i*SIMDD);
+                                r2 = r0 * rtheta;
+                                MM_STORE(u+i*SIMDD, MM_DIV(r2, r0+r1-r2));
+                                MM_STORE(w+i*SIMDD, MM_LOAD(w+i*SIMDD) * rtheta1);
                         }
-                        return 0;
                 }
         } else {
+                //:theta = omega * omega / (omega * omega + a0);
+                r0 = MM_SET1(omega);
+                r0 = MM_MUL(r0, r0);
+                r1 = MM_LOAD(a0);
+                __MD rtheta = MM_DIV(r0, MM_ADD(r0, r1));
+                MM_STORE(x, MM_MUL(rtheta, MM_LOAD(x)));
+                MM_STORE(fac1, MM_MUL(MM_LOAD(fac1), MM_SQRT(rtheta)));
                 _CINTrys_roots_batch(nroots, x, u, w, count);
+                r1 = MM_SET1(1.);
+                for (i = 0; i < nroots; i++) {
+                        r0 = MM_LOAD(u+i*SIMDD);
+                        r2 = r0 * rtheta;
+                        MM_STORE(u+i*SIMDD, MM_DIV(r2, r0+r1-r2));
+                }
         }
-#else
-        _CINTrys_roots_batch(nroots, x, u, w, count);
-#endif
 
-        double *gx = g;
-        double *gy = gx + envs->g_size * SIMDD;
-        double *gz = gy + envs->g_size * SIMDD;
         //:for (i = 0; i < nroots; i++) {
         //:for (k = 0; k < count; k++) {
         //:        gx[i*SIMDD+k] = 1;
@@ -2517,33 +5567,14 @@ int CINTg0_2e(double *g, double *cutoff, Rys2eT *bc, CINTEnvVars *envs, int coun
         r0 = MM_LOAD(fac1);
         r1 = MM_SET1(1.);
         for (i = 0; i < nroots; i++) {
-                //MM_STORE(gx+i*SIMDD, r1);
-                //MM_STORE(gy+i*SIMDD, r1);
+                MM_STORE(gx+i*SIMDD, r1);
+                MM_STORE(gy+i*SIMDD, r1);
                 MM_STORE(gz+i*SIMDD, MM_MUL(MM_LOAD(w+i*SIMDD), r0));
         }
         if (envs->g_size == 1) { // ssss
                 return 1;
         }
 
-#ifdef WITH_RANGE_COULOMB
-        if (omega > 0) {
-                /* u[:] = tau^2 / (1 - tau^2)
-                 * transform u[:] to theta^-1 tau^2 / (theta^-1 - tau^2)
-                 * so the rest code can be reused.
-                 */
-                //:for (irys = 0; irys < nroots; irys++) {
-                //:        u[irys] /= u[irys] + 1 - u[irys] * theta;
-                //:}
-                r0 = MM_LOAD(theta);
-                r1 = MM_SET1(1.);
-                for (i = 0; i < nroots; i++) {
-                        r2 = MM_LOAD(u+i*SIMDD);
-                        r3 = r2 + r1 - r2 * r0;
-                        MM_STORE(u+i*SIMDD, MM_DIV(r2, r3));
-                }
-        }
-#endif
-
         double *b00 = bc->b00;
         double *b10 = bc->b10;
         double *b01 = bc->b01;
@@ -2553,9 +5584,6 @@ int CINTg0_2e(double *g, double *cutoff, Rys2eT *bc, CINTEnvVars *envs, int coun
         double *c0px = bc->c0px;
         double *c0py = bc->c0py;
         double *c0pz = bc->c0pz;
-
-        ALIGNMM double tmp1[MXRYSROOTS*SIMDD];
-        ALIGNMM double tmp4[MXRYSROOTS*SIMDD];
         //:double u2[SIMDD];
         //:double tmp2, tmp3;
         //:double div[SIMDD];
@@ -2579,49 +5607,39 @@ int CINTg0_2e(double *g, double *cutoff, Rys2eT *bc, CINTEnvVars *envs, int coun
         //:        c0pz[i*SIMDD+k] = rklrx[2*SIMDD+k] + tmp3 * rijrkl[2*SIMDD+k];
         //:} }
 
-        ra = MM_LOAD(aijkl);
+        ra = MM_ADD(aij, akl);
         r0 = MM_LOAD(a0);
         r1 = MM_LOAD(a1);
         r2 = MM_SET1(.5);
         r3 = MM_SET1(1.);
-        for (i = 0; i < nroots; i++) {
-                r4 = MM_MUL(r0, MM_LOAD(u+i*SIMDD));
-                r5 = MM_DIV(r3, MM_FMA(r4, ra, r1));
-                MM_STORE(tmp4+i*SIMDD, MM_MUL(r2, r5));
-                MM_STORE(tmp1+i*SIMDD, MM_MUL(r4, r5));
-        }
-        ra = MM_SET1(.5);
-        r2 = MM_LOAD(akl);
-        r3 = MM_LOAD(aij);
-        for (i = 0; i < nroots; i++) {
-                r0 = MM_MUL(ra, MM_LOAD(tmp1+i*SIMDD));
-                MM_STORE(b00+i*SIMDD, r0);
-                r1 = MM_LOAD(tmp4+i*SIMDD);
-                MM_STORE(b10+i*SIMDD, MM_FMA(r1, r2, r0));
-                MM_STORE(b01+i*SIMDD, MM_FMA(r1, r3, r0));
-        }
         r4 = MM_LOAD(rijrkl+0*SIMDD);
         r5 = MM_LOAD(rijrkl+1*SIMDD);
         r6 = MM_LOAD(rijrkl+2*SIMDD);
-        ra = MM_LOAD(akl);
-        r1 = MM_LOAD(rijrx+0*SIMDD);
-        r2 = MM_LOAD(rijrx+1*SIMDD);
-        r3 = MM_LOAD(rijrx+2*SIMDD);
-        for (i = 0; i < nroots; i++) {
-                r0 = MM_MUL(MM_LOAD(tmp1+i*SIMDD), ra);
-                MM_STORE(c00x+i*SIMDD, MM_FNMA(r0, r4, r1));
-                MM_STORE(c00y+i*SIMDD, MM_FNMA(r0, r5, r2));
-                MM_STORE(c00z+i*SIMDD, MM_FNMA(r0, r6, r3));
-        }
-        ra = MM_LOAD(aij);
-        r1 = MM_LOAD(rklrx+0*SIMDD);
-        r2 = MM_LOAD(rklrx+1*SIMDD);
-        r3 = MM_LOAD(rklrx+2*SIMDD);
+        __MD _rijrx = MM_LOAD(rijrx+0*SIMDD);
+        __MD _rijry = MM_LOAD(rijrx+1*SIMDD);
+        __MD _rijrz = MM_LOAD(rijrx+2*SIMDD);
+        __MD _rklrx = MM_LOAD(rklrx+0*SIMDD);
+        __MD _rklry = MM_LOAD(rklrx+1*SIMDD);
+        __MD _rklrz = MM_LOAD(rklrx+2*SIMDD);
+        __MD tmp1, tmp2, tmp3, tmp4, tmp5;
         for (i = 0; i < nroots; i++) {
-                r0 = MM_MUL(MM_LOAD(tmp1+i*SIMDD), ra);
-                MM_STORE(c0px+i*SIMDD, MM_FMA (r0, r4, r1));
-                MM_STORE(c0py+i*SIMDD, MM_FMA (r0, r5, r2));
-                MM_STORE(c0pz+i*SIMDD, MM_FMA (r0, r6, r3));
+                tmp1 = MM_MUL(r0, MM_LOAD(u+i*SIMDD));
+                tmp5 = MM_DIV(r3, MM_FMA(tmp1, ra, r1));
+                tmp1 = MM_MUL(tmp1, tmp5);
+                tmp4 = MM_MUL(r2, tmp5);
+                tmp2 = MM_MUL(r2, tmp1);
+                MM_STORE(b00+i*SIMDD, tmp2);
+                MM_STORE(b10+i*SIMDD, MM_FMA(tmp4, akl, tmp2));
+                MM_STORE(b01+i*SIMDD, MM_FMA(tmp4, aij, tmp2));
+
+                tmp2 = MM_MUL(tmp1, akl);
+                tmp3 = MM_MUL(tmp1, aij);
+                MM_STORE(c00x+i*SIMDD, MM_FNMA(tmp2, r4, _rijrx));
+                MM_STORE(c00y+i*SIMDD, MM_FNMA(tmp2, r5, _rijry));
+                MM_STORE(c00z+i*SIMDD, MM_FNMA(tmp2, r6, _rijrz));
+                MM_STORE(c0px+i*SIMDD, MM_FMA (tmp3, r4, _rklrx));
+                MM_STORE(c0py+i*SIMDD, MM_FMA (tmp3, r5, _rklry));
+                MM_STORE(c0pz+i*SIMDD, MM_FMA (tmp3, r6, _rklrz));
         }
 
         (*envs->f_g0_2d4d)(g, bc, envs);
diff --git a/src/g2e.h b/src/g2e.h
index 421665d..c0a4149 100644
--- a/src/g2e.h
+++ b/src/g2e.h
@@ -33,8 +33,6 @@ typedef struct {
         ALIGNMM double b01[SIMDD*MXRYSROOTS];
         ALIGNMM double b00[SIMDD*MXRYSROOTS];
         ALIGNMM double b10[SIMDD*MXRYSROOTS];
-        ALIGNMM double u[MXRYSROOTS*SIMDD];
-        ALIGNMM double w[MXRYSROOTS*SIMDD];
 } Rys2eT;
 #endif
 
@@ -46,8 +44,6 @@ void CINTinit_int3c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
                               int *atm, int natm, int *bas, int nbas, double *env);
 void CINTinit_int2c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
                               int *atm, int natm, int *bas, int nbas, double *env);
-void CINTinit_int2e_coulerf_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
-                                    int *atm, int natm, int *bas, int nbas, double *env);
 void CINTinit_int2e_stg_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
                                 int *atm, int natm, int *bas, int nbas, double *env);
 void CINTinit_int2e_yp_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
@@ -61,6 +57,8 @@ void CINTg0_2e_2d(double *g, Rys2eT *bc, CINTEnvVars *envs);
 void CINTg0_2e_2d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs);
 void CINTg0_2e_2d4d_unrolled(double *g, Rys2eT *bc, CINTEnvVars *envs);
 void CINTg0_2e_2d4d_unrolled_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs);
+void CINTsrg0_2e_2d4d_unrolled(double *g, Rys2eT *bc, CINTEnvVars *envs);
+void CINTsrg0_2e_2d4d_unrolled_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs);
 void CINTg0_2e_lj2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs);
 void CINTg0_2e_kj2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs);
 void CINTg0_2e_il2d4d(double *g, Rys2eT *bc, CINTEnvVars *envs);
@@ -74,8 +72,6 @@ void CINTinit_int2e_stg_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
                                 int *atm, int natm, int *bas, int nbas, double *env);
 void CINTinit_int2e_yp_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
                                int *atm, int natm, int *bas, int nbas, double *env);
-void CINTinit_int2e_coulerf_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
-                                    int *atm, int natm, int *bas, int nbas, double *env);
 
 void CINTnabla1i_2e(double *f, double *g,
                     int li, int lj, int lk, int ll, CINTEnvVars *envs);
@@ -120,16 +116,6 @@ void CINTinit_int2e_stg_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
                                 int *atm, int natm, int *bas, int nbas, double *env);
 #endif
  
-#ifdef WITH_GTG
-void CINTinit_int2e_gtg_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
-                                int *atm, int natm, int *bas, int nbas, double *env);
-void CINTinit_int3c2e_gtg_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
-                                  int *atm, int natm, int *bas, int nbas, double *env);
-void CINTinit_int2c2e_gtg_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
-                                  int *atm, int natm, int *bas, int nbas, double *env);
-#endif
-
-
 #define G2E_D_I(f, g, li, lj, lk, ll)   CINTnabla1i_2e(f, g, li, lj, lk, ll, envs)
 #define G2E_D_J(f, g, li, lj, lk, ll)   CINTnabla1j_2e(f, g, li, lj, lk, ll, envs)
 #define G2E_D_K(f, g, li, lj, lk, ll)   CINTnabla1k_2e(f, g, li, lj, lk, ll, envs)
diff --git a/src/g2e_coulerf.c b/src/g2e_coulerf.c
deleted file mode 100644
index cc41e04..0000000
--- a/src/g2e_coulerf.c
+++ /dev/null
@@ -1,432 +0,0 @@
-/*
- * Qcint is a general GTO integral library for computational chemistry
- * Copyright (C) 2014- Qiming Sun <osirpt.sun@gmail.com>
- *
- * This file is part of Qcint.
- *
- * Qcint is free software: you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation, either version 3 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-/*
- * Attenuated coulomb operator exp(-w r_{12}^2) / r_{12}
- */
-
-#include <stdlib.h>
-#include <math.h>
-#include <assert.h>
-#include "config.h"
-#include "cint_bas.h"
-#include "simd.h"
-#include "rys_roots.h"
-#include "misc.h"
-#include "g2e.h"
-
-int CINTg0_2e_coulerf(double *g, double *cutoff,
-                      Rys2eT *bc, CINTEnvVars *envs, int count);
-int CINTg0_2e_coulerf_simd1(double *g, double *cutoff,
-                            Rys2eT *bc, CINTEnvVars *envs, int idsimd);
-
-void CINTinit_int2e_coulerf_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
-                                    int *atm, int natm, int *bas, int nbas, double *env)
-{
-        envs->natm = natm;
-        envs->nbas = nbas;
-        envs->atm = atm;
-        envs->bas = bas;
-        envs->env = env;
-        envs->shls = shls;
-
-        int i_sh = shls[0];
-        int j_sh = shls[1];
-        int k_sh = shls[2];
-        int l_sh = shls[3];
-        envs->i_l = bas(ANG_OF, i_sh);
-        envs->j_l = bas(ANG_OF, j_sh);
-        envs->k_l = bas(ANG_OF, k_sh);
-        envs->l_l = bas(ANG_OF, l_sh);
-        envs->x_ctr[0] = bas(NCTR_OF, i_sh);
-        envs->x_ctr[1] = bas(NCTR_OF, j_sh);
-        envs->x_ctr[2] = bas(NCTR_OF, k_sh);
-        envs->x_ctr[3] = bas(NCTR_OF, l_sh);
-        envs->nfi = (envs->i_l+1)*(envs->i_l+2)/2;
-        envs->nfj = (envs->j_l+1)*(envs->j_l+2)/2;
-        envs->nfk = (envs->k_l+1)*(envs->k_l+2)/2;
-        envs->nfl = (envs->l_l+1)*(envs->l_l+2)/2;
-        envs->nf = envs->nfi * envs->nfk * envs->nfl * envs->nfj;
-        envs->common_factor = 1;
-        if (env[PTR_EXPCUTOFF] == 0) {
-                envs->expcutoff = EXPCUTOFF;
-        } else {
-                envs->expcutoff = MAX(MIN_EXPCUTOFF, env[PTR_EXPCUTOFF]);
-        }
-
-        envs->gbits = ng[GSHIFT];
-        envs->ncomp_e1 = ng[POS_E1];
-        envs->ncomp_e2 = ng[POS_E2];
-        envs->ncomp_tensor = ng[TENSOR];
-
-        envs->li_ceil = envs->i_l + ng[IINC];
-        envs->lj_ceil = envs->j_l + ng[JINC];
-        envs->lk_ceil = envs->k_l + ng[KINC];
-        envs->ll_ceil = envs->l_l + ng[LINC];
-
-        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
-        envs->rj = env + atm(PTR_COORD, bas(ATOM_OF, j_sh));
-        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
-        envs->rl = env + atm(PTR_COORD, bas(ATOM_OF, l_sh));
-
-        int nroots = (envs->li_ceil + envs->lj_ceil +
-                      envs->lk_ceil + envs->ll_ceil)/2 + 1;
-        envs->nrys_roots = nroots;
-        assert(nroots < MXRYSROOTS);
-
-        int dli, dlj, dlk, dll;
-        int ibase = envs->li_ceil > envs->lj_ceil;
-        int kbase = envs->lk_ceil > envs->ll_ceil;
-        if (nroots <= 2) { // use the fully optimized lj_4d algorithm
-                ibase = 0;
-                kbase = 0;
-        }
-        if (kbase) {
-                if (ibase) {
-                        envs->f_g0_2d4d = &CINTg0_2e_ik2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_ik2d4d_simd1;
-                } else {
-                        envs->f_g0_2d4d = &CINTg0_2e_kj2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_kj2d4d_simd1;
-                }
-        } else {
-                if (ibase) {
-                        envs->f_g0_2d4d = &CINTg0_2e_il2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_il2d4d_simd1;
-                } else {
-                        envs->f_g0_2d4d = &CINTg0_2e_lj2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_lj2d4d_simd1;
-                }
-        }
-        envs->f_g0_2e = &CINTg0_2e_coulerf;
-        envs->f_g0_2e_simd1 = &CINTg0_2e_coulerf_simd1;
-
-        if (kbase) {
-                dlk = envs->lk_ceil + envs->ll_ceil + 1;
-                dll = envs->ll_ceil + 1;
-        } else {
-                dlk = envs->lk_ceil + 1;
-                dll = envs->lk_ceil + envs->ll_ceil + 1;
-        }
-
-        if (ibase) {
-                dli = envs->li_ceil + envs->lj_ceil + 1;
-                dlj = envs->lj_ceil + 1;
-        } else {
-                dli = envs->li_ceil + 1;
-                dlj = envs->li_ceil + envs->lj_ceil + 1;
-        }
-        envs->g_stride_i = nroots;
-        envs->g_stride_k = nroots * dli;
-        envs->g_stride_l = nroots * dli * dlk;
-        envs->g_stride_j = nroots * dli * dlk * dll;
-        envs->g_size     = nroots * dli * dlk * dll * dlj;
-
-        if (kbase) {
-                envs->g2d_klmax = envs->g_stride_k;
-                envs->rx_in_rklrx = envs->rk;
-                envs->rkrl[0] = envs->rk[0] - envs->rl[0];
-                envs->rkrl[1] = envs->rk[1] - envs->rl[1];
-                envs->rkrl[2] = envs->rk[2] - envs->rl[2];
-        } else {
-                envs->g2d_klmax = envs->g_stride_l;
-                envs->rx_in_rklrx = envs->rl;
-                envs->rkrl[0] = envs->rl[0] - envs->rk[0];
-                envs->rkrl[1] = envs->rl[1] - envs->rk[1];
-                envs->rkrl[2] = envs->rl[2] - envs->rk[2];
-        }
-
-        if (ibase) {
-                envs->g2d_ijmax = envs->g_stride_i;
-                envs->rx_in_rijrx = envs->ri;
-                envs->rirj[0] = envs->ri[0] - envs->rj[0];
-                envs->rirj[1] = envs->ri[1] - envs->rj[1];
-                envs->rirj[2] = envs->ri[2] - envs->rj[2];
-        } else {
-                envs->g2d_ijmax = envs->g_stride_j;
-                envs->rx_in_rijrx = envs->rj;
-                envs->rirj[0] = envs->rj[0] - envs->ri[0];
-                envs->rirj[1] = envs->rj[1] - envs->ri[1];
-                envs->rirj[2] = envs->rj[2] - envs->ri[2];
-        }
-}
-
-int CINTg0_2e_coulerf(double *g, double *cutoff,
-                      Rys2eT *bc, CINTEnvVars *envs, int count)
-{
-        ALIGNMM double aij[SIMDD];
-        ALIGNMM double akl[SIMDD];
-        ALIGNMM double a0[SIMDD];
-        ALIGNMM double a1[SIMDD];
-        ALIGNMM double aijkl[SIMDD];
-        ALIGNMM double fac1[SIMDD];
-        ALIGNMM double x[SIMDD];
-        ALIGNMM double rijrkl[SIMDD*3];
-        ALIGNMM double rijrx[SIMDD*3];
-        ALIGNMM double rklrx[SIMDD*3];
-        double *rij = envs->rij;
-        double *rkl = envs->rkl;
-        double *u = bc->u;
-        double *w = bc->w;
-        __MD ra, r0, r1, r2, r3, r4, r5, r6, r7, r8;
-        double omega = envs->env[PTR_RANGE_OMEGA];
-        int nroots = envs->nrys_roots;
-        int i;
-
-        r2 = MM_ADD(MM_LOAD(envs->ai),  MM_LOAD(envs->aj));
-        r3 = MM_ADD(MM_LOAD(envs->ak),  MM_LOAD(envs->al));
-        MM_STORE(aij, r2);
-        MM_STORE(akl, r3);
-        r1 = MM_MUL(r2, r3);
-        MM_STORE(a1, r1);
-        ra = MM_ADD(r2, r3);
-        MM_STORE(aijkl, ra);
-        r0 = MM_DIV(r1, ra);
-        MM_STORE(a0, r0);
-
-        ALIGNMM double theta[SIMDD];
-        if (omega > 0) {
-// For long-range part of range-separated Coulomb operator
-                //:theta = omega * omega / (omega * omega + a0);
-                //:a0 *= theta;
-                r0 = MM_SET1(omega);
-                r1 = MM_LOAD(a0);
-                r0 = MM_MUL(r0, r0);
-                r0 = MM_DIV(r0, MM_ADD(r0, r1));
-                MM_STORE(theta, r0);
-                MM_STORE(a0, MM_MUL(r0, r1));
-        }
-        r1 = MM_LOAD(a1);
-        r0 = MM_DIV(MM_LOAD(a0), MM_MUL(r1, MM_MUL(r1, r1)));
-        MM_STORE(fac1, MM_MUL(MM_SQRT(r0), MM_LOAD(envs->fac)));
-
-        r0 = MM_LOAD(rij+0*SIMDD);
-        r1 = MM_LOAD(rij+1*SIMDD);
-        r2 = MM_LOAD(rij+2*SIMDD);
-        r3 = MM_LOAD(rkl+0*SIMDD);
-        r4 = MM_LOAD(rkl+1*SIMDD);
-        r5 = MM_LOAD(rkl+2*SIMDD);
-
-        r6 = MM_SUB(r0, r3); MM_STORE(rijrkl+0*SIMDD, r6);
-        r7 = MM_SUB(r1, r4); MM_STORE(rijrkl+1*SIMDD, r7);
-        r8 = MM_SUB(r2, r5); MM_STORE(rijrkl+2*SIMDD, r8);
-        ra = MM_FMA(r6, r6, MM_FMA(r7, r7, MM_MUL(r8, r8)));
-        MM_STORE(x, MM_MUL(MM_LOAD(a0), ra));
-        MM_STORE(rijrx+0*SIMDD, MM_SUB(r0, MM_SET1(envs->rx_in_rijrx[0])));
-        MM_STORE(rijrx+1*SIMDD, MM_SUB(r1, MM_SET1(envs->rx_in_rijrx[1])));
-        MM_STORE(rijrx+2*SIMDD, MM_SUB(r2, MM_SET1(envs->rx_in_rijrx[2])));
-        MM_STORE(rklrx+0*SIMDD, MM_SUB(r3, MM_SET1(envs->rx_in_rklrx[0])));
-        MM_STORE(rklrx+1*SIMDD, MM_SUB(r4, MM_SET1(envs->rx_in_rklrx[1])));
-        MM_STORE(rklrx+2*SIMDD, MM_SUB(r5, MM_SET1(envs->rx_in_rklrx[2])));
-
-        _CINTrys_roots_batch(nroots, x, u, w, count);
-
-        double *gx = g;
-        double *gy = gx + envs->g_size * SIMDD;
-        double *gz = gy + envs->g_size * SIMDD;
-        r0 = MM_LOAD(fac1);
-        r1 = MM_SET1(1.);
-        for (i = 0; i < nroots; i++) {
-                //MM_STORE(gx+i*SIMDD, r1);
-                //MM_STORE(gy+i*SIMDD, r1);
-                MM_STORE(gz+i*SIMDD, MM_MUL(MM_LOAD(w+i*SIMDD), r0));
-        }
-        if (envs->g_size == 1) {
-                return 1;
-        }
-
-        if (omega > 0) {
-                /* u[:] = tau^2 / (1 - tau^2)
-                 * transform u[:] to theta^-1 tau^2 / (theta^-1 - tau^2)
-                 * so the rest code can be reused.
-                 */
-                //:for (irys = 0; irys < nroots; irys++) {
-                //:        u[irys] /= u[irys] + 1 - u[irys] * theta;
-                //:}
-                r0 = MM_LOAD(theta);
-                r1 = MM_SET1(1.);
-                for (i = 0; i < nroots; i++) {
-                        r2 = MM_LOAD(u+i*SIMDD);
-                        r3 = r2 + r1 - r2 * r0;
-                        MM_STORE(u+i*SIMDD, MM_DIV(r2, r3));
-                }
-        }
-
-        double *b00 = bc->b00;
-        double *b10 = bc->b10;
-        double *b01 = bc->b01;
-        double *c00x = bc->c00x;
-        double *c00y = bc->c00y;
-        double *c00z = bc->c00z;
-        double *c0px = bc->c0px;
-        double *c0py = bc->c0py;
-        double *c0pz = bc->c0pz;
-
-        ALIGNMM double tmp1[MXRYSROOTS*SIMDD];
-        ALIGNMM double tmp4[MXRYSROOTS*SIMDD];
-
-        ra = MM_LOAD(aijkl);
-        r0 = MM_LOAD(a0);
-        r1 = MM_LOAD(a1);
-        r2 = MM_SET1(.5);
-        r3 = MM_SET1(2.);
-        for (i = 0; i < nroots; i++) {
-                r4 = MM_MUL(r0, MM_LOAD(u+i*SIMDD));
-                r5 = MM_DIV(r2, MM_FMA(r4, ra, r1));
-                MM_STORE(tmp4+i*SIMDD, r5);
-                r6 = MM_MUL(MM_MUL(r3, r4), r5);
-                MM_STORE(tmp1+i*SIMDD, r6);
-        }
-        ra = MM_SET1(.5);
-        r2 = MM_LOAD(akl);
-        r3 = MM_LOAD(aij);
-        for (i = 0; i < nroots; i++) {
-                r0 = MM_MUL(ra, MM_LOAD(tmp1+i*SIMDD));
-                MM_STORE(b00+i*SIMDD, r0);
-                r1 = MM_LOAD(tmp4+i*SIMDD);
-                MM_STORE(b10+i*SIMDD, MM_FMA(r1, r2, r0));
-                MM_STORE(b01+i*SIMDD, MM_FMA(r1, r3, r0));
-        }
-        r4 = MM_LOAD(rijrkl+0*SIMDD);
-        r5 = MM_LOAD(rijrkl+1*SIMDD);
-        r6 = MM_LOAD(rijrkl+2*SIMDD);
-        ra = MM_LOAD(akl);
-        r1 = MM_LOAD(rijrx+0*SIMDD);
-        r2 = MM_LOAD(rijrx+1*SIMDD);
-        r3 = MM_LOAD(rijrx+2*SIMDD);
-        for (i = 0; i < nroots; i++) {
-                r0 = MM_MUL(MM_LOAD(tmp1+i*SIMDD), ra);
-                MM_STORE(c00x+i*SIMDD, MM_FNMA(r0, r4, r1));
-                MM_STORE(c00y+i*SIMDD, MM_FNMA(r0, r5, r2));
-                MM_STORE(c00z+i*SIMDD, MM_FNMA(r0, r6, r3));
-        }
-        ra = MM_LOAD(aij);
-        r1 = MM_LOAD(rklrx+0*SIMDD);
-        r2 = MM_LOAD(rklrx+1*SIMDD);
-        r3 = MM_LOAD(rklrx+2*SIMDD);
-        for (i = 0; i < nroots; i++) {
-                r0 = MM_MUL(MM_LOAD(tmp1+i*SIMDD), ra);
-                MM_STORE(c0px+i*SIMDD, MM_FMA (r0, r4, r1));
-                MM_STORE(c0py+i*SIMDD, MM_FMA (r0, r5, r2));
-                MM_STORE(c0pz+i*SIMDD, MM_FMA (r0, r6, r3));
-        }
-
-        (*envs->f_g0_2d4d)(g, bc, envs);
-        return 1;
-}
-
-int CINTg0_2e_coulerf_simd1(double *g, double *cutoff,
-                            Rys2eT *bc, CINTEnvVars *envs, int idsimd)
-{
-        double aij, akl, a0, a1, fac1;
-        double *rij = envs->rij;
-        double *rkl = envs->rkl;
-        double rijrkl[3];
-        double rijrx[3];
-        double rklrx[3];
-        double *u = bc->u;
-        double *w = bc->w;
-        double omega = envs->env[PTR_RANGE_OMEGA];
-        int nroots = envs->nrys_roots;
-        int i;
-
-        aij = envs->ai[idsimd] + envs->aj[idsimd];
-        akl = envs->ak[idsimd] + envs->al[idsimd];
-        a1 = aij * akl;
-        a0 = a1 / (aij + akl);
-
-        double theta = 0;
-        if (omega > 0) {
-// For long-range part of range-separated Coulomb operator
-                theta = omega * omega / (omega * omega + a0);
-                a0 *= theta;
-        }
-        fac1 = sqrt(a0 / (a1 * a1 * a1)) * envs->fac[idsimd];
-
-        rijrkl[0] = rij[0*SIMDD+idsimd] - rkl[0*SIMDD+idsimd];
-        rijrkl[1] = rij[1*SIMDD+idsimd] - rkl[1*SIMDD+idsimd];
-        rijrkl[2] = rij[2*SIMDD+idsimd] - rkl[2*SIMDD+idsimd];
-        CINTrys_roots(nroots, a0 * SQUARE(rijrkl), u, w);
-
-        double *gx = g;
-        double *gy = g + envs->g_size;
-        double *gz = g + envs->g_size * 2;
-        for (i = 0; i < nroots; i++) {
-                gx[i] = 1;
-                gy[i] = 1;
-                gz[i] = w[i] * fac1;
-        }
-        if (envs->g_size == 1) {
-                return 1;
-        }
-
-        if (omega > 0) {
-                /* u[:] = tau^2 / (1 - tau^2)
-                 * transform u[:] to theta^-1 tau^2 / (theta^-1 - tau^2)
-                 * so the rest code can be reused.
-                 */
-                for (i = 0; i < nroots; i++) {
-                        u[i] /= u[i] + 1 - u[i] * theta;
-                }
-        }
-
-        double u2, div, tmp1, tmp2, tmp3, tmp4;
-        double *b00 = bc->b00;
-        double *b10 = bc->b10;
-        double *b01 = bc->b01;
-        double *c00x = bc->c00x;
-        double *c00y = bc->c00y;
-        double *c00z = bc->c00z;
-        double *c0px = bc->c0px;
-        double *c0py = bc->c0py;
-        double *c0pz = bc->c0pz;
-
-        rijrx[0] = rij[0*SIMDD+idsimd] - envs->rx_in_rijrx[0];
-        rijrx[1] = rij[1*SIMDD+idsimd] - envs->rx_in_rijrx[1];
-        rijrx[2] = rij[2*SIMDD+idsimd] - envs->rx_in_rijrx[2];
-        rklrx[0] = rkl[0*SIMDD+idsimd] - envs->rx_in_rklrx[0];
-        rklrx[1] = rkl[1*SIMDD+idsimd] - envs->rx_in_rklrx[1];
-        rklrx[2] = rkl[2*SIMDD+idsimd] - envs->rx_in_rklrx[2];
-        for (i = 0; i < nroots; i++) {
-                /*
-                 *t2 = u(i)/(1+u(i))
-                 *u2 = aij*akl/(aij+akl)*t2/(1-t2)
-                 */
-                u2 = a0 * u[i];
-                div = 1 / (u2 * (aij + akl) + a1);
-                tmp1 = u2 * div;
-                tmp4 = .5 * div;
-                b00[i] = 0.5 * tmp1;
-                tmp2 = tmp1 * akl;
-                tmp3 = tmp1 * aij;
-                b10[i] = b00[i] + tmp4 * akl;
-                b01[i] = b00[i] + tmp4 * aij;
-                c00x[i] = rijrx[0] - tmp2 * rijrkl[0];
-                c00y[i] = rijrx[1] - tmp2 * rijrkl[1];
-                c00z[i] = rijrx[2] - tmp2 * rijrkl[2];
-                c0px[i] = rklrx[0] + tmp3 * rijrkl[0];
-                c0py[i] = rklrx[1] + tmp3 * rijrkl[1];
-                c0pz[i] = rklrx[2] + tmp3 * rijrkl[2];
-        }
-
-        (*envs->f_g0_2d4d_simd1)(g, bc, envs);
-        return 1;
-}
-
diff --git a/src/g2e_f12.c b/src/g2e_f12.c
index 1afbbd0..7ce266c 100644
--- a/src/g2e_f12.c
+++ b/src/g2e_f12.c
@@ -203,10 +203,10 @@ int CINTg0_2e_yp(double *g, double *cutoff,
         ALIGNMM double rijrkl[SIMDD*3];
         ALIGNMM double rijrx[SIMDD*3];
         ALIGNMM double rklrx[SIMDD*3];
+        ALIGNMM double u[MXRYSROOTS*SIMDD];
         double *rij = envs->rij;
         double *rkl = envs->rkl;
-        double *u = bc->u;
-        double *w = bc->w;
+        double *w = g + envs->g_size * 2; // ~ gz
         __MD ra, r0, r1, r2, r3, r4, r5, r6, r7, r8;
         double zeta = envs->env[PTR_F12_ZETA];
         int nroots = envs->nrys_roots;
@@ -345,18 +345,18 @@ int CINTg0_2e_yp(double *g, double *cutoff,
 int CINTg0_2e_yp_simd1(double *g, double *cutoff,
                        Rys2eT *bc, CINTEnvVars *envs, int idsimd)
 {
-        const double aij = envs->ai[idsimd] + envs->aj[idsimd];
-        const double akl = envs->ak[idsimd] + envs->al[idsimd];
-        const double zeta = envs->env[PTR_F12_ZETA];
-        const int nroots = envs->nrys_roots;
+        double aij = envs->ai[idsimd] + envs->aj[idsimd];
+        double akl = envs->ak[idsimd] + envs->al[idsimd];
+        double zeta = envs->env[PTR_F12_ZETA];
+        int nroots = envs->nrys_roots;
         double a0, a1, fac1, x, ua;
         double *rij = envs->rij;
         double *rkl = envs->rkl;
         double rijrkl[3];
         double rijrx[3];
         double rklrx[3];
-        double *u = bc->u;
-        double *w = bc->w;
+        ALIGNMM double u[MXRYSROOTS];
+        double *w = g + envs->g_size * 2; // ~ gz
         int i;
 
         a1 = aij * akl;
@@ -453,10 +453,10 @@ int CINTg0_2e_stg(double *g, double *cutoff,
         ALIGNMM double rijrkl[SIMDD*3];
         ALIGNMM double rijrx[SIMDD*3];
         ALIGNMM double rklrx[SIMDD*3];
+        ALIGNMM double u[MXRYSROOTS*SIMDD];
         double *rij = envs->rij;
         double *rkl = envs->rkl;
-        double *u = bc->u;
-        double *w = bc->w;
+        double *w = g + envs->g_size * 2; // ~ gz
         __MD ra, r0, r1, r2, r3, r4, r5, r6, r7, r8;
         double zeta = envs->env[PTR_F12_ZETA];
         int nroots = envs->nrys_roots;
@@ -594,18 +594,18 @@ int CINTg0_2e_stg(double *g, double *cutoff,
 int CINTg0_2e_stg_simd1(double *g, double *cutoff,
                         Rys2eT *bc, CINTEnvVars *envs, int idsimd)
 {
-        const double aij = envs->ai[idsimd] + envs->aj[idsimd];
-        const double akl = envs->ak[idsimd] + envs->al[idsimd];
-        const double zeta = envs->env[PTR_F12_ZETA];
-        const int nroots = envs->nrys_roots;
+        double aij = envs->ai[idsimd] + envs->aj[idsimd];
+        double akl = envs->ak[idsimd] + envs->al[idsimd];
+        double zeta = envs->env[PTR_F12_ZETA];
+        int nroots = envs->nrys_roots;
         double a0, a1, fac1, x, ua;
         double *rij = envs->rij;
         double *rkl = envs->rkl;
         double rijrkl[3];
         double rijrx[3];
         double rklrx[3];
-        double *u = bc->u;
-        double *w = bc->w;
+        ALIGNMM double u[MXRYSROOTS*SIMDD];
+        double *w = g + envs->g_size * 2; // ~ gz
         int i;
 
         a1 = aij * akl;
diff --git a/src/g2e_gtg.c b/src/g2e_gtg.c
deleted file mode 100644
index cbc6404..0000000
--- a/src/g2e_gtg.c
+++ /dev/null
@@ -1,446 +0,0 @@
-/*
- * Qcint is a general GTO integral library for computational chemistry
- * Copyright (C) 2014- Qiming Sun <osirpt.sun@gmail.com>
- *
- * This file is part of Qcint.
- *
- * Qcint is free software: you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation, either version 3 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-
-#include <stdlib.h>
-#include <math.h>
-#include <assert.h>
-#include "config.h"
-#include "cint_bas.h"
-#include "simd.h"
-#include "g2e.h"
-#include "rys_roots.h"
-#include "misc.h"
-
-void CINTg0_2e_lj2d4d_regular(double *g, double *cutoff,
-                              Rys2eT *bc, CINTEnvVars *envs);
-void CINTg0_2e_lj2d4d_simd1_regular(double *g, double *cutoff,
-                                    Rys2eT *bc, CINTEnvVars *envs);
-int CINTg0_2e_gtg(double *g, double *cutoff,
-                  Rys2eT *bc, CINTEnvVars *envs, int count);
-int CINTg0_2e_gtg_simd1(double *g, double *cutoff,
-                        Rys2eT *bc, CINTEnvVars *envs, int idsimd);
-void CINTg0_lj_4d(double *g, CINTEnvVars *envs);
-void CINTg0_lj_4d_simd1(double *g, CINTEnvVars *envs);
-
-void CINTinit_int2e_gtg_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
-                                int *atm, int natm, int *bas, int nbas, double *env)
-{
-        envs->natm = natm;
-        envs->nbas = nbas;
-        envs->atm = atm;
-        envs->bas = bas;
-        envs->env = env;
-        envs->shls = shls;
-
-        const int i_sh = shls[0];
-        const int j_sh = shls[1];
-        const int k_sh = shls[2];
-        const int l_sh = shls[3];
-        envs->i_l = bas(ANG_OF, i_sh);
-        envs->j_l = bas(ANG_OF, j_sh);
-        envs->k_l = bas(ANG_OF, k_sh);
-        envs->l_l = bas(ANG_OF, l_sh);
-        envs->x_ctr[0] = bas(NCTR_OF, i_sh);
-        envs->x_ctr[1] = bas(NCTR_OF, j_sh);
-        envs->x_ctr[2] = bas(NCTR_OF, k_sh);
-        envs->x_ctr[3] = bas(NCTR_OF, l_sh);
-        envs->nfi = (envs->i_l+1)*(envs->i_l+2)/2;
-        envs->nfj = (envs->j_l+1)*(envs->j_l+2)/2;
-        envs->nfk = (envs->k_l+1)*(envs->k_l+2)/2;
-        envs->nfl = (envs->l_l+1)*(envs->l_l+2)/2;
-        envs->nf = envs->nfi * envs->nfk * envs->nfl * envs->nfj;
-
-        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
-        envs->rj = env + atm(PTR_COORD, bas(ATOM_OF, j_sh));
-        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
-        envs->rl = env + atm(PTR_COORD, bas(ATOM_OF, l_sh));
-
-        envs->common_factor = SQRTPI * .5;
-        if (env[PTR_EXPCUTOFF] == 0) {
-                envs->expcutoff = EXPCUTOFF;
-        } else {
-                envs->expcutoff = MAX(MIN_EXPCUTOFF, env[PTR_EXPCUTOFF]);
-        }
-
-        envs->gbits = ng[GSHIFT];
-        envs->ncomp_e1 = ng[POS_E1];
-        envs->ncomp_e2 = ng[POS_E2];
-        envs->ncomp_tensor = ng[TENSOR];
-
-        envs->li_ceil = envs->i_l + ng[IINC];
-        envs->lj_ceil = envs->j_l + ng[JINC];
-        envs->lk_ceil = envs->k_l + ng[KINC];
-        envs->ll_ceil = envs->l_l + ng[LINC];
-        envs->nrys_roots = 1;
-
-        assert(i_sh < SHLS_MAX);
-        assert(j_sh < SHLS_MAX);
-        assert(k_sh < SHLS_MAX);
-        assert(l_sh < SHLS_MAX);
-        assert(envs->i_l < ANG_MAX);
-        assert(envs->j_l < ANG_MAX);
-        assert(envs->k_l < ANG_MAX);
-        assert(envs->l_l < ANG_MAX);
-        assert(bas(ATOM_OF,i_sh) >= 0);
-        assert(bas(ATOM_OF,j_sh) >= 0);
-        assert(bas(ATOM_OF,k_sh) >= 0);
-        assert(bas(ATOM_OF,l_sh) >= 0);
-        assert(bas(ATOM_OF,i_sh) < natm);
-        assert(bas(ATOM_OF,j_sh) < natm);
-        assert(bas(ATOM_OF,k_sh) < natm);
-        assert(bas(ATOM_OF,l_sh) < natm);
-        assert(envs->nrys_roots < MXRYSROOTS);
-
-        int dli, dlj, dlk, dll;
-        int ibase = envs->li_ceil > envs->lj_ceil;
-        int kbase = envs->lk_ceil > envs->ll_ceil;
-
-        if (kbase) {
-                dlk = envs->lk_ceil + envs->ll_ceil + 1;
-                dll = envs->ll_ceil + 1;
-        } else {
-                dlk = envs->lk_ceil + 1;
-                dll = envs->lk_ceil + envs->ll_ceil + 1;
-        }
-
-        if (ibase) {
-                dli = envs->li_ceil + envs->lj_ceil + 1;
-                dlj = envs->lj_ceil + 1;
-        } else {
-                dli = envs->li_ceil + 1;
-                dlj = envs->li_ceil + envs->lj_ceil + 1;
-        }
-        envs->g_stride_i = envs->nrys_roots;
-        envs->g_stride_k = envs->nrys_roots * dli;
-        envs->g_stride_l = envs->nrys_roots * dli * dlk;
-        envs->g_stride_j = envs->nrys_roots * dli * dlk * dll;
-        envs->g_size     = envs->nrys_roots * dli * dlk * dll * dlj;
-
-        if (kbase) {
-                envs->g2d_klmax = envs->g_stride_k;
-                envs->rx_in_rklrx = envs->rk;
-                envs->rkrl[0] = envs->rk[0] - envs->rl[0];
-                envs->rkrl[1] = envs->rk[1] - envs->rl[1];
-                envs->rkrl[2] = envs->rk[2] - envs->rl[2];
-        } else {
-                envs->g2d_klmax = envs->g_stride_l;
-                envs->rx_in_rklrx = envs->rl;
-                envs->rkrl[0] = envs->rl[0] - envs->rk[0];
-                envs->rkrl[1] = envs->rl[1] - envs->rk[1];
-                envs->rkrl[2] = envs->rl[2] - envs->rk[2];
-        }
-
-        if (ibase) {
-                envs->g2d_ijmax = envs->g_stride_i;
-                envs->rx_in_rijrx = envs->ri;
-                envs->rirj[0] = envs->ri[0] - envs->rj[0];
-                envs->rirj[1] = envs->ri[1] - envs->rj[1];
-                envs->rirj[2] = envs->ri[2] - envs->rj[2];
-        } else {
-                envs->g2d_ijmax = envs->g_stride_j;
-                envs->rx_in_rijrx = envs->rj;
-                envs->rirj[0] = envs->rj[0] - envs->ri[0];
-                envs->rirj[1] = envs->rj[1] - envs->ri[1];
-                envs->rirj[2] = envs->rj[2] - envs->ri[2];
-        }
-
-        if (kbase) {
-                if (ibase) {
-                        envs->f_g0_2d4d = &CINTg0_2e_ik2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_ik2d4d_simd1;
-                } else {
-                        envs->f_g0_2d4d = &CINTg0_2e_kj2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_kj2d4d_simd1;
-                }
-        } else {
-                if (ibase) {
-                        envs->f_g0_2d4d = &CINTg0_2e_il2d4d;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_il2d4d_simd1;
-                } else {
-                        envs->f_g0_2d4d = &CINTg0_2e_lj2d4d_regular;
-                        envs->f_g0_2d4d_simd1 = &CINTg0_2e_lj2d4d_simd1_regular;
-                }
-        }
-        envs->f_g0_2e = &CINTg0_2e_gtg;
-        envs->f_g0_2e_simd1 = &CINTg0_2e_gtg_simd1;
-}
-
-
-void CINTg0_2e_lj2d4d_regular(double *g, Rys2eT *bc, CINTEnvVars *envs)
-{
-        CINTg0_2e_2d(g, bc, envs);
-        CINTg0_lj_4d(g, envs);
-        return;
-}
-
-void CINTg0_2e_lj2d4d_simd1_regular(double *g, Rys2eT *bc, CINTEnvVars *envs)
-{
-        CINTg0_2e_2d_simd1(g, bc, envs);
-        CINTg0_lj_4d_simd1(g, envs);
-        return;
-}
-
-/*
- * g[i,k,l,j] = < ik | lj > = ( i j | k l )
- */
-int CINTg0_2e_gtg(double *g, double *cutoff,
-                  Rys2eT *bc, CINTEnvVars *envs, int count)
-{
-        const double zeta = envs->env[PTR_GTG_ZETA];
-        double *gx = g;
-        double *gy = gx + envs->g_size * SIMDD;
-        double *gz = gy + envs->g_size * SIMDD;
-        ALIGNMM double aij[SIMDD];
-        ALIGNMM double akl[SIMDD];
-        ALIGNMM double a0[SIMDD];
-        ALIGNMM double a1[SIMDD];
-        ALIGNMM double aijkl[SIMDD];
-        ALIGNMM double fac1[SIMDD];
-        ALIGNMM double x[SIMDD];
-        ALIGNMM double t[SIMDD];
-        ALIGNMM double rijrkl[SIMDD*3];
-        ALIGNMM double rijrx[SIMDD*3];
-        ALIGNMM double rklrx[SIMDD*3];
-        double *rij = envs->rij;
-        double *rkl = envs->rkl;
-        __MD ra, r0, r1, r2, r3, r4, r5, r6, r7, r8;
-
-        //:for (int k = 0; k < count; k++) {
-        //:        aij[k] = envs->ai[k] + envs->aj[k];
-        //:        akl[k] = envs->ak[k] + envs->al[k];
-        //:        aijkl[k] = aij[k] + akl[k];
-        //:        a1[k] = aij[k] * akl[k];
-        //:        a0[k] = a1[k] / aijkl[k];
-        //:}
-        r2 = MM_ADD(MM_LOAD(envs->ai),  MM_LOAD(envs->aj));
-        r3 = MM_ADD(MM_LOAD(envs->ak),  MM_LOAD(envs->al));
-        MM_STORE(aij, r2);
-        MM_STORE(akl, r3);
-        r1 = MM_MUL(r2, r3);
-        MM_STORE(a1, r1);
-        ra = MM_ADD(r2, r3);
-        MM_STORE(aijkl, ra);
-        r0 = MM_DIV(r1, ra);
-        MM_STORE(a0, r0);
-
-        // t = zeta / (zeta + a0);
-        // fac1 = (1-t) / a1;
-        r2 = MM_SET1(zeta);
-        r2 = MM_DIV(r2, MM_ADD(r2, r0));
-        MM_STORE(t, r2);
-        MM_STORE(fac1, MM_DIV(MM_SUB(MM_SET1(1), r2), r1));
-
-        r0 = MM_LOAD(rij+0*SIMDD);
-        r1 = MM_LOAD(rij+1*SIMDD);
-        r2 = MM_LOAD(rij+2*SIMDD);
-        r3 = MM_LOAD(rkl+0*SIMDD);
-        r4 = MM_LOAD(rkl+1*SIMDD);
-        r5 = MM_LOAD(rkl+2*SIMDD);
-
-        //:for (k = 0; k < count; k++) {
-        //:        rijrkl[0*SIMDD+k] = rij[0*SIMDD+k] - rkl[0*SIMDD+k];
-        //:        rijrkl[1*SIMDD+k] = rij[1*SIMDD+k] - rkl[1*SIMDD+k];
-        //:        rijrkl[2*SIMDD+k] = rij[2*SIMDD+k] - rkl[2*SIMDD+k];
-        //:        rijrx[0*SIMDD+k] = rij[0*SIMDD+k] - envs->rx_in_rijrx[0];
-        //:        rijrx[1*SIMDD+k] = rij[1*SIMDD+k] - envs->rx_in_rijrx[1];
-        //:        rijrx[2*SIMDD+k] = rij[2*SIMDD+k] - envs->rx_in_rijrx[2];
-        //:        rklrx[0*SIMDD+k] = rkl[0*SIMDD+k] - envs->rx_in_rklrx[0];
-        //:        rklrx[1*SIMDD+k] = rkl[1*SIMDD+k] - envs->rx_in_rklrx[1];
-        //:        rklrx[2*SIMDD+k] = rkl[2*SIMDD+k] - envs->rx_in_rklrx[2];
-        //:        x[k] = a0[k] *(rijrkl[0*SIMDD+k] * rijrkl[0*SIMDD+k]
-        //:                     + rijrkl[1*SIMDD+k] * rijrkl[1*SIMDD+k]
-        //:                     + rijrkl[2*SIMDD+k] * rijrkl[2*SIMDD+k]);
-        //:}
-        MM_STORE(rijrx+0*SIMDD, MM_SUB(r0, MM_SET1(envs->rx_in_rijrx[0])));
-        MM_STORE(rijrx+1*SIMDD, MM_SUB(r1, MM_SET1(envs->rx_in_rijrx[1])));
-        MM_STORE(rijrx+2*SIMDD, MM_SUB(r2, MM_SET1(envs->rx_in_rijrx[2])));
-        MM_STORE(rklrx+0*SIMDD, MM_SUB(r3, MM_SET1(envs->rx_in_rklrx[0])));
-        MM_STORE(rklrx+1*SIMDD, MM_SUB(r4, MM_SET1(envs->rx_in_rklrx[1])));
-        MM_STORE(rklrx+2*SIMDD, MM_SUB(r5, MM_SET1(envs->rx_in_rklrx[2])));
-
-        r6 = MM_SUB(r0, r3); MM_STORE(rijrkl+0*SIMDD, r6);
-        r7 = MM_SUB(r1, r4); MM_STORE(rijrkl+1*SIMDD, r7);
-        r8 = MM_SUB(r2, r5); MM_STORE(rijrkl+2*SIMDD, r8);
-        ra = MM_FMA(r6, r6, MM_FMA(r7, r7, MM_MUL(r8, r8)));
-
-        // x = -x * t
-        // gz[0] = fac1*sqrt(fac1) * exp(x) * fac;
-        ra = MM_MUL(MM_LOAD(a0), ra);
-        ra = MM_MUL(MM_LOAD(t), ra);
-        MM_STORE(x, -ra);
-        int k;
-        for (k = 0; k < count; k++) {
-                t[k] = exp(x[k]);
-        }
-        ra = MM_LOAD(fac1);
-        r1 = MM_MUL(MM_SQRT(ra), ra);
-        r2 = MM_MUL(MM_LOAD(t), MM_LOAD(envs->fac));
-        MM_STORE(gz, MM_MUL(r1, r2));
-        r1 = MM_SET1(1.);
-        MM_STORE(gx, r1);
-        MM_STORE(gy, r1);
-        if (envs->g_size == 1) {
-                return 1;
-        }
-
-        double *b00 = bc->b00;
-        double *b10 = bc->b10;
-        double *b01 = bc->b01;
-        double *c00x = bc->c00x;
-        double *c00y = bc->c00y;
-        double *c00z = bc->c00z;
-        double *c0px = bc->c0px;
-        double *c0py = bc->c0py;
-        double *c0pz = bc->c0pz;
-
-        ALIGNMM double tmp1[SIMDD];
-        ALIGNMM double tmp4[SIMDD];
-        //:double u2[SIMDD];
-        //:double tmp2, tmp3;
-        //:double div[SIMDD];
-        //:for (k = 0; k < count; k++) {
-        //:        div[k] = 1 / (zeta * aijkl[k] + a1[k]);
-        //:        tmp1[k] = zeta * div[k];
-        //:        tmp4[k] = .5 * div[k];
-        //:
-        //:        b00[k] = 0.5 * tmp1[k];
-        //:        tmp2 = tmp1[k] * akl[k];
-        //:        tmp3 = tmp1[k] * aij[k];
-        //:        b10[k] = b00[k] + tmp4[k] * akl[k];
-        //:        b01[k] = b00[k] + tmp4[k] * aij[k];
-        //:        c00x[k] = rijrx[0*SIMDD+k] - tmp2 * rijrkl[0*SIMDD+k];
-        //:        c00y[k] = rijrx[1*SIMDD+k] - tmp2 * rijrkl[1*SIMDD+k];
-        //:        c00z[k] = rijrx[2*SIMDD+k] - tmp2 * rijrkl[2*SIMDD+k];
-        //:        c0px[k] = rklrx[0*SIMDD+k] + tmp3 * rijrkl[0*SIMDD+k];
-        //:        c0py[k] = rklrx[1*SIMDD+k] + tmp3 * rijrkl[1*SIMDD+k];
-        //:        c0pz[k] = rklrx[2*SIMDD+k] + tmp3 * rijrkl[2*SIMDD+k];
-        //:}
-
-        ra = MM_LOAD(aijkl);
-        r0 = MM_SET1(zeta);
-        r1 = MM_LOAD(a1);
-        r2 = MM_SET1(.5);
-        r3 = MM_SET1(1.);
-
-        r5 = MM_DIV(r3, MM_FMA(r0, ra, r1));
-        r1 = MM_MUL(r0, r5);
-        r4 = MM_MUL(r2, r5);
-        MM_STORE(tmp1, r1);
-        MM_STORE(tmp4, r4);
-
-        r0 = MM_MUL(r2, r1);
-        MM_STORE(b00, r0);
-        MM_STORE(b10, MM_FMA(r4, MM_LOAD(akl), r0));
-        MM_STORE(b01, MM_FMA(r4, MM_LOAD(aij), r0));
-
-        r4 = MM_LOAD(rijrkl+0*SIMDD);
-        r5 = MM_LOAD(rijrkl+1*SIMDD);
-        r6 = MM_LOAD(rijrkl+2*SIMDD);
-        ra = MM_LOAD(akl);
-        r1 = MM_LOAD(rijrx+0*SIMDD);
-        r2 = MM_LOAD(rijrx+1*SIMDD);
-        r3 = MM_LOAD(rijrx+2*SIMDD);
-        r0 = MM_MUL(MM_LOAD(tmp1), ra);
-        MM_STORE(c00x, MM_FNMA(r0, r4, r1));
-        MM_STORE(c00y, MM_FNMA(r0, r5, r2));
-        MM_STORE(c00z, MM_FNMA(r0, r6, r3));
-
-        ra = MM_LOAD(aij);
-        r1 = MM_LOAD(rklrx+0*SIMDD);
-        r2 = MM_LOAD(rklrx+1*SIMDD);
-        r3 = MM_LOAD(rklrx+2*SIMDD);
-        r0 = MM_MUL(MM_LOAD(tmp1), ra);
-        MM_STORE(c0px, MM_FMA (r0, r4, r1));
-        MM_STORE(c0py, MM_FMA (r0, r5, r2));
-        MM_STORE(c0pz, MM_FMA (r0, r6, r3));
-
-        (*envs->f_g0_2d4d)(g, bc, envs);
-        return 1;
-}
-
-int CINTg0_2e_gtg_simd1(double *g, double *cutoff,
-                        Rys2eT *bc, CINTEnvVars *envs, int idsimd)
-{
-        const double aij = envs->ai[idsimd] + envs->aj[idsimd];
-        const double akl = envs->ak[idsimd] + envs->al[idsimd];
-        const double zeta = envs->env[PTR_GTG_ZETA];
-        double *gx = g;
-        double *gy = gx + envs->g_size;
-        double *gz = gy + envs->g_size;
-        double a0, a1, fac1, x, t;
-        double *rij = envs->rij;
-        double *rkl = envs->rkl;
-        double rijrkl[3];
-        double rijrx[3];
-        double rklrx[3];
-        rijrkl[0] = rij[0*SIMDD+idsimd] - rkl[0*SIMDD+idsimd];
-        rijrkl[1] = rij[1*SIMDD+idsimd] - rkl[1*SIMDD+idsimd];
-        rijrkl[2] = rij[2*SIMDD+idsimd] - rkl[2*SIMDD+idsimd];
-
-        a1 = aij * akl;
-        a0 = a1 / (aij + akl);
-        t = zeta / (zeta + a0);
-        x = a0 *(rijrkl[0] * rijrkl[0]
-               + rijrkl[1] * rijrkl[1]
-               + rijrkl[2] * rijrkl[2]);
-        fac1 = (1-t) / a1;
-        gx[0] = 1;
-        gy[0] = 1;
-        gz[0] = fac1*sqrt(fac1) * exp(-t * x) * envs->fac[idsimd];
-        if (envs->g_size == 1) {
-                return 1;
-        }
-
-        double div, tmp1, tmp2, tmp3, tmp4;
-        double *b00 = bc->b00;
-        double *b10 = bc->b10;
-        double *b01 = bc->b01;
-        double *c00x = bc->c00x;
-        double *c00y = bc->c00y;
-        double *c00z = bc->c00z;
-        double *c0px = bc->c0px;
-        double *c0py = bc->c0py;
-        double *c0pz = bc->c0pz;
-
-        rijrx[0] = rij[0*SIMDD+idsimd] - envs->rx_in_rijrx[0];
-        rijrx[1] = rij[1*SIMDD+idsimd] - envs->rx_in_rijrx[1];
-        rijrx[2] = rij[2*SIMDD+idsimd] - envs->rx_in_rijrx[2];
-        rklrx[0] = rkl[0*SIMDD+idsimd] - envs->rx_in_rklrx[0];
-        rklrx[1] = rkl[1*SIMDD+idsimd] - envs->rx_in_rklrx[1];
-        rklrx[2] = rkl[2*SIMDD+idsimd] - envs->rx_in_rklrx[2];
-        div = 1 / (zeta * (aij + akl) + a1);
-        tmp1 = zeta * div;
-        tmp2 = tmp1 * akl;
-        tmp3 = tmp1 * aij;
-        tmp4 = .5 * div;
-        b00[0] = 0.5 * tmp1;
-        b10[0] = b00[0] + tmp4 * akl;
-        b01[0] = b00[0] + tmp4 * aij;
-        c00x[0] = rijrx[0] - tmp2 * rijrkl[0];
-        c00y[0] = rijrx[1] - tmp2 * rijrkl[1];
-        c00z[0] = rijrx[2] - tmp2 * rijrkl[2];
-        c0px[0] = rklrx[0] + tmp3 * rijrkl[0];
-        c0py[0] = rklrx[1] + tmp3 * rijrkl[1];
-        c0pz[0] = rklrx[2] + tmp3 * rijrkl[2];
-
-        (*envs->f_g0_2d4d_simd1)(g, bc, envs);
-        return 1;
-}
diff --git a/src/g2e_simd1.c b/src/g2e_simd1.c
index 969b4f3..599b8f3 100644
--- a/src/g2e_simd1.c
+++ b/src/g2e_simd1.c
@@ -39,12 +39,12 @@
 void CINTg0_2e_2d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
 {
         int nroots = envs->nrys_roots;
-        int i, m, n;
-        DEF_GXYZ(double, g, gx, gy, gz);
         int nmax = envs->li_ceil + envs->lj_ceil;
         int mmax = envs->lk_ceil + envs->ll_ceil;
         int dm = envs->g2d_klmax;
         int dn = envs->g2d_ijmax;
+        int i, m, n;
+        DEF_GXYZ(double, g, gx, gy, gz);
         double *c00x = bc->c00x;
         double *c00y = bc->c00y;
         double *c00z = bc->c00z;
@@ -57,29 +57,39 @@ void CINTg0_2e_2d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
         double *p0x, *p0y, *p0z;
         double *p1x, *p1y, *p1z;
         double nb1, mb0;
+
+        for (i = 0; i < nroots; i++) {
+                gx[i] = 1;
+                gy[i] = 1;
+                //gz[i] = w[i];
+        }
+
         double r0x, r0y, r0z;
         double r1x, r1y, r1z;
         double r2x, r2y, r2z;
-
         if (nmax > 0) {
                 p0x = gx + dn;
                 p0y = gy + dn;
                 p0z = gz + dn;
                 for (i = 0; i < nroots; i++) {
+                        double c00x_i = c00x[i];
+                        double c00y_i = c00y[i];
+                        double c00z_i = c00z[i];
+                        double b10_i = b10[i];
                         r2x = gx[i];
                         r2y = gy[i];
                         r2z = gz[i];
-                        r0x = c00x[i] * r2x;
-                        r0y = c00y[i] * r2y;
-                        r0z = c00z[i] * r2z;
+                        r0x = c00x_i * r2x;
+                        r0y = c00y_i * r2y;
+                        r0z = c00z_i * r2z;
                         p0x[i] = r0x;
                         p0y[i] = r0y;
                         p0z[i] = r0z;
                         for (n = 1; n < nmax; n++) {
-                                nb1 = n * b10[i];
-                                r1x = c00x[i] * r0x + nb1 * r2x;
-                                r1y = c00y[i] * r0y + nb1 * r2y;
-                                r1z = c00z[i] * r0z + nb1 * r2z;
+                                nb1 = n * b10_i;
+                                r1x = c00x_i * r0x + nb1 * r2x;
+                                r1y = c00y_i * r0y + nb1 * r2y;
+                                r1z = c00z_i * r0z + nb1 * r2z;
                                 p0x[i+n*dn] = r1x;
                                 p0y[i+n*dn] = r1y;
                                 p0z[i+n*dn] = r1z;
@@ -98,20 +108,24 @@ void CINTg0_2e_2d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
                 p0y = gy + dm;
                 p0z = gz + dm;
                 for (i = 0; i < nroots; i++) {
+                        double c0px_i = c0px[i];
+                        double c0py_i = c0py[i];
+                        double c0pz_i = c0pz[i];
+                        double b01_i = b01[i];
                         r2x = gx[i];
                         r2y = gy[i];
                         r2z = gz[i];
-                        r0x = c0px[i] * r2x;
-                        r0y = c0py[i] * r2y;
-                        r0z = c0pz[i] * r2z;
+                        r0x = c0px_i * r2x;
+                        r0y = c0py_i * r2y;
+                        r0z = c0pz_i * r2z;
                         p0x[i] = r0x;
                         p0y[i] = r0y;
                         p0z[i] = r0z;
                         for (m = 1; m < mmax; m++) {
-                                mb0 = m * b01[i];
-                                r1x = c0px[i] * r0x + mb0 * r2x;
-                                r1y = c0py[i] * r0y + mb0 * r2y;
-                                r1z = c0pz[i] * r0z + mb0 * r2z;
+                                mb0 = m * b01_i;
+                                r1x = c0px_i * r0x + mb0 * r2x;
+                                r1y = c0py_i * r0y + mb0 * r2y;
+                                r1z = c0pz_i * r0z + mb0 * r2z;
                                 p0x[i+m*dm] = r1x;
                                 p0y[i+m*dm] = r1y;
                                 p0z[i+m*dm] = r1z;
@@ -133,9 +147,17 @@ void CINTg0_2e_2d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
                 p1y = gy + dn;
                 p1z = gz + dn;
                 for (i = 0; i < nroots; i++) {
-                        r1x = c0px[i] * gx[i+dn] + b00[i] * gx[i];
-                        r1y = c0py[i] * gy[i+dn] + b00[i] * gy[i];
-                        r1z = c0pz[i] * gz[i+dn] + b00[i] * gz[i];
+                        double c00x_i = c00x[i];
+                        double c00y_i = c00y[i];
+                        double c00z_i = c00z[i];
+                        double c0px_i = c0px[i];
+                        double c0py_i = c0py[i];
+                        double c0pz_i = c0pz[i];
+                        double b10_i = b10[i];
+                        double b00_i = b00[i];
+                        r1x = c0px_i * gx[i+dn] + b00_i * gx[i];
+                        r1y = c0py_i * gy[i+dn] + b00_i * gy[i];
+                        r1z = c0pz_i * gz[i+dn] + b00_i * gz[i];
                         p0x[i] = r1x;
                         p0y[i] = r1y;
                         p0z[i] = r1z;
@@ -143,9 +165,9 @@ void CINTg0_2e_2d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
                         r2y = gy[i+dm];
                         r2z = gz[i+dm];
                         for (n = 1; n < nmax; n++) {
-                                r0x = c00x[i] * r1x + n * b10[i] * r2x + b00[i] * gx[i+n*dn];
-                                r0y = c00y[i] * r1y + n * b10[i] * r2y + b00[i] * gy[i+n*dn];
-                                r0z = c00z[i] * r1z + n * b10[i] * r2z + b00[i] * gz[i+n*dn];
+                                r0x = c00x_i * r1x + n * b10_i * r2x + b00_i * gx[i+n*dn];
+                                r0y = c00y_i * r1y + n * b10_i * r2y + b00_i * gy[i+n*dn];
+                                r0z = c00z_i * r1z + n * b10_i * r2z + b00_i * gz[i+n*dn];
                                 p0x[i+n*dn] = r0x;
                                 p0y[i+n*dn] = r0y;
                                 p0z[i+n*dn] = r0z;
@@ -160,9 +182,18 @@ void CINTg0_2e_2d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
 
                 for (m = 1; m < mmax; m++) {
                 for (i = 0; i < nroots; i++) {
-                        r1x = c0px[i] * p1x[i+m*dm] + m * b01[i] * p1x[i+(m-1)*dm] + b00[i] * gx[i+m*dm];
-                        r1y = c0py[i] * p1y[i+m*dm] + m * b01[i] * p1y[i+(m-1)*dm] + b00[i] * gy[i+m*dm];
-                        r1z = c0pz[i] * p1z[i+m*dm] + m * b01[i] * p1z[i+(m-1)*dm] + b00[i] * gz[i+m*dm];
+                        double c00x_i = c00x[i];
+                        double c00y_i = c00y[i];
+                        double c00z_i = c00z[i];
+                        double c0px_i = c0px[i];
+                        double c0py_i = c0py[i];
+                        double c0pz_i = c0pz[i];
+                        double b10_i = b10[i];
+                        double b01_i = b01[i];
+                        double b00_i = b00[i];
+                        r1x = c0px_i * p1x[i+m*dm] + m * b01_i * p1x[i+(m-1)*dm] + b00_i * gx[i+m*dm];
+                        r1y = c0py_i * p1y[i+m*dm] + m * b01_i * p1y[i+(m-1)*dm] + b00_i * gy[i+m*dm];
+                        r1z = c0pz_i * p1z[i+m*dm] + m * b01_i * p1z[i+(m-1)*dm] + b00_i * gz[i+m*dm];
                         p0x[i+m*dm] = r1x;
                         p0y[i+m*dm] = r1y;
                         p0z[i+m*dm] = r1z;
@@ -170,11 +201,11 @@ void CINTg0_2e_2d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
                         r2x = gx[i+(m+1)*dm];
                         r2y = gy[i+(m+1)*dm];
                         r2z = gz[i+(m+1)*dm];
-                        mb0 = (m + 1) * b00[i];
+                        mb0 = (m + 1) * b00_i;
                         for (n = 1; n < nmax; n++) {
-                                r0x = c00x[i] * r1x + n * b10[i] * r2x + mb0 * gx[i+m*dm+n*dn];
-                                r0y = c00y[i] * r1y + n * b10[i] * r2y + mb0 * gy[i+m*dm+n*dn];
-                                r0z = c00z[i] * r1z + n * b10[i] * r2z + mb0 * gz[i+m*dm+n*dn];
+                                r0x = c00x_i * r1x + n * b10_i * r2x + mb0 * gx[i+m*dm+n*dn];
+                                r0y = c00y_i * r1y + n * b10_i * r2y + mb0 * gy[i+m*dm+n*dn];
+                                r0z = c00z_i * r1z + n * b10_i * r2z + mb0 * gz[i+m*dm+n*dn];
                                 p0x[i+m*dm+n*dn] = r0x;
                                 p0y[i+m*dm+n*dn] = r0y;
                                 p0z[i+m*dm+n*dn] = r0z;
@@ -215,8 +246,12 @@ void CINTg0_lj_4d_simd1(double *g, CINTEnvVars *envs)
         double *rkrl = envs->rkrl;
         DEF_GXYZ(double, g, gx, gy, gz);
         double *p1x, *p1y, *p1z, *p2x, *p2y, *p2z;
+        double rx, ry, rz;
 
         // g(i,...,j) = rirj * g(i-1,...,j) +  g(i-1,...,j+1)
+        rx = rirj[0];
+        ry = rirj[1];
+        rz = rirj[2];
         p1x = gx - di;
         p1y = gy - di;
         p1z = gz - di;
@@ -228,13 +263,16 @@ void CINTg0_lj_4d_simd1(double *g, CINTEnvVars *envs)
         for (l = 0; l <= mmax; l++) {
                 ptr = j*dj + l*dl + i*di;
                 for (n = ptr; n < ptr+nroots; n++) {
-                        gx[n] = rirj[0] * p1x[n] + p2x[n];
-                        gy[n] = rirj[1] * p1y[n] + p2y[n];
-                        gz[n] = rirj[2] * p1z[n] + p2z[n];
+                        gx[n] = rx * p1x[n] + p2x[n];
+                        gy[n] = ry * p1y[n] + p2y[n];
+                        gz[n] = rz * p1z[n] + p2z[n];
                 }
         } } }
 
         // g(...,k,l,..) = rkrl * g(...,k-1,l,..) + g(...,k-1,l+1,..)
+        rx = rkrl[0];
+        ry = rkrl[1];
+        rz = rkrl[2];
         p1x = gx - dk;
         p1y = gy - dk;
         p1z = gz - dk;
@@ -246,9 +284,9 @@ void CINTg0_lj_4d_simd1(double *g, CINTEnvVars *envs)
         for (l = 0; l <= mmax-k; l++) {
                 ptr = j*dj + l*dl + k*dk;
                 for (n = ptr; n < ptr+dk; n++) {
-                        gx[n] = rkrl[0] * p1x[n] + p2x[n];
-                        gy[n] = rkrl[1] * p1y[n] + p2y[n];
-                        gz[n] = rkrl[2] * p1z[n] + p2z[n];
+                        gx[n] = rx * p1x[n] + p2x[n];
+                        gy[n] = ry * p1y[n] + p2y[n];
+                        gz[n] = rz * p1z[n] + p2z[n];
                 }
         } } }
 }
@@ -274,8 +312,12 @@ void CINTg0_kj_4d_simd1(double *g, CINTEnvVars *envs)
         double *rkrl = envs->rkrl;
         DEF_GXYZ(double, g, gx, gy, gz);
         double *p1x, *p1y, *p1z, *p2x, *p2y, *p2z;
+        double rx, ry, rz;
 
         // g(i,...,j) = rirj * g(i-1,...,j) +  g(i-1,...,j+1)
+        rx = rirj[0];
+        ry = rirj[1];
+        rz = rirj[2];
         p1x = gx - di;
         p1y = gy - di;
         p1z = gz - di;
@@ -287,13 +329,16 @@ void CINTg0_kj_4d_simd1(double *g, CINTEnvVars *envs)
         for (k = 0; k <= mmax; k++) {
                 ptr = j*dj + k*dk + i*di;
                 for (n = ptr; n < ptr+nroots; n++) {
-                        gx[n] = rirj[0] * p1x[n] + p2x[n];
-                        gy[n] = rirj[1] * p1y[n] + p2y[n];
-                        gz[n] = rirj[2] * p1z[n] + p2z[n];
+                        gx[n] = rx * p1x[n] + p2x[n];
+                        gy[n] = ry * p1y[n] + p2y[n];
+                        gz[n] = rz * p1z[n] + p2z[n];
                 }
         } } }
 
         // g(...,k,l,..) = rkrl * g(...,k,l-1,..) + g(...,k+1,l-1,..)
+        rx = rkrl[0];
+        ry = rkrl[1];
+        rz = rkrl[2];
         p1x = gx - dl;
         p1y = gy - dl;
         p1z = gz - dl;
@@ -305,9 +350,9 @@ void CINTg0_kj_4d_simd1(double *g, CINTEnvVars *envs)
         for (k = 0; k <= mmax-l; k++) {
                 ptr = j*dj + l*dl + k*dk;
                 for (n = ptr; n < ptr+dk; n++) {
-                        gx[n] = rkrl[0] * p1x[n] + p2x[n];
-                        gy[n] = rkrl[1] * p1y[n] + p2y[n];
-                        gz[n] = rkrl[2] * p1z[n] + p2z[n];
+                        gx[n] = rx * p1x[n] + p2x[n];
+                        gy[n] = ry * p1y[n] + p2y[n];
+                        gz[n] = rz * p1z[n] + p2z[n];
                 }
         } } }
 }
@@ -333,8 +378,12 @@ void CINTg0_il_4d_simd1(double *g, CINTEnvVars *envs)
         double *rkrl = envs->rkrl;
         DEF_GXYZ(double, g, gx, gy, gz);
         double *p1x, *p1y, *p1z, *p2x, *p2y, *p2z;
+        double rx, ry, rz;
 
         // g(...,k,l,..) = rkrl * g(...,k-1,l,..) + g(...,k-1,l+1,..)
+        rx = rkrl[0];
+        ry = rkrl[1];
+        rz = rkrl[2];
         p1x = gx - dk;
         p1y = gy - dk;
         p1z = gz - dk;
@@ -346,13 +395,16 @@ void CINTg0_il_4d_simd1(double *g, CINTEnvVars *envs)
         for (i = 0; i <= nmax; i++) {
                 ptr = l*dl + k*dk + i*di;
                 for (n = ptr; n < ptr+nroots; n++) {
-                        gx[n] = rkrl[0] * p1x[n] + p2x[n];
-                        gy[n] = rkrl[1] * p1y[n] + p2y[n];
-                        gz[n] = rkrl[2] * p1z[n] + p2z[n];
+                        gx[n] = rx * p1x[n] + p2x[n];
+                        gy[n] = ry * p1y[n] + p2y[n];
+                        gz[n] = rz * p1z[n] + p2z[n];
                 }
         } } }
 
         // g(i,...,j) = rirj * g(i,...,j-1) +  g(i+1,...,j-1)
+        rx = rirj[0];
+        ry = rirj[1];
+        rz = rirj[2];
         p1x = gx - dj;
         p1y = gy - dj;
         p1z = gz - dj;
@@ -364,9 +416,9 @@ void CINTg0_il_4d_simd1(double *g, CINTEnvVars *envs)
         for (k = 0; k <= lk; k++) {
                 ptr = j*dj + l*dl + k*dk;
                 for (n = ptr; n < ptr+dk-di*j; n++) {
-                        gx[n] = rirj[0] * p1x[n] + p2x[n];
-                        gy[n] = rirj[1] * p1y[n] + p2y[n];
-                        gz[n] = rirj[2] * p1z[n] + p2z[n];
+                        gx[n] = rx * p1x[n] + p2x[n];
+                        gy[n] = ry * p1y[n] + p2y[n];
+                        gz[n] = rz * p1z[n] + p2z[n];
                 }
         } } }
 }
@@ -392,8 +444,12 @@ void CINTg0_ik_4d_simd1(double *g, CINTEnvVars *envs)
         double *rkrl = envs->rkrl;
         DEF_GXYZ(double, g, gx, gy, gz);
         double *p1x, *p1y, *p1z, *p2x, *p2y, *p2z;
+        double rx, ry, rz;
 
         // g(...,k,l,..) = rkrl * g(...,k,l-1,..) + g(...,k+1,l-1,..)
+        rx = rkrl[0];
+        ry = rkrl[1];
+        rz = rkrl[2];
         p1x = gx - dl;
         p1y = gy - dl;
         p1z = gz - dl;
@@ -407,14 +463,17 @@ void CINTg0_ik_4d_simd1(double *g, CINTEnvVars *envs)
                 for (i = 0; i <= nmax; i++) {
                         ptr = l*dl + k*dk + i*di;
                         for (n = ptr; n < ptr+nroots; n++) {
-                                gx[n] = rkrl[0] * p1x[n] + p2x[n];
-                                gy[n] = rkrl[1] * p1y[n] + p2y[n];
-                                gz[n] = rkrl[2] * p1z[n] + p2z[n];
+                                gx[n] = rx * p1x[n] + p2x[n];
+                                gy[n] = ry * p1y[n] + p2y[n];
+                                gz[n] = rz * p1z[n] + p2z[n];
                         }
                 } }
         }
 
         // g(i,...,j) = rirj * g(i,...,j-1) +  g(i+1,...,j-1)
+        rx = rirj[0];
+        ry = rirj[1];
+        rz = rirj[2];
         p1x = gx - dj;
         p1y = gy - dj;
         p1z = gz - dj;
@@ -426,17 +485,17 @@ void CINTg0_ik_4d_simd1(double *g, CINTEnvVars *envs)
         for (k = 0; k <= lk; k++) {
                 ptr = j*dj + l*dl + k*dk;
                 for (n = ptr; n < ptr+dk-di*j; n++) {
-                        gx[n] = rirj[0] * p1x[n] + p2x[n];
-                        gy[n] = rirj[1] * p1y[n] + p2y[n];
-                        gz[n] = rirj[2] * p1z[n] + p2z[n];
+                        gx[n] = rx * p1x[n] + p2x[n];
+                        gy[n] = ry * p1y[n] + p2y[n];
+                        gz[n] = rz * p1z[n] + p2z[n];
                 }
         } } }
 }
 
 static inline void _g0_2d4d_0000(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         //g[2] = w[0];
 }
 
@@ -445,9 +504,9 @@ static inline void _g0_2d4d_0001(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpx = bc->c0px;
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
-        //g[0] = 1;
+        g[0] = 1;
         g[1] = cpx[0];
-        //g[2] = 1;
+        g[2] = 1;
         g[3] = cpy[0];
         //g[4] = w[0];
         g[5] = cpz[0] * g[4];
@@ -459,14 +518,14 @@ static inline void _g0_2d4d_0002(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b01 = bc->b01;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[4] = cpx[0] * cpx[0] + b01[0];
         g[5] = cpx[1] * cpx[1] + b01[1];
-        //g[6] = 1;
-        //g[7] = 1;
+        g[6] = 1;
+        g[7] = 1;
         g[8] = cpy[0];
         g[9] = cpy[1];
         g[10] = cpy[0] * cpy[0] + b01[0];
@@ -485,16 +544,16 @@ static inline void _g0_2d4d_0003(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b01 = bc->b01;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[4] = cpx[0] * cpx[0] + b01[0];
         g[5] = cpx[1] * cpx[1] + b01[1];
         g[6] = cpx[0] * (g[4] + 2 * b01[0]);
         g[7] = cpx[1] * (g[5] + 2 * b01[1]);
-        //g[8] = 1;
-        //g[9] = 1;
+        g[8] = 1;
+        g[9] = 1;
         g[10] = cpy[0];
         g[11] = cpy[1];
         g[12] = cpy[0] * cpy[0] + b01[0];
@@ -516,9 +575,9 @@ static inline void _g0_2d4d_0010(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpx = bc->c0px;
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
-        //g[0] = 1;
+        g[0] = 1;
         g[1] = cpx[0];
-        //g[2] = 1;
+        g[2] = 1;
         g[3] = cpy[0];
         //g[4] = w[0];
         g[5] = cpz[0] * g[4];
@@ -533,16 +592,16 @@ static inline void _g0_2d4d_0011(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xkxl = envs->rkrl[0];
         double ykyl = envs->rkrl[1];
         double zkzl = envs->rkrl[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[4] = cpx[0];
         g[5] = cpx[1];
         g[6] = cpx[0] * (xkxl + cpx[0]) + b01[0];
         g[7] = cpx[1] * (xkxl + cpx[1]) + b01[1];
         g[2] = xkxl + cpx[0];
         g[3] = xkxl + cpx[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[16] = cpy[0];
         g[17] = cpy[1];
         g[18] = cpy[0] * (ykyl + cpy[0]) + b01[0];
@@ -568,8 +627,8 @@ static inline void _g0_2d4d_0012(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xkxl = envs->rkrl[0];
         double ykyl = envs->rkrl[1];
         double zkzl = envs->rkrl[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[4] = cpx[0];
         g[5] = cpx[1];
         g[8] = cpx[0] * cpx[0] + b01[0];
@@ -580,8 +639,8 @@ static inline void _g0_2d4d_0012(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[7] = cpx[1] * (xkxl + cpx[1]) + b01[1];
         g[2] = xkxl + cpx[0];
         g[3] = xkxl + cpx[1];
-        //g[16] = 1;
-        //g[17] = 1;
+        g[16] = 1;
+        g[17] = 1;
         g[20] = cpy[0];
         g[21] = cpy[1];
         g[24] = cpy[0] * cpy[0] + b01[0];
@@ -612,14 +671,14 @@ static inline void _g0_2d4d_0020(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b01 = bc->b01;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[4] = cpx[0] * cpx[0] + b01[0];
         g[5] = cpx[1] * cpx[1] + b01[1];
-        //g[6] = 1;
-        //g[7] = 1;
+        g[6] = 1;
+        g[7] = 1;
         g[8] = cpy[0];
         g[9] = cpy[1];
         g[10] = cpy[0] * cpy[0] + b01[0];
@@ -641,8 +700,8 @@ static inline void _g0_2d4d_0021(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xkxl = envs->rkrl[0];
         double ykyl = envs->rkrl[1];
         double zkzl = envs->rkrl[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[4] = cpx[0] * cpx[0] + b01[0];
@@ -653,8 +712,8 @@ static inline void _g0_2d4d_0021(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[11] = cpx[1] * (xkxl + cpx[1]) + b01[1];
         g[12] = g[4] * (xkxl + cpx[0]) + cpx[0] * 2 * b01[0];
         g[13] = g[5] * (xkxl + cpx[1]) + cpx[1] * 2 * b01[1];
-        //g[16] = 1;
-        //g[17] = 1;
+        g[16] = 1;
+        g[17] = 1;
         g[18] = cpy[0];
         g[19] = cpy[1];
         g[20] = cpy[0] * cpy[0] + b01[0];
@@ -685,16 +744,16 @@ static inline void _g0_2d4d_0030(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b01 = bc->b01;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[4] = cpx[0] * cpx[0] + b01[0];
         g[5] = cpx[1] * cpx[1] + b01[1];
         g[6] = cpx[0] * (g[4] + 2 * b01[0]);
         g[7] = cpx[1] * (g[5] + 2 * b01[1]);
-        //g[8] = 1;
-        //g[9] = 1;
+        g[8] = 1;
+        g[9] = 1;
         g[10] = cpy[0];
         g[11] = cpy[1];
         g[12] = cpy[0] * cpy[0] + b01[0];
@@ -716,9 +775,9 @@ static inline void _g0_2d4d_0100(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *c0x = bc->c00x;
         double *c0y = bc->c00y;
         double *c0z = bc->c00z;
-        //g[0] = 1;
+        g[0] = 1;
         g[1] = c0x[0];
-        //g[2] = 1;
+        g[2] = 1;
         g[3] = c0y[0];
         //g[4] = w[0];
         g[5] = c0z[0] * g[4];
@@ -733,16 +792,16 @@ static inline void _g0_2d4d_0101(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[4] = c0x[0];
         g[5] = c0x[1];
         g[6] = cpx[0] * c0x[0] + b00[0];
         g[7] = cpx[1] * c0x[1] + b00[1];
-        //g[8] = 1;
-        //g[9] = 1;
+        g[8] = 1;
+        g[9] = 1;
         g[10] = cpy[0];
         g[11] = cpy[1];
         g[12] = c0y[0];
@@ -769,8 +828,8 @@ static inline void _g0_2d4d_0102(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
         double *b01 = bc->b01;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[6] = c0x[0];
@@ -781,8 +840,8 @@ static inline void _g0_2d4d_0102(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[9] = cpx[1] * c0x[1] + b00[1];
         g[10] = cpx[0] * (g[8] + b00[0]) + b01[0] * c0x[0];
         g[11] = cpx[1] * (g[9] + b00[1]) + b01[1] * c0x[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[14] = cpy[0];
         g[15] = cpy[1];
         g[18] = c0y[0];
@@ -816,16 +875,16 @@ static inline void _g0_2d4d_0110(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[4] = c0x[0];
         g[5] = c0x[1];
         g[6] = cpx[0] * c0x[0] + b00[0];
         g[7] = cpx[1] * c0x[1] + b00[1];
-        //g[8] = 1;
-        //g[9] = 1;
+        g[8] = 1;
+        g[9] = 1;
         g[10] = cpy[0];
         g[11] = cpy[1];
         g[12] = c0y[0];
@@ -855,8 +914,8 @@ static inline void _g0_2d4d_0111(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xkxl = envs->rkrl[0];
         double ykyl = envs->rkrl[1];
         double zkzl = envs->rkrl[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[12] = c0x[0];
         g[13] = c0x[1];
         g[4] = cpx[0];
@@ -871,8 +930,8 @@ static inline void _g0_2d4d_0111(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[3] = xkxl + cpx[1];
         g[14] = c0x[0] * (xkxl + cpx[0]) + b00[0];
         g[15] = c0x[1] * (xkxl + cpx[1]) + b00[1];
-        //g[24] = 1;
-        //g[25] = 1;
+        g[24] = 1;
+        g[25] = 1;
         g[36] = c0y[0];
         g[37] = c0y[1];
         g[28] = cpy[0];
@@ -915,8 +974,8 @@ static inline void _g0_2d4d_0120(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
         double *b01 = bc->b01;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[6] = c0x[0];
@@ -927,8 +986,8 @@ static inline void _g0_2d4d_0120(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[9] = cpx[1] * c0x[1] + b00[1];
         g[10] = cpx[0] * (g[8] + b00[0]) + b01[0] * c0x[0];
         g[11] = cpx[1] * (g[9] + b00[1]) + b01[1] * c0x[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[14] = cpy[0];
         g[15] = cpy[1];
         g[18] = c0y[0];
@@ -959,14 +1018,14 @@ static inline void _g0_2d4d_0200(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *c0y = bc->c00y;
         double *c0z = bc->c00z;
         double *b10 = bc->b10;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = c0x[0] * c0x[0] + b10[0];
         g[5] = c0x[1] * c0x[1] + b10[1];
-        //g[6] = 1;
-        //g[7] = 1;
+        g[6] = 1;
+        g[7] = 1;
         g[8] = c0y[0];
         g[9] = c0y[1];
         g[10] = c0y[0] * c0y[0] + b10[0];
@@ -989,8 +1048,8 @@ static inline void _g0_2d4d_0201(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
         double *b10 = bc->b10;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[4] = c0x[0];
         g[5] = c0x[1];
         g[8] = c0x[0] * c0x[0] + b10[0];
@@ -1001,8 +1060,8 @@ static inline void _g0_2d4d_0201(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[7] = cpx[1] * c0x[1] + b00[1];
         g[10] = c0x[0] * (g[6] + b00[0]) + b10[0] * cpx[0];
         g[11] = c0x[1] * (g[7] + b00[1]) + b10[1] * cpx[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[16] = c0y[0];
         g[17] = c0y[1];
         g[20] = c0y[0] * c0y[0] + b10[0];
@@ -1037,8 +1096,8 @@ static inline void _g0_2d4d_0210(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
         double *b10 = bc->b10;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = cpx[0];
         g[3] = cpx[1];
         g[4] = c0x[0];
@@ -1049,8 +1108,8 @@ static inline void _g0_2d4d_0210(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[9] = c0x[1] * c0x[1] + b10[1];
         g[10] = c0x[0] * (g[6] + b00[0]) + b10[0] * cpx[0];
         g[11] = c0x[1] * (g[7] + b00[1]) + b10[1] * cpx[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[14] = cpy[0];
         g[15] = cpy[1];
         g[16] = c0y[0];
@@ -1084,16 +1143,16 @@ static inline void _g0_2d4d_0300(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b10 = bc->b10;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = c0x[0] * c0x[0] + b10[0];
         g[5] = c0x[1] * c0x[1] + b10[1];
         g[6] = c0x[0] * (g[4] + 2 * b10[0]);
         g[7] = c0x[1] * (g[5] + 2 * b10[1]);
-        //g[8] = 1;
-        //g[9] = 1;
+        g[8] = 1;
+        g[9] = 1;
         g[10] = c0y[0];
         g[11] = c0y[1];
         g[12] = c0y[0] * c0y[0] + b10[0];
@@ -1115,9 +1174,9 @@ static inline void _g0_2d4d_1000(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *c0x = bc->c00x;
         double *c0y = bc->c00y;
         double *c0z = bc->c00z;
-        //g[0] = 1;
+        g[0] = 1;
         g[1] = c0x[0];
-        //g[2] = 1;
+        g[2] = 1;
         g[3] = c0y[0];
         //g[4] = w[0];
         g[5] = c0z[0] * g[4];
@@ -1132,16 +1191,16 @@ static inline void _g0_2d4d_1001(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = cpx[0];
         g[5] = cpx[1];
         g[6] = cpx[0] * c0x[0] + b00[0];
         g[7] = cpx[1] * c0x[1] + b00[1];
-        //g[8] = 1;
-        //g[9] = 1;
+        g[8] = 1;
+        g[9] = 1;
         g[10] = c0y[0];
         g[11] = c0y[1];
         g[12] = cpy[0];
@@ -1168,8 +1227,8 @@ static inline void _g0_2d4d_1002(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
         double *b01 = bc->b01;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = cpx[0];
@@ -1180,8 +1239,8 @@ static inline void _g0_2d4d_1002(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[9] = cpx[1] * cpx[1] + b01[1];
         g[10] = cpx[0] * (g[6] + b00[0]) + b01[0] * c0x[0];
         g[11] = cpx[1] * (g[7] + b00[1]) + b01[1] * c0x[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[14] = c0y[0];
         g[15] = c0y[1];
         g[16] = cpy[0];
@@ -1215,16 +1274,16 @@ static inline void _g0_2d4d_1010(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpy = bc->c0py;
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = cpx[0];
         g[5] = cpx[1];
         g[6] = cpx[0] * c0x[0] + b00[0];
         g[7] = cpx[1] * c0x[1] + b00[1];
-        //g[8] = 1;
-        //g[9] = 1;
+        g[8] = 1;
+        g[9] = 1;
         g[10] = c0y[0];
         g[11] = c0y[1];
         g[12] = cpy[0];
@@ -1254,8 +1313,8 @@ static inline void _g0_2d4d_1011(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xkxl = envs->rkrl[0];
         double ykyl = envs->rkrl[1];
         double zkzl = envs->rkrl[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[8] = cpx[0];
@@ -1270,8 +1329,8 @@ static inline void _g0_2d4d_1011(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[5] = xkxl + cpx[1];
         g[6] = c0x[0] * (xkxl + cpx[0]) + b00[0];
         g[7] = c0x[1] * (xkxl + cpx[1]) + b00[1];
-        //g[24] = 1;
-        //g[25] = 1;
+        g[24] = 1;
+        g[25] = 1;
         g[26] = c0y[0];
         g[27] = c0y[1];
         g[32] = cpy[0];
@@ -1314,8 +1373,8 @@ static inline void _g0_2d4d_1020(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
         double *b01 = bc->b01;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = cpx[0];
@@ -1326,8 +1385,8 @@ static inline void _g0_2d4d_1020(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[9] = cpx[1] * cpx[1] + b01[1];
         g[10] = cpx[0] * (g[6] + b00[0]) + b01[0] * c0x[0];
         g[11] = cpx[1] * (g[7] + b00[1]) + b01[1] * c0x[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[14] = c0y[0];
         g[15] = c0y[1];
         g[16] = cpy[0];
@@ -1361,16 +1420,16 @@ static inline void _g0_2d4d_1100(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xixj = envs->rirj[0];
         double yiyj = envs->rirj[1];
         double zizj = envs->rirj[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[4] = c0x[0];
         g[5] = c0x[1];
         g[6] = c0x[0] * (xixj + c0x[0]) + b10[0];
         g[7] = c0x[1] * (xixj + c0x[1]) + b10[1];
         g[2] = xixj + c0x[0];
         g[3] = xixj + c0x[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[16] = c0y[0];
         g[17] = c0y[1];
         g[18] = c0y[0] * (yiyj + c0y[0]) + b10[0];
@@ -1400,8 +1459,8 @@ static inline void _g0_2d4d_1101(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xixj = envs->rirj[0];
         double yiyj = envs->rirj[1];
         double zizj = envs->rirj[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[8] = c0x[0];
         g[9] = c0x[1];
         g[4] = cpx[0];
@@ -1416,8 +1475,8 @@ static inline void _g0_2d4d_1101(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[15] = g[13] * (xixj + c0x[1]) + c0x[1] * b00[1] + b10[1] * cpx[1];
         g[6] = cpx[0] * (xixj + c0x[0]) + b00[0];
         g[7] = cpx[1] * (xixj + c0x[1]) + b00[1];
-        //g[24] = 1;
-        //g[25] = 1;
+        g[24] = 1;
+        g[25] = 1;
         g[32] = c0y[0];
         g[33] = c0y[1];
         g[28] = cpy[0];
@@ -1463,8 +1522,8 @@ static inline void _g0_2d4d_1110(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xixj = envs->rirj[0];
         double yiyj = envs->rirj[1];
         double zizj = envs->rirj[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[8] = c0x[0];
         g[9] = c0x[1];
         g[4] = cpx[0];
@@ -1479,8 +1538,8 @@ static inline void _g0_2d4d_1110(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[15] = g[13] * (xixj + c0x[1]) + c0x[1] * b00[1] + b10[1] * cpx[1];
         g[6] = cpx[0] * (xixj + c0x[0]) + b00[0];
         g[7] = cpx[1] * (xixj + c0x[1]) + b00[1];
-        //g[24] = 1;
-        //g[25] = 1;
+        g[24] = 1;
+        g[25] = 1;
         g[32] = c0y[0];
         g[33] = c0y[1];
         g[28] = cpy[0];
@@ -1522,8 +1581,8 @@ static inline void _g0_2d4d_1200(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xixj = envs->rirj[0];
         double yiyj = envs->rirj[1];
         double zizj = envs->rirj[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[4] = c0x[0];
         g[5] = c0x[1];
         g[8] = c0x[0] * c0x[0] + b10[0];
@@ -1534,8 +1593,8 @@ static inline void _g0_2d4d_1200(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[7] = c0x[1] * (xixj + c0x[1]) + b10[1];
         g[2] = xixj + c0x[0];
         g[3] = xixj + c0x[1];
-        //g[16] = 1;
-        //g[17] = 1;
+        g[16] = 1;
+        g[17] = 1;
         g[20] = c0y[0];
         g[21] = c0y[1];
         g[24] = c0y[0] * c0y[0] + b10[0];
@@ -1566,14 +1625,14 @@ static inline void _g0_2d4d_2000(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *c0y = bc->c00y;
         double *c0z = bc->c00z;
         double *b10 = bc->b10;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = c0x[0] * c0x[0] + b10[0];
         g[5] = c0x[1] * c0x[1] + b10[1];
-        //g[6] = 1;
-        //g[7] = 1;
+        g[6] = 1;
+        g[7] = 1;
         g[8] = c0y[0];
         g[9] = c0y[1];
         g[10] = c0y[0] * c0y[0] + b10[0];
@@ -1596,8 +1655,8 @@ static inline void _g0_2d4d_2001(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
         double *b10 = bc->b10;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = c0x[0] * c0x[0] + b10[0];
@@ -1608,8 +1667,8 @@ static inline void _g0_2d4d_2001(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[9] = cpx[1] * c0x[1] + b00[1];
         g[10] = c0x[0] * (g[8] + b00[0]) + b10[0] * cpx[0];
         g[11] = c0x[1] * (g[9] + b00[1]) + b10[1] * cpx[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[14] = c0y[0];
         g[15] = c0y[1];
         g[16] = c0y[0] * c0y[0] + b10[0];
@@ -1644,8 +1703,8 @@ static inline void _g0_2d4d_2010(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *cpz = bc->c0pz;
         double *b00 = bc->b00;
         double *b10 = bc->b10;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = c0x[0] * c0x[0] + b10[0];
@@ -1656,8 +1715,8 @@ static inline void _g0_2d4d_2010(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[9] = cpx[1] * c0x[1] + b00[1];
         g[10] = c0x[0] * (g[8] + b00[0]) + b10[0] * cpx[0];
         g[11] = c0x[1] * (g[9] + b00[1]) + b10[1] * cpx[1];
-        //g[12] = 1;
-        //g[13] = 1;
+        g[12] = 1;
+        g[13] = 1;
         g[14] = c0y[0];
         g[15] = c0y[1];
         g[16] = c0y[0] * c0y[0] + b10[0];
@@ -1691,8 +1750,8 @@ static inline void _g0_2d4d_2100(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double xixj = envs->rirj[0];
         double yiyj = envs->rirj[1];
         double zizj = envs->rirj[2];
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = c0x[0] * c0x[0] + b10[0];
@@ -1703,8 +1762,8 @@ static inline void _g0_2d4d_2100(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[11] = c0x[1] * (xixj + c0x[1]) + b10[1];
         g[8] = xixj + c0x[0];
         g[9] = xixj + c0x[1];
-        //g[16] = 1;
-        //g[17] = 1;
+        g[16] = 1;
+        g[17] = 1;
         g[18] = c0y[0];
         g[19] = c0y[1];
         g[20] = c0y[0] * c0y[0] + b10[0];
@@ -1735,16 +1794,16 @@ static inline void _g0_2d4d_3000(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         double *c0y = bc->c00y;
         double *c0z = bc->c00z;
         double *b10 = bc->b10;
-        //g[0] = 1;
-        //g[1] = 1;
+        g[0] = 1;
+        g[1] = 1;
         g[2] = c0x[0];
         g[3] = c0x[1];
         g[4] = c0x[0] * c0x[0] + b10[0];
         g[5] = c0x[1] * c0x[1] + b10[1];
         g[6] = c0x[0] * (g[4] + 2 * b10[0]);
         g[7] = c0x[1] * (g[5] + 2 * b10[1]);
-        //g[8] = 1;
-        //g[9] = 1;
+        g[8] = 1;
+        g[9] = 1;
         g[10] = c0y[0];
         g[11] = c0y[1];
         g[12] = c0y[0] * c0y[0] + b10[0];
@@ -1761,7 +1820,7 @@ static inline void _g0_2d4d_3000(double *restrict g, Rys2eT *bc, CINTEnvVars *en
         g[23] = c0z[1] * g[21] + 2 * b10[1] * g[19];
 }
 
-void CINTg0_2e_2d4d_unrolled_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+void CINTg0_2e_2d4d_unrolled_simd1(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
 {
         int type_ijkl = ((envs->li_ceil << 6) | (envs->lj_ceil << 4) |
                          (envs->lk_ceil << 2) | (envs->ll_ceil));
@@ -1807,146 +1866,2481 @@ void CINTg0_2e_2d4d_unrolled_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
                (int)envs->ll_ceil, (int)envs->lj_ceil);
 }
 
-void CINTg0_2e_lj2d4d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+static inline void _srg0_2d4d_0000(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        CINTg0_2e_2d_simd1(g, bc, envs);
-        CINTg0_lj_4d_simd1(g, envs);
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        //g[4] = w[0];
+        //g[5] = w[1];
 }
 
-void CINTg0_2e_kj2d4d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+static inline void _srg0_2d4d_0001(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        CINTg0_2e_2d_simd1(g, bc, envs);
-        CINTg0_kj_4d_simd1(g, envs);
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = cpx[0];
+        g[3] = cpx[1];
+        g[4] = 1;
+        g[5] = 1;
+        g[6] = cpy[0];
+        g[7] = cpy[1];
+        //g[8] = w[0];
+        //g[9] = w[0];
+        g[10] = cpz[0] * g[8];
+        g[11] = cpz[1] * g[9];
 }
-void CINTg0_2e_ik2d4d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+
+static inline void _srg0_2d4d_0002(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        CINTg0_2e_2d_simd1(g, bc, envs);
-        CINTg0_ik_4d_simd1(g, envs);
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b01 = bc->b01;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[8] = cpx[0] * cpx[0] + b01[0];
+        g[9] = cpx[1] * cpx[1] + b01[1];
+        g[10] = cpx[2] * cpx[2] + b01[2];
+        g[11] = cpx[3] * cpx[3] + b01[3];
+        g[12] = 1;
+        g[13] = 1;
+        g[14] = 1;
+        g[15] = 1;
+        g[16] = cpy[0];
+        g[17] = cpy[1];
+        g[18] = cpy[2];
+        g[19] = cpy[3];
+        g[20] = cpy[0] * cpy[0] + b01[0];
+        g[21] = cpy[1] * cpy[1] + b01[1];
+        g[22] = cpy[2] * cpy[2] + b01[2];
+        g[23] = cpy[3] * cpy[3] + b01[3];
+        //g[24] = w[0];
+        //g[25] = w[0];
+        //g[26] = w[1];
+        //g[27] = w[1];
+        g[28] = cpz[0] * g[24];
+        g[29] = cpz[1] * g[25];
+        g[30] = cpz[2] * g[26];
+        g[31] = cpz[3] * g[27];
+        g[32] = cpz[0] * g[28] + b01[0] * g[24];
+        g[33] = cpz[1] * g[29] + b01[1] * g[25];
+        g[34] = cpz[2] * g[30] + b01[2] * g[26];
+        g[35] = cpz[3] * g[31] + b01[3] * g[27];
 }
-void CINTg0_2e_il2d4d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+
+static inline void _srg0_2d4d_0003(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        CINTg0_2e_2d_simd1(g, bc, envs);
-        CINTg0_il_4d_simd1(g, envs);
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b01 = bc->b01;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[8] = cpx[0] * cpx[0] + b01[0];
+        g[9] = cpx[1] * cpx[1] + b01[1];
+        g[10] = cpx[2] * cpx[2] + b01[2];
+        g[11] = cpx[3] * cpx[3] + b01[3];
+        g[12] = cpx[0] * (g[8] + 2 * b01[0]);
+        g[13] = cpx[1] * (g[9] + 2 * b01[1]);
+        g[14] = cpx[2] * (g[10] + 2 * b01[2]);
+        g[15] = cpx[3] * (g[11] + 2 * b01[3]);
+        g[16] = 1;
+        g[17] = 1;
+        g[18] = 1;
+        g[19] = 1;
+        g[20] = cpy[0];
+        g[21] = cpy[1];
+        g[22] = cpy[2];
+        g[23] = cpy[3];
+        g[24] = cpy[0] * cpy[0] + b01[0];
+        g[25] = cpy[1] * cpy[1] + b01[1];
+        g[26] = cpy[2] * cpy[2] + b01[2];
+        g[27] = cpy[3] * cpy[3] + b01[3];
+        g[28] = cpy[0] * (g[24] + 2 * b01[0]);
+        g[29] = cpy[1] * (g[25] + 2 * b01[1]);
+        g[30] = cpy[2] * (g[26] + 2 * b01[2]);
+        g[31] = cpy[3] * (g[27] + 2 * b01[3]);
+        //g[32] = w[0];
+        //g[33] = w[0];
+        //g[34] = w[1];
+        //g[35] = w[1];
+        g[36] = cpz[0] * g[32];
+        g[37] = cpz[1] * g[33];
+        g[38] = cpz[2] * g[34];
+        g[39] = cpz[3] * g[35];
+        g[40] = cpz[0] * g[36] + b01[0] * g[32];
+        g[41] = cpz[1] * g[37] + b01[1] * g[33];
+        g[42] = cpz[2] * g[38] + b01[2] * g[34];
+        g[43] = cpz[3] * g[39] + b01[3] * g[35];
+        g[44] = cpz[0] * g[40] + 2 * b01[0] * g[36];
+        g[45] = cpz[1] * g[41] + 2 * b01[1] * g[37];
+        g[46] = cpz[2] * g[42] + 2 * b01[2] * g[38];
+        g[47] = cpz[3] * g[43] + 2 * b01[3] * g[39];
 }
 
-/*
- * g[i,k,l,j] = < ik | lj > = ( i j | k l )
- */
-int CINTg0_2e_simd1(double *g, double *cutoff,
-                    Rys2eT *bc, CINTEnvVars *envs, int idsimd)
+static inline void _srg0_2d4d_0010(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
 {
-        int nroots = envs->nrys_roots;
-        double aij, akl, a0, a1, fac1;
-        double x;
-        double *rij = envs->rij;
-        double *rkl = envs->rkl;
-        double rijrkl[3];
-        double rijrx[3];
-        double rklrx[3];
-        double *u = bc->u;
-        double *w = bc->w;
-
-        int i;
-        aij = envs->ai[idsimd] + envs->aj[idsimd];
-        akl = envs->ak[idsimd] + envs->al[idsimd];
-        a1 = aij * akl;
-        a0 = a1 / (aij + akl);
-#ifdef WITH_RANGE_COULOMB
-        const double omega = envs->env[PTR_RANGE_OMEGA];
-        double theta = 1;
-        if (omega != 0) {
-                theta = omega * omega / (omega * omega + a0);
-                if (omega > 0) { // long-range part of range-separated Coulomb
-                        a0 *= theta;
-                }
-        }
-#endif
-        fac1 = sqrt(a0 / (a1 * a1 * a1)) * envs->fac[idsimd];
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = cpx[0];
+        g[3] = cpx[1];
+        g[4] = 1;
+        g[5] = 1;
+        g[6] = cpy[0];
+        g[7] = cpy[1];
+        //g[8] = w[0];
+        //g[9] = w[0];
+        g[10] = cpz[0] * g[8];
+        g[11] = cpz[1] * g[9];
+}
 
-        rijrkl[0] = rij[0*SIMDD+idsimd] - rkl[0*SIMDD+idsimd];
-        rijrkl[1] = rij[1*SIMDD+idsimd] - rkl[1*SIMDD+idsimd];
-        rijrkl[2] = rij[2*SIMDD+idsimd] - rkl[2*SIMDD+idsimd];
-        x = a0 * SQUARE(rijrkl);
+static inline void _srg0_2d4d_0011(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b01 = bc->b01;
+        double xkxl = envs->rkrl[0];
+        double ykyl = envs->rkrl[1];
+        double zkzl = envs->rkrl[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[8] = cpx[0];
+        g[9] = cpx[1];
+        g[10] = cpx[2];
+        g[11] = cpx[3];
+        g[12] = cpx[0] * (xkxl + cpx[0]) + b01[0];
+        g[13] = cpx[1] * (xkxl + cpx[1]) + b01[1];
+        g[14] = cpx[2] * (xkxl + cpx[2]) + b01[2];
+        g[15] = cpx[3] * (xkxl + cpx[3]) + b01[3];
+        g[4] = xkxl + cpx[0];
+        g[5] = xkxl + cpx[1];
+        g[6] = xkxl + cpx[2];
+        g[7] = xkxl + cpx[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[32] = cpy[0];
+        g[33] = cpy[1];
+        g[34] = cpy[2];
+        g[35] = cpy[3];
+        g[36] = cpy[0] * (ykyl + cpy[0]) + b01[0];
+        g[37] = cpy[1] * (ykyl + cpy[1]) + b01[1];
+        g[38] = cpy[2] * (ykyl + cpy[2]) + b01[2];
+        g[39] = cpy[3] * (ykyl + cpy[3]) + b01[3];
+        g[28] = ykyl + cpy[0];
+        g[29] = ykyl + cpy[1];
+        g[30] = ykyl + cpy[2];
+        g[31] = ykyl + cpy[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[56] = cpz[0] * g[48];
+        g[57] = cpz[1] * g[49];
+        g[58] = cpz[2] * g[50];
+        g[59] = cpz[3] * g[51];
+        g[60] = g[56] * (zkzl + cpz[0]) + b01[0] * g[48];
+        g[61] = g[57] * (zkzl + cpz[1]) + b01[1] * g[49];
+        g[62] = g[58] * (zkzl + cpz[2]) + b01[2] * g[50];
+        g[63] = g[59] * (zkzl + cpz[3]) + b01[3] * g[51];
+        g[52] = g[48] * (zkzl + cpz[0]);
+        g[53] = g[49] * (zkzl + cpz[1]);
+        g[54] = g[50] * (zkzl + cpz[2]);
+        g[55] = g[51] * (zkzl + cpz[3]);
+}
 
-#ifdef WITH_RANGE_COULOMB
-        if (omega < 0) { // short-range part of range-separated Coulomb
-                // very small erfc() leads to ~0 weights. They can cause
-                // numerical issue in sr_rys_roots.
-                if (theta * x > cutoff[idsimd] || theta * x > EXPCUTOFF_SR) {
-                        for (i = 0; i < envs->g_size * 3; i++) {
-                                g[i] = 0;
-                        }
-                        return 0;
-                }
-                CINTsr_rys_roots(nroots, x, sqrt(theta), u, w);
-        } else {
-                CINTrys_roots(nroots, x, u, w);
-        }
-#else
-        CINTrys_roots(nroots, x, u, w);
-#endif
+static inline void _srg0_2d4d_0012(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b01 = bc->b01;
+        double xkxl = envs->rkrl[0];
+        double ykyl = envs->rkrl[1];
+        double zkzl = envs->rkrl[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[8] = cpx[0];
+        g[9] = cpx[1];
+        g[10] = cpx[2];
+        g[11] = cpx[3];
+        g[16] = cpx[0] * cpx[0] + b01[0];
+        g[17] = cpx[1] * cpx[1] + b01[1];
+        g[18] = cpx[2] * cpx[2] + b01[2];
+        g[19] = cpx[3] * cpx[3] + b01[3];
+        g[20] = g[16] * (xkxl + cpx[0]) + cpx[0] * 2 * b01[0];
+        g[21] = g[17] * (xkxl + cpx[1]) + cpx[1] * 2 * b01[1];
+        g[22] = g[18] * (xkxl + cpx[2]) + cpx[2] * 2 * b01[2];
+        g[23] = g[19] * (xkxl + cpx[3]) + cpx[3] * 2 * b01[3];
+        g[12] = cpx[0] * (xkxl + cpx[0]) + b01[0];
+        g[13] = cpx[1] * (xkxl + cpx[1]) + b01[1];
+        g[14] = cpx[2] * (xkxl + cpx[2]) + b01[2];
+        g[15] = cpx[3] * (xkxl + cpx[3]) + b01[3];
+        g[4] = xkxl + cpx[0];
+        g[5] = xkxl + cpx[1];
+        g[6] = xkxl + cpx[2];
+        g[7] = xkxl + cpx[3];
+        g[32] = 1;
+        g[33] = 1;
+        g[34] = 1;
+        g[35] = 1;
+        g[40] = cpy[0];
+        g[41] = cpy[1];
+        g[42] = cpy[2];
+        g[43] = cpy[3];
+        g[48] = cpy[0] * cpy[0] + b01[0];
+        g[49] = cpy[1] * cpy[1] + b01[1];
+        g[50] = cpy[2] * cpy[2] + b01[2];
+        g[51] = cpy[3] * cpy[3] + b01[3];
+        g[52] = g[48] * (ykyl + cpy[0]) + cpy[0] * 2 * b01[0];
+        g[53] = g[49] * (ykyl + cpy[1]) + cpy[1] * 2 * b01[1];
+        g[54] = g[50] * (ykyl + cpy[2]) + cpy[2] * 2 * b01[2];
+        g[55] = g[51] * (ykyl + cpy[3]) + cpy[3] * 2 * b01[3];
+        g[44] = cpy[0] * (ykyl + cpy[0]) + b01[0];
+        g[45] = cpy[1] * (ykyl + cpy[1]) + b01[1];
+        g[46] = cpy[2] * (ykyl + cpy[2]) + b01[2];
+        g[47] = cpy[3] * (ykyl + cpy[3]) + b01[3];
+        g[36] = ykyl + cpy[0];
+        g[37] = ykyl + cpy[1];
+        g[38] = ykyl + cpy[2];
+        g[39] = ykyl + cpy[3];
+        //g[64] = w[0];
+        //g[65] = w[0];
+        //g[66] = w[1];
+        //g[67] = w[1];
+        g[72] = cpz[0] * g[64];
+        g[73] = cpz[1] * g[65];
+        g[74] = cpz[2] * g[66];
+        g[75] = cpz[3] * g[67];
+        g[80] = cpz[0] * g[72] + b01[0] * g[64];
+        g[81] = cpz[1] * g[73] + b01[1] * g[65];
+        g[82] = cpz[2] * g[74] + b01[2] * g[66];
+        g[83] = cpz[3] * g[75] + b01[3] * g[67];
+        g[84] = g[80] * (zkzl + cpz[0]) + 2 * b01[0] * g[72];
+        g[85] = g[81] * (zkzl + cpz[1]) + 2 * b01[1] * g[73];
+        g[86] = g[82] * (zkzl + cpz[2]) + 2 * b01[2] * g[74];
+        g[87] = g[83] * (zkzl + cpz[3]) + 2 * b01[3] * g[75];
+        g[76] = g[72] * (zkzl + cpz[0]) + b01[0] * g[64];
+        g[77] = g[73] * (zkzl + cpz[1]) + b01[1] * g[65];
+        g[78] = g[74] * (zkzl + cpz[2]) + b01[2] * g[66];
+        g[79] = g[75] * (zkzl + cpz[3]) + b01[3] * g[67];
+        g[68] = g[64] * (zkzl + cpz[0]);
+        g[69] = g[65] * (zkzl + cpz[1]);
+        g[70] = g[66] * (zkzl + cpz[2]);
+        g[71] = g[67] * (zkzl + cpz[3]);
+}
 
-        double *gx = g;
-        double *gy = g + envs->g_size;
-        double *gz = g + envs->g_size * 2;
-        for (i = 0; i < nroots; i++) {
-                gx[i] = 1;
-                gy[i] = 1;
-                gz[i] = w[i] * fac1;
-        }
-        if (envs->g_size == 1) {
-                return 1;
-        }
+static inline void _srg0_2d4d_0020(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b01 = bc->b01;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[8] = cpx[0] * cpx[0] + b01[0];
+        g[9] = cpx[1] * cpx[1] + b01[1];
+        g[10] = cpx[2] * cpx[2] + b01[2];
+        g[11] = cpx[3] * cpx[3] + b01[3];
+        g[12] = 1;
+        g[13] = 1;
+        g[14] = 1;
+        g[15] = 1;
+        g[16] = cpy[0];
+        g[17] = cpy[1];
+        g[18] = cpy[2];
+        g[19] = cpy[3];
+        g[20] = cpy[0] * cpy[0] + b01[0];
+        g[21] = cpy[1] * cpy[1] + b01[1];
+        g[22] = cpy[2] * cpy[2] + b01[2];
+        g[23] = cpy[3] * cpy[3] + b01[3];
+        //g[24] = w[0];
+        //g[25] = w[0];
+        //g[26] = w[1];
+        //g[27] = w[1];
+        g[28] = cpz[0] * g[24];
+        g[29] = cpz[1] * g[25];
+        g[30] = cpz[2] * g[26];
+        g[31] = cpz[3] * g[27];
+        g[32] = cpz[0] * g[28] + b01[0] * g[24];
+        g[33] = cpz[1] * g[29] + b01[1] * g[25];
+        g[34] = cpz[2] * g[30] + b01[2] * g[26];
+        g[35] = cpz[3] * g[31] + b01[3] * g[27];
+}
 
-#ifdef WITH_RANGE_COULOMB
-        if (omega > 0) {
-                /* u[:] = tau^2 / (1 - tau^2)
-                 * transform u[:] to theta^-1 tau^2 / (theta^-1 - tau^2)
-                 * so the rest code can be reused.
-                 */
-                for (i = 0; i < nroots; i++) {
-                        u[i] /= u[i] + 1 - u[i] * theta;
-                }
-        }
-#endif
+static inline void _srg0_2d4d_0021(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b01 = bc->b01;
+        double xkxl = envs->rkrl[0];
+        double ykyl = envs->rkrl[1];
+        double zkzl = envs->rkrl[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[8] = cpx[0] * cpx[0] + b01[0];
+        g[9] = cpx[1] * cpx[1] + b01[1];
+        g[10] = cpx[2] * cpx[2] + b01[2];
+        g[11] = cpx[3] * cpx[3] + b01[3];
+        g[16] = xkxl + cpx[0];
+        g[17] = xkxl + cpx[1];
+        g[18] = xkxl + cpx[2];
+        g[19] = xkxl + cpx[3];
+        g[20] = cpx[0] * (xkxl + cpx[0]) + b01[0];
+        g[21] = cpx[1] * (xkxl + cpx[1]) + b01[1];
+        g[22] = cpx[2] * (xkxl + cpx[2]) + b01[2];
+        g[23] = cpx[3] * (xkxl + cpx[3]) + b01[3];
+        g[24] = g[8] * (xkxl + cpx[0]) + cpx[0] * 2 * b01[0];
+        g[25] = g[9] * (xkxl + cpx[1]) + cpx[1] * 2 * b01[1];
+        g[26] = g[10] * (xkxl + cpx[2]) + cpx[2] * 2 * b01[2];
+        g[27] = g[11] * (xkxl + cpx[3]) + cpx[3] * 2 * b01[3];
+        g[32] = 1;
+        g[33] = 1;
+        g[34] = 1;
+        g[35] = 1;
+        g[36] = cpy[0];
+        g[37] = cpy[1];
+        g[38] = cpy[2];
+        g[39] = cpy[3];
+        g[40] = cpy[0] * cpy[0] + b01[0];
+        g[41] = cpy[1] * cpy[1] + b01[1];
+        g[42] = cpy[2] * cpy[2] + b01[2];
+        g[43] = cpy[3] * cpy[3] + b01[3];
+        g[48] = ykyl + cpy[0];
+        g[49] = ykyl + cpy[1];
+        g[50] = ykyl + cpy[2];
+        g[51] = ykyl + cpy[3];
+        g[52] = cpy[0] * (ykyl + cpy[0]) + b01[0];
+        g[53] = cpy[1] * (ykyl + cpy[1]) + b01[1];
+        g[54] = cpy[2] * (ykyl + cpy[2]) + b01[2];
+        g[55] = cpy[3] * (ykyl + cpy[3]) + b01[3];
+        g[56] = g[40] * (ykyl + cpy[0]) + cpy[0] * 2 * b01[0];
+        g[57] = g[41] * (ykyl + cpy[1]) + cpy[1] * 2 * b01[1];
+        g[58] = g[42] * (ykyl + cpy[2]) + cpy[2] * 2 * b01[2];
+        g[59] = g[43] * (ykyl + cpy[3]) + cpy[3] * 2 * b01[3];
+        //g[64] = w[0];
+        //g[65] = w[0];
+        //g[66] = w[1];
+        //g[67] = w[1];
+        g[68] = cpz[0] * g[64];
+        g[69] = cpz[1] * g[65];
+        g[70] = cpz[2] * g[66];
+        g[71] = cpz[3] * g[67];
+        g[72] = cpz[0] * g[68] + b01[0] * g[64];
+        g[73] = cpz[1] * g[69] + b01[1] * g[65];
+        g[74] = cpz[2] * g[70] + b01[2] * g[66];
+        g[75] = cpz[3] * g[71] + b01[3] * g[67];
+        g[80] = g[64] * (zkzl + cpz[0]);
+        g[81] = g[65] * (zkzl + cpz[1]);
+        g[82] = g[66] * (zkzl + cpz[2]);
+        g[83] = g[67] * (zkzl + cpz[3]);
+        g[84] = g[68] * (zkzl + cpz[0]) + b01[0] * g[64];
+        g[85] = g[69] * (zkzl + cpz[1]) + b01[1] * g[65];
+        g[86] = g[70] * (zkzl + cpz[2]) + b01[2] * g[66];
+        g[87] = g[71] * (zkzl + cpz[3]) + b01[3] * g[67];
+        g[88] = g[72] * (zkzl + cpz[0]) + 2 * b01[0] * g[68];
+        g[89] = g[73] * (zkzl + cpz[1]) + 2 * b01[1] * g[69];
+        g[90] = g[74] * (zkzl + cpz[2]) + 2 * b01[2] * g[70];
+        g[91] = g[75] * (zkzl + cpz[3]) + 2 * b01[3] * g[71];
+}
 
-        double u2, div, tmp1, tmp2, tmp3, tmp4;
-        double *b00 = bc->b00;
-        double *b10 = bc->b10;
+static inline void _srg0_2d4d_0030(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
         double *b01 = bc->b01;
-        double *c00x = bc->c00x;
-        double *c00y = bc->c00y;
-        double *c00z = bc->c00z;
-        double *c0px = bc->c0px;
-        double *c0py = bc->c0py;
-        double *c0pz = bc->c0pz;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[8] = cpx[0] * cpx[0] + b01[0];
+        g[9] = cpx[1] * cpx[1] + b01[1];
+        g[10] = cpx[2] * cpx[2] + b01[2];
+        g[11] = cpx[3] * cpx[3] + b01[3];
+        g[12] = cpx[0] * (g[8] + 2 * b01[0]);
+        g[13] = cpx[1] * (g[9] + 2 * b01[1]);
+        g[14] = cpx[2] * (g[10] + 2 * b01[2]);
+        g[15] = cpx[3] * (g[11] + 2 * b01[3]);
+        g[16] = 1;
+        g[17] = 1;
+        g[18] = 1;
+        g[19] = 1;
+        g[20] = cpy[0];
+        g[21] = cpy[1];
+        g[22] = cpy[2];
+        g[23] = cpy[3];
+        g[24] = cpy[0] * cpy[0] + b01[0];
+        g[25] = cpy[1] * cpy[1] + b01[1];
+        g[26] = cpy[2] * cpy[2] + b01[2];
+        g[27] = cpy[3] * cpy[3] + b01[3];
+        g[28] = cpy[0] * (g[24] + 2 * b01[0]);
+        g[29] = cpy[1] * (g[25] + 2 * b01[1]);
+        g[30] = cpy[2] * (g[26] + 2 * b01[2]);
+        g[31] = cpy[3] * (g[27] + 2 * b01[3]);
+        //g[32] = w[0];
+        //g[33] = w[0];
+        //g[34] = w[1];
+        //g[35] = w[1];
+        g[36] = cpz[0] * g[32];
+        g[37] = cpz[1] * g[33];
+        g[38] = cpz[2] * g[34];
+        g[39] = cpz[3] * g[35];
+        g[40] = cpz[0] * g[36] + b01[0] * g[32];
+        g[41] = cpz[1] * g[37] + b01[1] * g[33];
+        g[42] = cpz[2] * g[38] + b01[2] * g[34];
+        g[43] = cpz[3] * g[39] + b01[3] * g[35];
+        g[44] = cpz[0] * g[40] + 2 * b01[0] * g[36];
+        g[45] = cpz[1] * g[41] + 2 * b01[1] * g[37];
+        g[46] = cpz[2] * g[42] + 2 * b01[2] * g[38];
+        g[47] = cpz[3] * g[43] + 2 * b01[3] * g[39];
+}
 
-        rijrx[0] = rij[0*SIMDD+idsimd] - envs->rx_in_rijrx[0];
-        rijrx[1] = rij[1*SIMDD+idsimd] - envs->rx_in_rijrx[1];
-        rijrx[2] = rij[2*SIMDD+idsimd] - envs->rx_in_rijrx[2];
-        rklrx[0] = rkl[0*SIMDD+idsimd] - envs->rx_in_rklrx[0];
-        rklrx[1] = rkl[1*SIMDD+idsimd] - envs->rx_in_rklrx[1];
-        rklrx[2] = rkl[2*SIMDD+idsimd] - envs->rx_in_rklrx[2];
-        for (i = 0; i < nroots; i++) {
+static inline void _srg0_2d4d_0100(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = c0x[0];
+        g[3] = c0x[1];
+        g[4] = 1;
+        g[5] = 1;
+        g[6] = c0y[0];
+        g[7] = c0y[1];
+        //g[8] = w[0];
+        //g[9] = w[0];
+        g[10] = c0z[0] * g[8];
+        g[11] = c0z[1] * g[9];
+}
+
+static inline void _srg0_2d4d_0101(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[8] = c0x[0];
+        g[9] = c0x[1];
+        g[10] = c0x[2];
+        g[11] = c0x[3];
+        g[12] = cpx[0] * c0x[0] + b00[0];
+        g[13] = cpx[1] * c0x[1] + b00[1];
+        g[14] = cpx[2] * c0x[2] + b00[2];
+        g[15] = cpx[3] * c0x[3] + b00[3];
+        g[16] = 1;
+        g[17] = 1;
+        g[18] = 1;
+        g[19] = 1;
+        g[20] = cpy[0];
+        g[21] = cpy[1];
+        g[22] = cpy[2];
+        g[23] = cpy[3];
+        g[24] = c0y[0];
+        g[25] = c0y[1];
+        g[26] = c0y[2];
+        g[27] = c0y[3];
+        g[28] = cpy[0] * c0y[0] + b00[0];
+        g[29] = cpy[1] * c0y[1] + b00[1];
+        g[30] = cpy[2] * c0y[2] + b00[2];
+        g[31] = cpy[3] * c0y[3] + b00[3];
+        //g[32] = w[0];
+        //g[33] = w[0];
+        //g[34] = w[1];
+        //g[35] = w[1];
+        g[36] = cpz[0] * g[32];
+        g[37] = cpz[1] * g[33];
+        g[38] = cpz[2] * g[34];
+        g[39] = cpz[3] * g[35];
+        g[40] = c0z[0] * g[32];
+        g[41] = c0z[1] * g[33];
+        g[42] = c0z[2] * g[34];
+        g[43] = c0z[3] * g[35];
+        g[44] = cpz[0] * g[40] + b00[0] * g[32];
+        g[45] = cpz[1] * g[41] + b00[1] * g[33];
+        g[46] = cpz[2] * g[42] + b00[2] * g[34];
+        g[47] = cpz[3] * g[43] + b00[3] * g[35];
+}
+
+static inline void _srg0_2d4d_0102(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b01 = bc->b01;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[12] = c0x[0];
+        g[13] = c0x[1];
+        g[14] = c0x[2];
+        g[15] = c0x[3];
+        g[8] = cpx[0] * cpx[0] + b01[0];
+        g[9] = cpx[1] * cpx[1] + b01[1];
+        g[10] = cpx[2] * cpx[2] + b01[2];
+        g[11] = cpx[3] * cpx[3] + b01[3];
+        g[16] = cpx[0] * c0x[0] + b00[0];
+        g[17] = cpx[1] * c0x[1] + b00[1];
+        g[18] = cpx[2] * c0x[2] + b00[2];
+        g[19] = cpx[3] * c0x[3] + b00[3];
+        g[20] = cpx[0] * (g[16] + b00[0]) + b01[0] * c0x[0];
+        g[21] = cpx[1] * (g[17] + b00[1]) + b01[1] * c0x[1];
+        g[22] = cpx[2] * (g[18] + b00[2]) + b01[2] * c0x[2];
+        g[23] = cpx[3] * (g[19] + b00[3]) + b01[3] * c0x[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[28] = cpy[0];
+        g[29] = cpy[1];
+        g[30] = cpy[2];
+        g[31] = cpy[3];
+        g[36] = c0y[0];
+        g[37] = c0y[1];
+        g[38] = c0y[2];
+        g[39] = c0y[3];
+        g[32] = cpy[0] * cpy[0] + b01[0];
+        g[33] = cpy[1] * cpy[1] + b01[1];
+        g[34] = cpy[2] * cpy[2] + b01[2];
+        g[35] = cpy[3] * cpy[3] + b01[3];
+        g[40] = cpy[0] * c0y[0] + b00[0];
+        g[41] = cpy[1] * c0y[1] + b00[1];
+        g[42] = cpy[2] * c0y[2] + b00[2];
+        g[43] = cpy[3] * c0y[3] + b00[3];
+        g[44] = cpy[0] * (g[40] + b00[0]) + b01[0] * c0y[0];
+        g[45] = cpy[1] * (g[41] + b00[1]) + b01[1] * c0y[1];
+        g[46] = cpy[2] * (g[42] + b00[2]) + b01[2] * c0y[2];
+        g[47] = cpy[3] * (g[43] + b00[3]) + b01[3] * c0y[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[52] = cpz[0] * g[48];
+        g[53] = cpz[1] * g[49];
+        g[54] = cpz[2] * g[50];
+        g[55] = cpz[3] * g[51];
+        g[60] = c0z[0] * g[48];
+        g[61] = c0z[1] * g[49];
+        g[62] = c0z[2] * g[50];
+        g[63] = c0z[3] * g[51];
+        g[56] = cpz[0] * g[52] + b01[0] * g[48];
+        g[57] = cpz[1] * g[53] + b01[1] * g[49];
+        g[58] = cpz[2] * g[54] + b01[2] * g[50];
+        g[59] = cpz[3] * g[55] + b01[3] * g[51];
+        g[64] = cpz[0] * g[60] + b00[0] * g[48];
+        g[65] = cpz[1] * g[61] + b00[1] * g[49];
+        g[66] = cpz[2] * g[62] + b00[2] * g[50];
+        g[67] = cpz[3] * g[63] + b00[3] * g[51];
+        g[68] = cpz[0] * g[64] + b01[0] * g[60] + b00[0] * g[52];
+        g[69] = cpz[1] * g[65] + b01[1] * g[61] + b00[1] * g[53];
+        g[70] = cpz[2] * g[66] + b01[2] * g[62] + b00[2] * g[54];
+        g[71] = cpz[3] * g[67] + b01[3] * g[63] + b00[3] * g[55];
+}
+
+static inline void _srg0_2d4d_0110(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[8] = c0x[0];
+        g[9] = c0x[1];
+        g[10] = c0x[2];
+        g[11] = c0x[3];
+        g[12] = cpx[0] * c0x[0] + b00[0];
+        g[13] = cpx[1] * c0x[1] + b00[1];
+        g[14] = cpx[2] * c0x[2] + b00[2];
+        g[15] = cpx[3] * c0x[3] + b00[3];
+        g[16] = 1;
+        g[17] = 1;
+        g[18] = 1;
+        g[19] = 1;
+        g[20] = cpy[0];
+        g[21] = cpy[1];
+        g[22] = cpy[2];
+        g[23] = cpy[3];
+        g[24] = c0y[0];
+        g[25] = c0y[1];
+        g[26] = c0y[2];
+        g[27] = c0y[3];
+        g[28] = cpy[0] * c0y[0] + b00[0];
+        g[29] = cpy[1] * c0y[1] + b00[1];
+        g[30] = cpy[2] * c0y[2] + b00[2];
+        g[31] = cpy[3] * c0y[3] + b00[3];
+        //g[32] = w[0];
+        //g[33] = w[0];
+        //g[34] = w[1];
+        //g[35] = w[1];
+        g[36] = cpz[0] * g[32];
+        g[37] = cpz[1] * g[33];
+        g[38] = cpz[2] * g[34];
+        g[39] = cpz[3] * g[35];
+        g[40] = c0z[0] * g[32];
+        g[41] = c0z[1] * g[33];
+        g[42] = c0z[2] * g[34];
+        g[43] = c0z[3] * g[35];
+        g[44] = cpz[0] * g[40] + b00[0] * g[32];
+        g[45] = cpz[1] * g[41] + b00[1] * g[33];
+        g[46] = cpz[2] * g[42] + b00[2] * g[34];
+        g[47] = cpz[3] * g[43] + b00[3] * g[35];
+}
+
+static inline void _srg0_2d4d_0111(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b01 = bc->b01;
+        double xkxl = envs->rkrl[0];
+        double ykyl = envs->rkrl[1];
+        double zkzl = envs->rkrl[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[24] = c0x[0];
+        g[25] = c0x[1];
+        g[26] = c0x[2];
+        g[27] = c0x[3];
+        g[8] = cpx[0];
+        g[9] = cpx[1];
+        g[10] = cpx[2];
+        g[11] = cpx[3];
+        g[32] = cpx[0] * c0x[0] + b00[0];
+        g[33] = cpx[1] * c0x[1] + b00[1];
+        g[34] = cpx[2] * c0x[2] + b00[2];
+        g[35] = cpx[3] * c0x[3] + b00[3];
+        g[12] = cpx[0] * (xkxl + cpx[0]) + b01[0];
+        g[13] = cpx[1] * (xkxl + cpx[1]) + b01[1];
+        g[14] = cpx[2] * (xkxl + cpx[2]) + b01[2];
+        g[15] = cpx[3] * (xkxl + cpx[3]) + b01[3];
+        g[36] = g[32] * (xkxl + cpx[0]) + cpx[0] * b00[0] + b01[0] * c0x[0];
+        g[37] = g[33] * (xkxl + cpx[1]) + cpx[1] * b00[1] + b01[1] * c0x[1];
+        g[38] = g[34] * (xkxl + cpx[2]) + cpx[2] * b00[2] + b01[2] * c0x[2];
+        g[39] = g[35] * (xkxl + cpx[3]) + cpx[3] * b00[3] + b01[3] * c0x[3];
+        g[4] = xkxl + cpx[0];
+        g[5] = xkxl + cpx[1];
+        g[6] = xkxl + cpx[2];
+        g[7] = xkxl + cpx[3];
+        g[28] = c0x[0] * (xkxl + cpx[0]) + b00[0];
+        g[29] = c0x[1] * (xkxl + cpx[1]) + b00[1];
+        g[30] = c0x[2] * (xkxl + cpx[2]) + b00[2];
+        g[31] = c0x[3] * (xkxl + cpx[3]) + b00[3];
+        g[48] = 1;
+        g[49] = 1;
+        g[50] = 1;
+        g[51] = 1;
+        g[72] = c0y[0];
+        g[73] = c0y[1];
+        g[74] = c0y[2];
+        g[75] = c0y[3];
+        g[56] = cpy[0];
+        g[57] = cpy[1];
+        g[58] = cpy[2];
+        g[59] = cpy[3];
+        g[80] = cpy[0] * c0y[0] + b00[0];
+        g[81] = cpy[1] * c0y[1] + b00[1];
+        g[82] = cpy[2] * c0y[2] + b00[2];
+        g[83] = cpy[3] * c0y[3] + b00[3];
+        g[60] = cpy[0] * (ykyl + cpy[0]) + b01[0];
+        g[61] = cpy[1] * (ykyl + cpy[1]) + b01[1];
+        g[62] = cpy[2] * (ykyl + cpy[2]) + b01[2];
+        g[63] = cpy[3] * (ykyl + cpy[3]) + b01[3];
+        g[84] = g[80] * (ykyl + cpy[0]) + cpy[0] * b00[0] + b01[0] * c0y[0];
+        g[85] = g[81] * (ykyl + cpy[1]) + cpy[1] * b00[1] + b01[1] * c0y[1];
+        g[86] = g[82] * (ykyl + cpy[2]) + cpy[2] * b00[2] + b01[2] * c0y[2];
+        g[87] = g[83] * (ykyl + cpy[3]) + cpy[3] * b00[3] + b01[3] * c0y[3];
+        g[52] = ykyl + cpy[0];
+        g[53] = ykyl + cpy[1];
+        g[54] = ykyl + cpy[2];
+        g[55] = ykyl + cpy[3];
+        g[76] = c0y[0] * (ykyl + cpy[0]) + b00[0];
+        g[77] = c0y[1] * (ykyl + cpy[1]) + b00[1];
+        g[78] = c0y[2] * (ykyl + cpy[2]) + b00[2];
+        g[79] = c0y[3] * (ykyl + cpy[3]) + b00[3];
+        //g[96] = w[0];
+        //g[97] = w[0];
+        //g[98] = w[1];
+        //g[99] = w[1];
+        g[120] = c0z[0] * g[96];
+        g[121] = c0z[1] * g[97];
+        g[122] = c0z[2] * g[98];
+        g[123] = c0z[3] * g[99];
+        g[104] = cpz[0] * g[96];
+        g[105] = cpz[1] * g[97];
+        g[106] = cpz[2] * g[98];
+        g[107] = cpz[3] * g[99];
+        g[128] = cpz[0] * g[120] + b00[0] * g[96];
+        g[129] = cpz[1] * g[121] + b00[1] * g[97];
+        g[130] = cpz[2] * g[122] + b00[2] * g[98];
+        g[131] = cpz[3] * g[123] + b00[3] * g[99];
+        g[108] = g[104] * (zkzl + cpz[0]) + b01[0] * g[96];
+        g[109] = g[105] * (zkzl + cpz[1]) + b01[1] * g[97];
+        g[110] = g[106] * (zkzl + cpz[2]) + b01[2] * g[98];
+        g[111] = g[107] * (zkzl + cpz[3]) + b01[3] * g[99];
+        g[132] = g[128] * (zkzl + cpz[0]) + b01[0] * g[120] + b00[0] * g[104];
+        g[133] = g[129] * (zkzl + cpz[1]) + b01[1] * g[121] + b00[1] * g[105];
+        g[134] = g[130] * (zkzl + cpz[2]) + b01[2] * g[122] + b00[2] * g[106];
+        g[135] = g[131] * (zkzl + cpz[3]) + b01[3] * g[123] + b00[3] * g[107];
+        g[100] = g[96] * (zkzl + cpz[0]);
+        g[101] = g[97] * (zkzl + cpz[1]);
+        g[102] = g[98] * (zkzl + cpz[2]);
+        g[103] = g[99] * (zkzl + cpz[3]);
+        g[124] = g[120] * (zkzl + cpz[0]) + b00[0] * g[96];
+        g[125] = g[121] * (zkzl + cpz[1]) + b00[1] * g[97];
+        g[126] = g[122] * (zkzl + cpz[2]) + b00[2] * g[98];
+        g[127] = g[123] * (zkzl + cpz[3]) + b00[3] * g[99];
+}
+
+static inline void _srg0_2d4d_0120(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b01 = bc->b01;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[12] = c0x[0];
+        g[13] = c0x[1];
+        g[14] = c0x[2];
+        g[15] = c0x[3];
+        g[8] = cpx[0] * cpx[0] + b01[0];
+        g[9] = cpx[1] * cpx[1] + b01[1];
+        g[10] = cpx[2] * cpx[2] + b01[2];
+        g[11] = cpx[3] * cpx[3] + b01[3];
+        g[16] = cpx[0] * c0x[0] + b00[0];
+        g[17] = cpx[1] * c0x[1] + b00[1];
+        g[18] = cpx[2] * c0x[2] + b00[2];
+        g[19] = cpx[3] * c0x[3] + b00[3];
+        g[20] = cpx[0] * (g[16] + b00[0]) + b01[0] * c0x[0];
+        g[21] = cpx[1] * (g[17] + b00[1]) + b01[1] * c0x[1];
+        g[22] = cpx[2] * (g[18] + b00[2]) + b01[2] * c0x[2];
+        g[23] = cpx[3] * (g[19] + b00[3]) + b01[3] * c0x[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[28] = cpy[0];
+        g[29] = cpy[1];
+        g[30] = cpy[2];
+        g[31] = cpy[3];
+        g[36] = c0y[0];
+        g[37] = c0y[1];
+        g[38] = c0y[2];
+        g[39] = c0y[3];
+        g[32] = cpy[0] * cpy[0] + b01[0];
+        g[33] = cpy[1] * cpy[1] + b01[1];
+        g[34] = cpy[2] * cpy[2] + b01[2];
+        g[35] = cpy[3] * cpy[3] + b01[3];
+        g[40] = cpy[0] * c0y[0] + b00[0];
+        g[41] = cpy[1] * c0y[1] + b00[1];
+        g[42] = cpy[2] * c0y[2] + b00[2];
+        g[43] = cpy[3] * c0y[3] + b00[3];
+        g[44] = cpy[0] * (g[40] + b00[0]) + b01[0] * c0y[0];
+        g[45] = cpy[1] * (g[41] + b00[1]) + b01[1] * c0y[1];
+        g[46] = cpy[2] * (g[42] + b00[2]) + b01[2] * c0y[2];
+        g[47] = cpy[3] * (g[43] + b00[3]) + b01[3] * c0y[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[52] = cpz[0] * g[48];
+        g[53] = cpz[1] * g[49];
+        g[54] = cpz[2] * g[50];
+        g[55] = cpz[3] * g[51];
+        g[60] = c0z[0] * g[48];
+        g[61] = c0z[1] * g[49];
+        g[62] = c0z[2] * g[50];
+        g[63] = c0z[3] * g[51];
+        g[56] = cpz[0] * g[52] + b01[0] * g[48];
+        g[57] = cpz[1] * g[53] + b01[1] * g[49];
+        g[58] = cpz[2] * g[54] + b01[2] * g[50];
+        g[59] = cpz[3] * g[55] + b01[3] * g[51];
+        g[64] = cpz[0] * g[60] + b00[0] * g[48];
+        g[65] = cpz[1] * g[61] + b00[1] * g[49];
+        g[66] = cpz[2] * g[62] + b00[2] * g[50];
+        g[67] = cpz[3] * g[63] + b00[3] * g[51];
+        g[68] = cpz[0] * g[64] + b01[0] * g[60] + b00[0] * g[52];
+        g[69] = cpz[1] * g[65] + b01[1] * g[61] + b00[1] * g[53];
+        g[70] = cpz[2] * g[66] + b01[2] * g[62] + b00[2] * g[54];
+        g[71] = cpz[3] * g[67] + b01[3] * g[63] + b00[3] * g[55];
+}
+
+static inline void _srg0_2d4d_0200(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *b10 = bc->b10;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = c0x[0] * c0x[0] + b10[0];
+        g[9] = c0x[1] * c0x[1] + b10[1];
+        g[10] = c0x[2] * c0x[2] + b10[2];
+        g[11] = c0x[3] * c0x[3] + b10[3];
+        g[12] = 1;
+        g[13] = 1;
+        g[14] = 1;
+        g[15] = 1;
+        g[16] = c0y[0];
+        g[17] = c0y[1];
+        g[18] = c0y[2];
+        g[19] = c0y[3];
+        g[20] = c0y[0] * c0y[0] + b10[0];
+        g[21] = c0y[1] * c0y[1] + b10[1];
+        g[22] = c0y[2] * c0y[2] + b10[2];
+        g[23] = c0y[3] * c0y[3] + b10[3];
+        //g[24] = w[0];
+        //g[25] = w[0];
+        //g[26] = w[1];
+        //g[27] = w[1];
+        g[28] = c0z[0] * g[24];
+        g[29] = c0z[1] * g[25];
+        g[30] = c0z[2] * g[26];
+        g[31] = c0z[3] * g[27];
+        g[32] = c0z[0] * g[28] + b10[0] * g[24];
+        g[33] = c0z[1] * g[29] + b10[1] * g[25];
+        g[34] = c0z[2] * g[30] + b10[2] * g[26];
+        g[35] = c0z[3] * g[31] + b10[3] * g[27];
+}
+
+static inline void _srg0_2d4d_0201(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b10 = bc->b10;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[8] = c0x[0];
+        g[9] = c0x[1];
+        g[10] = c0x[2];
+        g[11] = c0x[3];
+        g[16] = c0x[0] * c0x[0] + b10[0];
+        g[17] = c0x[1] * c0x[1] + b10[1];
+        g[18] = c0x[2] * c0x[2] + b10[2];
+        g[19] = c0x[3] * c0x[3] + b10[3];
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[12] = cpx[0] * c0x[0] + b00[0];
+        g[13] = cpx[1] * c0x[1] + b00[1];
+        g[14] = cpx[2] * c0x[2] + b00[2];
+        g[15] = cpx[3] * c0x[3] + b00[3];
+        g[20] = c0x[0] * (g[12] + b00[0]) + b10[0] * cpx[0];
+        g[21] = c0x[1] * (g[13] + b00[1]) + b10[1] * cpx[1];
+        g[22] = c0x[2] * (g[14] + b00[2]) + b10[2] * cpx[2];
+        g[23] = c0x[3] * (g[15] + b00[3]) + b10[3] * cpx[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[32] = c0y[0];
+        g[33] = c0y[1];
+        g[34] = c0y[2];
+        g[35] = c0y[3];
+        g[40] = c0y[0] * c0y[0] + b10[0];
+        g[41] = c0y[1] * c0y[1] + b10[1];
+        g[42] = c0y[2] * c0y[2] + b10[2];
+        g[43] = c0y[3] * c0y[3] + b10[3];
+        g[28] = cpy[0];
+        g[29] = cpy[1];
+        g[30] = cpy[2];
+        g[31] = cpy[3];
+        g[36] = cpy[0] * c0y[0] + b00[0];
+        g[37] = cpy[1] * c0y[1] + b00[1];
+        g[38] = cpy[2] * c0y[2] + b00[2];
+        g[39] = cpy[3] * c0y[3] + b00[3];
+        g[44] = c0y[0] * (g[36] + b00[0]) + b10[0] * cpy[0];
+        g[45] = c0y[1] * (g[37] + b00[1]) + b10[1] * cpy[1];
+        g[46] = c0y[2] * (g[38] + b00[2]) + b10[2] * cpy[2];
+        g[47] = c0y[3] * (g[39] + b00[3]) + b10[3] * cpy[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[56] = c0z[0] * g[48];
+        g[57] = c0z[1] * g[49];
+        g[58] = c0z[2] * g[50];
+        g[59] = c0z[3] * g[51];
+        g[64] = c0z[0] * g[56] + b10[0] * g[48];
+        g[65] = c0z[1] * g[57] + b10[1] * g[49];
+        g[66] = c0z[2] * g[58] + b10[2] * g[50];
+        g[67] = c0z[3] * g[59] + b10[3] * g[51];
+        g[52] = cpz[0] * g[48];
+        g[53] = cpz[1] * g[49];
+        g[54] = cpz[2] * g[50];
+        g[55] = cpz[3] * g[51];
+        g[60] = cpz[0] * g[56] + b00[0] * g[48];
+        g[61] = cpz[1] * g[57] + b00[1] * g[49];
+        g[62] = cpz[2] * g[58] + b00[2] * g[50];
+        g[63] = cpz[3] * g[59] + b00[3] * g[51];
+        g[68] = c0z[0] * g[60] + b10[0] * g[52] + b00[0] * g[56];
+        g[69] = c0z[1] * g[61] + b10[1] * g[53] + b00[1] * g[57];
+        g[70] = c0z[2] * g[62] + b10[2] * g[54] + b00[2] * g[58];
+        g[71] = c0z[3] * g[63] + b10[3] * g[55] + b00[3] * g[59];
+}
+
+static inline void _srg0_2d4d_0210(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b10 = bc->b10;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = cpx[0];
+        g[5] = cpx[1];
+        g[6] = cpx[2];
+        g[7] = cpx[3];
+        g[8] = c0x[0];
+        g[9] = c0x[1];
+        g[10] = c0x[2];
+        g[11] = c0x[3];
+        g[12] = cpx[0] * c0x[0] + b00[0];
+        g[13] = cpx[1] * c0x[1] + b00[1];
+        g[14] = cpx[2] * c0x[2] + b00[2];
+        g[15] = cpx[3] * c0x[3] + b00[3];
+        g[16] = c0x[0] * c0x[0] + b10[0];
+        g[17] = c0x[1] * c0x[1] + b10[1];
+        g[18] = c0x[2] * c0x[2] + b10[2];
+        g[19] = c0x[3] * c0x[3] + b10[3];
+        g[20] = c0x[0] * (g[12] + b00[0]) + b10[0] * cpx[0];
+        g[21] = c0x[1] * (g[13] + b00[1]) + b10[1] * cpx[1];
+        g[22] = c0x[2] * (g[14] + b00[2]) + b10[2] * cpx[2];
+        g[23] = c0x[3] * (g[15] + b00[3]) + b10[3] * cpx[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[28] = cpy[0];
+        g[29] = cpy[1];
+        g[30] = cpy[2];
+        g[31] = cpy[3];
+        g[32] = c0y[0];
+        g[33] = c0y[1];
+        g[34] = c0y[2];
+        g[35] = c0y[3];
+        g[36] = cpy[0] * c0y[0] + b00[0];
+        g[37] = cpy[1] * c0y[1] + b00[1];
+        g[38] = cpy[2] * c0y[2] + b00[2];
+        g[39] = cpy[3] * c0y[3] + b00[3];
+        g[40] = c0y[0] * c0y[0] + b10[0];
+        g[41] = c0y[1] * c0y[1] + b10[1];
+        g[42] = c0y[2] * c0y[2] + b10[2];
+        g[43] = c0y[3] * c0y[3] + b10[3];
+        g[44] = c0y[0] * (g[36] + b00[0]) + b10[0] * cpy[0];
+        g[45] = c0y[1] * (g[37] + b00[1]) + b10[1] * cpy[1];
+        g[46] = c0y[2] * (g[38] + b00[2]) + b10[2] * cpy[2];
+        g[47] = c0y[3] * (g[39] + b00[3]) + b10[3] * cpy[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[52] = cpz[0] * g[48];
+        g[53] = cpz[1] * g[49];
+        g[54] = cpz[2] * g[50];
+        g[55] = cpz[3] * g[51];
+        g[56] = c0z[0] * g[48];
+        g[57] = c0z[1] * g[49];
+        g[58] = c0z[2] * g[50];
+        g[59] = c0z[3] * g[51];
+        g[60] = cpz[0] * g[56] + b00[0] * g[48];
+        g[61] = cpz[1] * g[57] + b00[1] * g[49];
+        g[62] = cpz[2] * g[58] + b00[2] * g[50];
+        g[63] = cpz[3] * g[59] + b00[3] * g[51];
+        g[64] = c0z[0] * g[56] + b10[0] * g[48];
+        g[65] = c0z[1] * g[57] + b10[1] * g[49];
+        g[66] = c0z[2] * g[58] + b10[2] * g[50];
+        g[67] = c0z[3] * g[59] + b10[3] * g[51];
+        g[68] = c0z[0] * g[60] + b10[0] * g[52] + b00[0] * g[56];
+        g[69] = c0z[1] * g[61] + b10[1] * g[53] + b00[1] * g[57];
+        g[70] = c0z[2] * g[62] + b10[2] * g[54] + b00[2] * g[58];
+        g[71] = c0z[3] * g[63] + b10[3] * g[55] + b00[3] * g[59];
+}
+
+static inline void _srg0_2d4d_0300(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b10 = bc->b10;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = c0x[0] * c0x[0] + b10[0];
+        g[9] = c0x[1] * c0x[1] + b10[1];
+        g[10] = c0x[2] * c0x[2] + b10[2];
+        g[11] = c0x[3] * c0x[3] + b10[3];
+        g[12] = c0x[0] * (g[8] + 2 * b10[0]);
+        g[13] = c0x[1] * (g[9] + 2 * b10[1]);
+        g[14] = c0x[2] * (g[10] + 2 * b10[2]);
+        g[15] = c0x[3] * (g[11] + 2 * b10[3]);
+        g[16] = 1;
+        g[17] = 1;
+        g[18] = 1;
+        g[19] = 1;
+        g[20] = c0y[0];
+        g[21] = c0y[1];
+        g[22] = c0y[2];
+        g[23] = c0y[3];
+        g[24] = c0y[0] * c0y[0] + b10[0];
+        g[25] = c0y[1] * c0y[1] + b10[1];
+        g[26] = c0y[2] * c0y[2] + b10[2];
+        g[27] = c0y[3] * c0y[3] + b10[3];
+        g[28] = c0y[0] * (g[24] + 2 * b10[0]);
+        g[29] = c0y[1] * (g[25] + 2 * b10[1]);
+        g[30] = c0y[2] * (g[26] + 2 * b10[2]);
+        g[31] = c0y[3] * (g[27] + 2 * b10[3]);
+        //g[32] = w[0];
+        //g[33] = w[0];
+        //g[34] = w[1];
+        //g[35] = w[1];
+        g[36] = c0z[0] * g[32];
+        g[37] = c0z[1] * g[33];
+        g[38] = c0z[2] * g[34];
+        g[39] = c0z[3] * g[35];
+        g[40] = c0z[0] * g[36] + b10[0] * g[32];
+        g[41] = c0z[1] * g[37] + b10[1] * g[33];
+        g[42] = c0z[2] * g[38] + b10[2] * g[34];
+        g[43] = c0z[3] * g[39] + b10[3] * g[35];
+        g[44] = c0z[0] * g[40] + 2 * b10[0] * g[36];
+        g[45] = c0z[1] * g[41] + 2 * b10[1] * g[37];
+        g[46] = c0z[2] * g[42] + 2 * b10[2] * g[38];
+        g[47] = c0z[3] * g[43] + 2 * b10[3] * g[39];
+}
+
+static inline void _srg0_2d4d_1000(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = c0x[0];
+        g[3] = c0x[1];
+        g[4] = 1;
+        g[5] = 1;
+        g[6] = c0y[0];
+        g[7] = c0y[1];
+        //g[8] = w[0];
+        //g[9] = w[0];
+        g[10] = c0z[0] * g[8];
+        g[11] = c0z[1] * g[9];
+}
+
+static inline void _srg0_2d4d_1001(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = cpx[0];
+        g[9] = cpx[1];
+        g[10] = cpx[2];
+        g[11] = cpx[3];
+        g[12] = cpx[0] * c0x[0] + b00[0];
+        g[13] = cpx[1] * c0x[1] + b00[1];
+        g[14] = cpx[2] * c0x[2] + b00[2];
+        g[15] = cpx[3] * c0x[3] + b00[3];
+        g[16] = 1;
+        g[17] = 1;
+        g[18] = 1;
+        g[19] = 1;
+        g[20] = c0y[0];
+        g[21] = c0y[1];
+        g[22] = c0y[2];
+        g[23] = c0y[3];
+        g[24] = cpy[0];
+        g[25] = cpy[1];
+        g[26] = cpy[2];
+        g[27] = cpy[3];
+        g[28] = cpy[0] * c0y[0] + b00[0];
+        g[29] = cpy[1] * c0y[1] + b00[1];
+        g[30] = cpy[2] * c0y[2] + b00[2];
+        g[31] = cpy[3] * c0y[3] + b00[3];
+        //g[32] = w[0];
+        //g[33] = w[0];
+        //g[34] = w[1];
+        //g[35] = w[1];
+        g[36] = c0z[0] * g[32];
+        g[37] = c0z[1] * g[33];
+        g[38] = c0z[2] * g[34];
+        g[39] = c0z[3] * g[35];
+        g[40] = cpz[0] * g[32];
+        g[41] = cpz[1] * g[33];
+        g[42] = cpz[2] * g[34];
+        g[43] = cpz[3] * g[35];
+        g[44] = cpz[0] * g[36] + b00[0] * g[32];
+        g[45] = cpz[1] * g[37] + b00[1] * g[33];
+        g[46] = cpz[2] * g[38] + b00[2] * g[34];
+        g[47] = cpz[3] * g[39] + b00[3] * g[35];
+}
+
+static inline void _srg0_2d4d_1002(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b01 = bc->b01;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = cpx[0];
+        g[9] = cpx[1];
+        g[10] = cpx[2];
+        g[11] = cpx[3];
+        g[12] = cpx[0] * c0x[0] + b00[0];
+        g[13] = cpx[1] * c0x[1] + b00[1];
+        g[14] = cpx[2] * c0x[2] + b00[2];
+        g[15] = cpx[3] * c0x[3] + b00[3];
+        g[16] = cpx[0] * cpx[0] + b01[0];
+        g[17] = cpx[1] * cpx[1] + b01[1];
+        g[18] = cpx[2] * cpx[2] + b01[2];
+        g[19] = cpx[3] * cpx[3] + b01[3];
+        g[20] = cpx[0] * (g[12] + b00[0]) + b01[0] * c0x[0];
+        g[21] = cpx[1] * (g[13] + b00[1]) + b01[1] * c0x[1];
+        g[22] = cpx[2] * (g[14] + b00[2]) + b01[2] * c0x[2];
+        g[23] = cpx[3] * (g[15] + b00[3]) + b01[3] * c0x[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[28] = c0y[0];
+        g[29] = c0y[1];
+        g[30] = c0y[2];
+        g[31] = c0y[3];
+        g[32] = cpy[0];
+        g[33] = cpy[1];
+        g[34] = cpy[2];
+        g[35] = cpy[3];
+        g[36] = cpy[0] * c0y[0] + b00[0];
+        g[37] = cpy[1] * c0y[1] + b00[1];
+        g[38] = cpy[2] * c0y[2] + b00[2];
+        g[39] = cpy[3] * c0y[3] + b00[3];
+        g[40] = cpy[0] * cpy[0] + b01[0];
+        g[41] = cpy[1] * cpy[1] + b01[1];
+        g[42] = cpy[2] * cpy[2] + b01[2];
+        g[43] = cpy[3] * cpy[3] + b01[3];
+        g[44] = cpy[0] * (g[36] + b00[0]) + b01[0] * c0y[0];
+        g[45] = cpy[1] * (g[37] + b00[1]) + b01[1] * c0y[1];
+        g[46] = cpy[2] * (g[38] + b00[2]) + b01[2] * c0y[2];
+        g[47] = cpy[3] * (g[39] + b00[3]) + b01[3] * c0y[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[52] = c0z[0] * g[48];
+        g[53] = c0z[1] * g[49];
+        g[54] = c0z[2] * g[50];
+        g[55] = c0z[3] * g[51];
+        g[56] = cpz[0] * g[48];
+        g[57] = cpz[1] * g[49];
+        g[58] = cpz[2] * g[50];
+        g[59] = cpz[3] * g[51];
+        g[60] = cpz[0] * g[52] + b00[0] * g[48];
+        g[61] = cpz[1] * g[53] + b00[1] * g[49];
+        g[62] = cpz[2] * g[54] + b00[2] * g[50];
+        g[63] = cpz[3] * g[55] + b00[3] * g[51];
+        g[64] = cpz[0] * g[56] + b01[0] * g[48];
+        g[65] = cpz[1] * g[57] + b01[1] * g[49];
+        g[66] = cpz[2] * g[58] + b01[2] * g[50];
+        g[67] = cpz[3] * g[59] + b01[3] * g[51];
+        g[68] = cpz[0] * g[60] + b01[0] * g[52] + b00[0] * g[56];
+        g[69] = cpz[1] * g[61] + b01[1] * g[53] + b00[1] * g[57];
+        g[70] = cpz[2] * g[62] + b01[2] * g[54] + b00[2] * g[58];
+        g[71] = cpz[3] * g[63] + b01[3] * g[55] + b00[3] * g[59];
+}
+
+static inline void _srg0_2d4d_1010(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = cpx[0];
+        g[9] = cpx[1];
+        g[10] = cpx[2];
+        g[11] = cpx[3];
+        g[12] = cpx[0] * c0x[0] + b00[0];
+        g[13] = cpx[1] * c0x[1] + b00[1];
+        g[14] = cpx[2] * c0x[2] + b00[2];
+        g[15] = cpx[3] * c0x[3] + b00[3];
+        g[16] = 1;
+        g[17] = 1;
+        g[18] = 1;
+        g[19] = 1;
+        g[20] = c0y[0];
+        g[21] = c0y[1];
+        g[22] = c0y[2];
+        g[23] = c0y[3];
+        g[24] = cpy[0];
+        g[25] = cpy[1];
+        g[26] = cpy[2];
+        g[27] = cpy[3];
+        g[28] = cpy[0] * c0y[0] + b00[0];
+        g[29] = cpy[1] * c0y[1] + b00[1];
+        g[30] = cpy[2] * c0y[2] + b00[2];
+        g[31] = cpy[3] * c0y[3] + b00[3];
+        //g[32] = w[0];
+        //g[33] = w[0];
+        //g[34] = w[1];
+        //g[35] = w[1];
+        g[36] = c0z[0] * g[32];
+        g[37] = c0z[1] * g[33];
+        g[38] = c0z[2] * g[34];
+        g[39] = c0z[3] * g[35];
+        g[40] = cpz[0] * g[32];
+        g[41] = cpz[1] * g[33];
+        g[42] = cpz[2] * g[34];
+        g[43] = cpz[3] * g[35];
+        g[44] = cpz[0] * g[36] + b00[0] * g[32];
+        g[45] = cpz[1] * g[37] + b00[1] * g[33];
+        g[46] = cpz[2] * g[38] + b00[2] * g[34];
+        g[47] = cpz[3] * g[39] + b00[3] * g[35];
+}
+
+static inline void _srg0_2d4d_1011(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b01 = bc->b01;
+        double xkxl = envs->rkrl[0];
+        double ykyl = envs->rkrl[1];
+        double zkzl = envs->rkrl[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[16] = cpx[0];
+        g[17] = cpx[1];
+        g[18] = cpx[2];
+        g[19] = cpx[3];
+        g[20] = cpx[0] * c0x[0] + b00[0];
+        g[21] = cpx[1] * c0x[1] + b00[1];
+        g[22] = cpx[2] * c0x[2] + b00[2];
+        g[23] = cpx[3] * c0x[3] + b00[3];
+        g[24] = cpx[0] * (xkxl + cpx[0]) + b01[0];
+        g[25] = cpx[1] * (xkxl + cpx[1]) + b01[1];
+        g[26] = cpx[2] * (xkxl + cpx[2]) + b01[2];
+        g[27] = cpx[3] * (xkxl + cpx[3]) + b01[3];
+        g[28] = g[20] * (xkxl + cpx[0]) + cpx[0] * b00[0] + b01[0] * c0x[0];
+        g[29] = g[21] * (xkxl + cpx[1]) + cpx[1] * b00[1] + b01[1] * c0x[1];
+        g[30] = g[22] * (xkxl + cpx[2]) + cpx[2] * b00[2] + b01[2] * c0x[2];
+        g[31] = g[23] * (xkxl + cpx[3]) + cpx[3] * b00[3] + b01[3] * c0x[3];
+        g[8] = xkxl + cpx[0];
+        g[9] = xkxl + cpx[1];
+        g[10] = xkxl + cpx[2];
+        g[11] = xkxl + cpx[3];
+        g[12] = c0x[0] * (xkxl + cpx[0]) + b00[0];
+        g[13] = c0x[1] * (xkxl + cpx[1]) + b00[1];
+        g[14] = c0x[2] * (xkxl + cpx[2]) + b00[2];
+        g[15] = c0x[3] * (xkxl + cpx[3]) + b00[3];
+        g[48] = 1;
+        g[49] = 1;
+        g[50] = 1;
+        g[51] = 1;
+        g[52] = c0y[0];
+        g[53] = c0y[1];
+        g[54] = c0y[2];
+        g[55] = c0y[3];
+        g[64] = cpy[0];
+        g[65] = cpy[1];
+        g[66] = cpy[2];
+        g[67] = cpy[3];
+        g[68] = cpy[0] * c0y[0] + b00[0];
+        g[69] = cpy[1] * c0y[1] + b00[1];
+        g[70] = cpy[2] * c0y[2] + b00[2];
+        g[71] = cpy[3] * c0y[3] + b00[3];
+        g[72] = cpy[0] * (ykyl + cpy[0]) + b01[0];
+        g[73] = cpy[1] * (ykyl + cpy[1]) + b01[1];
+        g[74] = cpy[2] * (ykyl + cpy[2]) + b01[2];
+        g[75] = cpy[3] * (ykyl + cpy[3]) + b01[3];
+        g[76] = g[68] * (ykyl + cpy[0]) + cpy[0] * b00[0] + b01[0] * c0y[0];
+        g[77] = g[69] * (ykyl + cpy[1]) + cpy[1] * b00[1] + b01[1] * c0y[1];
+        g[78] = g[70] * (ykyl + cpy[2]) + cpy[2] * b00[2] + b01[2] * c0y[2];
+        g[79] = g[71] * (ykyl + cpy[3]) + cpy[3] * b00[3] + b01[3] * c0y[3];
+        g[56] = ykyl + cpy[0];
+        g[57] = ykyl + cpy[1];
+        g[58] = ykyl + cpy[2];
+        g[59] = ykyl + cpy[3];
+        g[60] = c0y[0] * (ykyl + cpy[0]) + b00[0];
+        g[61] = c0y[1] * (ykyl + cpy[1]) + b00[1];
+        g[62] = c0y[2] * (ykyl + cpy[2]) + b00[2];
+        g[63] = c0y[3] * (ykyl + cpy[3]) + b00[3];
+        //g[96] = w[0];
+        //g[97] = w[0];
+        //g[98] = w[1];
+        //g[99] = w[1];
+        g[100] = c0z[0] * g[96];
+        g[101] = c0z[1] * g[97];
+        g[102] = c0z[2] * g[98];
+        g[103] = c0z[3] * g[99];
+        g[112] = cpz[0] * g[96];
+        g[113] = cpz[1] * g[97];
+        g[114] = cpz[2] * g[98];
+        g[115] = cpz[3] * g[99];
+        g[116] = cpz[0] * g[100] + b00[0] * g[96];
+        g[117] = cpz[1] * g[101] + b00[1] * g[97];
+        g[118] = cpz[2] * g[102] + b00[2] * g[98];
+        g[119] = cpz[3] * g[103] + b00[3] * g[99];
+        g[120] = g[112] * (zkzl + cpz[0]) + b01[0] * g[96];
+        g[121] = g[113] * (zkzl + cpz[1]) + b01[1] * g[97];
+        g[122] = g[114] * (zkzl + cpz[2]) + b01[2] * g[98];
+        g[123] = g[115] * (zkzl + cpz[3]) + b01[3] * g[99];
+        g[124] = g[116] * (zkzl + cpz[0]) + b01[0] * g[100] + b00[0] * g[112];
+        g[125] = g[117] * (zkzl + cpz[1]) + b01[1] * g[101] + b00[1] * g[113];
+        g[126] = g[118] * (zkzl + cpz[2]) + b01[2] * g[102] + b00[2] * g[114];
+        g[127] = g[119] * (zkzl + cpz[3]) + b01[3] * g[103] + b00[3] * g[115];
+        g[104] = g[96] * (zkzl + cpz[0]);
+        g[105] = g[97] * (zkzl + cpz[1]);
+        g[106] = g[98] * (zkzl + cpz[2]);
+        g[107] = g[99] * (zkzl + cpz[3]);
+        g[108] = g[100] * (zkzl + cpz[0]) + b00[0] * g[96];
+        g[109] = g[101] * (zkzl + cpz[1]) + b00[1] * g[97];
+        g[110] = g[102] * (zkzl + cpz[2]) + b00[2] * g[98];
+        g[111] = g[103] * (zkzl + cpz[3]) + b00[3] * g[99];
+}
+
+static inline void _srg0_2d4d_1020(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b01 = bc->b01;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = cpx[0];
+        g[9] = cpx[1];
+        g[10] = cpx[2];
+        g[11] = cpx[3];
+        g[12] = cpx[0] * c0x[0] + b00[0];
+        g[13] = cpx[1] * c0x[1] + b00[1];
+        g[14] = cpx[2] * c0x[2] + b00[2];
+        g[15] = cpx[3] * c0x[3] + b00[3];
+        g[16] = cpx[0] * cpx[0] + b01[0];
+        g[17] = cpx[1] * cpx[1] + b01[1];
+        g[18] = cpx[2] * cpx[2] + b01[2];
+        g[19] = cpx[3] * cpx[3] + b01[3];
+        g[20] = cpx[0] * (g[12] + b00[0]) + b01[0] * c0x[0];
+        g[21] = cpx[1] * (g[13] + b00[1]) + b01[1] * c0x[1];
+        g[22] = cpx[2] * (g[14] + b00[2]) + b01[2] * c0x[2];
+        g[23] = cpx[3] * (g[15] + b00[3]) + b01[3] * c0x[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[28] = c0y[0];
+        g[29] = c0y[1];
+        g[30] = c0y[2];
+        g[31] = c0y[3];
+        g[32] = cpy[0];
+        g[33] = cpy[1];
+        g[34] = cpy[2];
+        g[35] = cpy[3];
+        g[36] = cpy[0] * c0y[0] + b00[0];
+        g[37] = cpy[1] * c0y[1] + b00[1];
+        g[38] = cpy[2] * c0y[2] + b00[2];
+        g[39] = cpy[3] * c0y[3] + b00[3];
+        g[40] = cpy[0] * cpy[0] + b01[0];
+        g[41] = cpy[1] * cpy[1] + b01[1];
+        g[42] = cpy[2] * cpy[2] + b01[2];
+        g[43] = cpy[3] * cpy[3] + b01[3];
+        g[44] = cpy[0] * (g[36] + b00[0]) + b01[0] * c0y[0];
+        g[45] = cpy[1] * (g[37] + b00[1]) + b01[1] * c0y[1];
+        g[46] = cpy[2] * (g[38] + b00[2]) + b01[2] * c0y[2];
+        g[47] = cpy[3] * (g[39] + b00[3]) + b01[3] * c0y[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[52] = c0z[0] * g[48];
+        g[53] = c0z[1] * g[49];
+        g[54] = c0z[2] * g[50];
+        g[55] = c0z[3] * g[51];
+        g[56] = cpz[0] * g[48];
+        g[57] = cpz[1] * g[49];
+        g[58] = cpz[2] * g[50];
+        g[59] = cpz[3] * g[51];
+        g[60] = cpz[0] * g[52] + b00[0] * g[48];
+        g[61] = cpz[1] * g[53] + b00[1] * g[49];
+        g[62] = cpz[2] * g[54] + b00[2] * g[50];
+        g[63] = cpz[3] * g[55] + b00[3] * g[51];
+        g[64] = cpz[0] * g[56] + b01[0] * g[48];
+        g[65] = cpz[1] * g[57] + b01[1] * g[49];
+        g[66] = cpz[2] * g[58] + b01[2] * g[50];
+        g[67] = cpz[3] * g[59] + b01[3] * g[51];
+        g[68] = cpz[0] * g[60] + b01[0] * g[52] + b00[0] * g[56];
+        g[69] = cpz[1] * g[61] + b01[1] * g[53] + b00[1] * g[57];
+        g[70] = cpz[2] * g[62] + b01[2] * g[54] + b00[2] * g[58];
+        g[71] = cpz[3] * g[63] + b01[3] * g[55] + b00[3] * g[59];
+}
+
+static inline void _srg0_2d4d_1100(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *b10 = bc->b10;
+        double xixj = envs->rirj[0];
+        double yiyj = envs->rirj[1];
+        double zizj = envs->rirj[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[8] = c0x[0];
+        g[9] = c0x[1];
+        g[10] = c0x[2];
+        g[11] = c0x[3];
+        g[12] = c0x[0] * (xixj + c0x[0]) + b10[0];
+        g[13] = c0x[1] * (xixj + c0x[1]) + b10[1];
+        g[14] = c0x[2] * (xixj + c0x[2]) + b10[2];
+        g[15] = c0x[3] * (xixj + c0x[3]) + b10[3];
+        g[4] = xixj + c0x[0];
+        g[5] = xixj + c0x[1];
+        g[6] = xixj + c0x[2];
+        g[7] = xixj + c0x[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[32] = c0y[0];
+        g[33] = c0y[1];
+        g[34] = c0y[2];
+        g[35] = c0y[3];
+        g[36] = c0y[0] * (yiyj + c0y[0]) + b10[0];
+        g[37] = c0y[1] * (yiyj + c0y[1]) + b10[1];
+        g[38] = c0y[2] * (yiyj + c0y[2]) + b10[2];
+        g[39] = c0y[3] * (yiyj + c0y[3]) + b10[3];
+        g[28] = yiyj + c0y[0];
+        g[29] = yiyj + c0y[1];
+        g[30] = yiyj + c0y[2];
+        g[31] = yiyj + c0y[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[56] = c0z[0] * g[48];
+        g[57] = c0z[1] * g[49];
+        g[58] = c0z[2] * g[50];
+        g[59] = c0z[3] * g[51];
+        g[60] = g[56] * (zizj + c0z[0]) + b10[0] * g[48];
+        g[61] = g[57] * (zizj + c0z[1]) + b10[1] * g[49];
+        g[62] = g[58] * (zizj + c0z[2]) + b10[2] * g[50];
+        g[63] = g[59] * (zizj + c0z[3]) + b10[3] * g[51];
+        g[52] = g[48] * (zizj + c0z[0]);
+        g[53] = g[49] * (zizj + c0z[1]);
+        g[54] = g[50] * (zizj + c0z[2]);
+        g[55] = g[51] * (zizj + c0z[3]);
+}
+
+static inline void _srg0_2d4d_1101(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b10 = bc->b10;
+        double xixj = envs->rirj[0];
+        double yiyj = envs->rirj[1];
+        double zizj = envs->rirj[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[16] = c0x[0];
+        g[17] = c0x[1];
+        g[18] = c0x[2];
+        g[19] = c0x[3];
+        g[8] = cpx[0];
+        g[9] = cpx[1];
+        g[10] = cpx[2];
+        g[11] = cpx[3];
+        g[24] = cpx[0] * c0x[0] + b00[0];
+        g[25] = cpx[1] * c0x[1] + b00[1];
+        g[26] = cpx[2] * c0x[2] + b00[2];
+        g[27] = cpx[3] * c0x[3] + b00[3];
+        g[20] = c0x[0] * (xixj + c0x[0]) + b10[0];
+        g[21] = c0x[1] * (xixj + c0x[1]) + b10[1];
+        g[22] = c0x[2] * (xixj + c0x[2]) + b10[2];
+        g[23] = c0x[3] * (xixj + c0x[3]) + b10[3];
+        g[4] = xixj + c0x[0];
+        g[5] = xixj + c0x[1];
+        g[6] = xixj + c0x[2];
+        g[7] = xixj + c0x[3];
+        g[28] = g[24] * (xixj + c0x[0]) + c0x[0] * b00[0] + b10[0] * cpx[0];
+        g[29] = g[25] * (xixj + c0x[1]) + c0x[1] * b00[1] + b10[1] * cpx[1];
+        g[30] = g[26] * (xixj + c0x[2]) + c0x[2] * b00[2] + b10[2] * cpx[2];
+        g[31] = g[27] * (xixj + c0x[3]) + c0x[3] * b00[3] + b10[3] * cpx[3];
+        g[12] = cpx[0] * (xixj + c0x[0]) + b00[0];
+        g[13] = cpx[1] * (xixj + c0x[1]) + b00[1];
+        g[14] = cpx[2] * (xixj + c0x[2]) + b00[2];
+        g[15] = cpx[3] * (xixj + c0x[3]) + b00[3];
+        g[48] = 1;
+        g[49] = 1;
+        g[50] = 1;
+        g[51] = 1;
+        g[64] = c0y[0];
+        g[65] = c0y[1];
+        g[66] = c0y[2];
+        g[67] = c0y[3];
+        g[56] = cpy[0];
+        g[57] = cpy[1];
+        g[58] = cpy[2];
+        g[59] = cpy[3];
+        g[72] = cpy[0] * c0y[0] + b00[0];
+        g[73] = cpy[1] * c0y[1] + b00[1];
+        g[74] = cpy[2] * c0y[2] + b00[2];
+        g[75] = cpy[3] * c0y[3] + b00[3];
+        g[68] = c0y[0] * (yiyj + c0y[0]) + b10[0];
+        g[69] = c0y[1] * (yiyj + c0y[1]) + b10[1];
+        g[70] = c0y[2] * (yiyj + c0y[2]) + b10[2];
+        g[71] = c0y[3] * (yiyj + c0y[3]) + b10[3];
+        g[52] = yiyj + c0y[0];
+        g[53] = yiyj + c0y[1];
+        g[54] = yiyj + c0y[2];
+        g[55] = yiyj + c0y[3];
+        g[76] = g[72] * (yiyj + c0y[0]) + c0y[0] * b00[0] + b10[0] * cpy[0];
+        g[77] = g[73] * (yiyj + c0y[1]) + c0y[1] * b00[1] + b10[1] * cpy[1];
+        g[78] = g[74] * (yiyj + c0y[2]) + c0y[2] * b00[2] + b10[2] * cpy[2];
+        g[79] = g[75] * (yiyj + c0y[3]) + c0y[3] * b00[3] + b10[3] * cpy[3];
+        g[60] = cpy[0] * (yiyj + c0y[0]) + b00[0];
+        g[61] = cpy[1] * (yiyj + c0y[1]) + b00[1];
+        g[62] = cpy[2] * (yiyj + c0y[2]) + b00[2];
+        g[63] = cpy[3] * (yiyj + c0y[3]) + b00[3];
+        //g[96] = w[0];
+        //g[97] = w[0];
+        //g[98] = w[1];
+        //g[99] = w[1];
+        g[112] = c0z[0] * g[96];
+        g[113] = c0z[1] * g[97];
+        g[114] = c0z[2] * g[98];
+        g[115] = c0z[3] * g[99];
+        g[104] = cpz[0] * g[96];
+        g[105] = cpz[1] * g[97];
+        g[106] = cpz[2] * g[98];
+        g[107] = cpz[3] * g[99];
+        g[120] = cpz[0] * g[112] + b00[0] * g[96];
+        g[121] = cpz[1] * g[113] + b00[1] * g[97];
+        g[122] = cpz[2] * g[114] + b00[2] * g[98];
+        g[123] = cpz[3] * g[115] + b00[3] * g[99];
+        g[116] = g[112] * (zizj + c0z[0]) + b10[0] * g[96];
+        g[117] = g[113] * (zizj + c0z[1]) + b10[1] * g[97];
+        g[118] = g[114] * (zizj + c0z[2]) + b10[2] * g[98];
+        g[119] = g[115] * (zizj + c0z[3]) + b10[3] * g[99];
+        g[100] = g[96] * (zizj + c0z[0]);
+        g[101] = g[97] * (zizj + c0z[1]);
+        g[102] = g[98] * (zizj + c0z[2]);
+        g[103] = g[99] * (zizj + c0z[3]);
+        g[124] = g[120] * (zizj + c0z[0]) + b10[0] * g[104] + b00[0] * g[112];
+        g[125] = g[121] * (zizj + c0z[1]) + b10[1] * g[105] + b00[1] * g[113];
+        g[126] = g[122] * (zizj + c0z[2]) + b10[2] * g[106] + b00[2] * g[114];
+        g[127] = g[123] * (zizj + c0z[3]) + b10[3] * g[107] + b00[3] * g[115];
+        g[108] = zizj * g[104] + cpz[0] * g[112] + b00[0] * g[96];
+        g[109] = zizj * g[105] + cpz[1] * g[113] + b00[1] * g[97];
+        g[110] = zizj * g[106] + cpz[2] * g[114] + b00[2] * g[98];
+        g[111] = zizj * g[107] + cpz[3] * g[115] + b00[3] * g[99];
+}
+
+static inline void _srg0_2d4d_1110(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b10 = bc->b10;
+        double xixj = envs->rirj[0];
+        double yiyj = envs->rirj[1];
+        double zizj = envs->rirj[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[16] = c0x[0];
+        g[17] = c0x[1];
+        g[18] = c0x[2];
+        g[19] = c0x[3];
+        g[8] = cpx[0];
+        g[9] = cpx[1];
+        g[10] = cpx[2];
+        g[11] = cpx[3];
+        g[24] = cpx[0] * c0x[0] + b00[0];
+        g[25] = cpx[1] * c0x[1] + b00[1];
+        g[26] = cpx[2] * c0x[2] + b00[2];
+        g[27] = cpx[3] * c0x[3] + b00[3];
+        g[20] = c0x[0] * (xixj + c0x[0]) + b10[0];
+        g[21] = c0x[1] * (xixj + c0x[1]) + b10[1];
+        g[22] = c0x[2] * (xixj + c0x[2]) + b10[2];
+        g[23] = c0x[3] * (xixj + c0x[3]) + b10[3];
+        g[4] = xixj + c0x[0];
+        g[5] = xixj + c0x[1];
+        g[6] = xixj + c0x[2];
+        g[7] = xixj + c0x[3];
+        g[28] = g[24] * (xixj + c0x[0]) + c0x[0] * b00[0] + b10[0] * cpx[0];
+        g[29] = g[25] * (xixj + c0x[1]) + c0x[1] * b00[1] + b10[1] * cpx[1];
+        g[30] = g[26] * (xixj + c0x[2]) + c0x[2] * b00[2] + b10[2] * cpx[2];
+        g[31] = g[27] * (xixj + c0x[3]) + c0x[3] * b00[3] + b10[3] * cpx[3];
+        g[12] = cpx[0] * (xixj + c0x[0]) + b00[0];
+        g[13] = cpx[1] * (xixj + c0x[1]) + b00[1];
+        g[14] = cpx[2] * (xixj + c0x[2]) + b00[2];
+        g[15] = cpx[3] * (xixj + c0x[3]) + b00[3];
+        g[48] = 1;
+        g[49] = 1;
+        g[50] = 1;
+        g[51] = 1;
+        g[64] = c0y[0];
+        g[65] = c0y[1];
+        g[66] = c0y[2];
+        g[67] = c0y[3];
+        g[56] = cpy[0];
+        g[57] = cpy[1];
+        g[58] = cpy[2];
+        g[59] = cpy[3];
+        g[72] = cpy[0] * c0y[0] + b00[0];
+        g[73] = cpy[1] * c0y[1] + b00[1];
+        g[74] = cpy[2] * c0y[2] + b00[2];
+        g[75] = cpy[3] * c0y[3] + b00[3];
+        g[68] = c0y[0] * (yiyj + c0y[0]) + b10[0];
+        g[69] = c0y[1] * (yiyj + c0y[1]) + b10[1];
+        g[70] = c0y[2] * (yiyj + c0y[2]) + b10[2];
+        g[71] = c0y[3] * (yiyj + c0y[3]) + b10[3];
+        g[52] = yiyj + c0y[0];
+        g[53] = yiyj + c0y[1];
+        g[54] = yiyj + c0y[2];
+        g[55] = yiyj + c0y[3];
+        g[76] = g[72] * (yiyj + c0y[0]) + c0y[0] * b00[0] + b10[0] * cpy[0];
+        g[77] = g[73] * (yiyj + c0y[1]) + c0y[1] * b00[1] + b10[1] * cpy[1];
+        g[78] = g[74] * (yiyj + c0y[2]) + c0y[2] * b00[2] + b10[2] * cpy[2];
+        g[79] = g[75] * (yiyj + c0y[3]) + c0y[3] * b00[3] + b10[3] * cpy[3];
+        g[60] = cpy[0] * (yiyj + c0y[0]) + b00[0];
+        g[61] = cpy[1] * (yiyj + c0y[1]) + b00[1];
+        g[62] = cpy[2] * (yiyj + c0y[2]) + b00[2];
+        g[63] = cpy[3] * (yiyj + c0y[3]) + b00[3];
+        //g[96] = w[0];
+        //g[97] = w[0];
+        //g[98] = w[1];
+        //g[99] = w[1];
+        g[112] = c0z[0] * g[96];
+        g[113] = c0z[1] * g[97];
+        g[114] = c0z[2] * g[98];
+        g[115] = c0z[3] * g[99];
+        g[104] = cpz[0] * g[96];
+        g[105] = cpz[1] * g[97];
+        g[106] = cpz[2] * g[98];
+        g[107] = cpz[3] * g[99];
+        g[120] = cpz[0] * g[112] + b00[0] * g[96];
+        g[121] = cpz[1] * g[113] + b00[1] * g[97];
+        g[122] = cpz[2] * g[114] + b00[2] * g[98];
+        g[123] = cpz[3] * g[115] + b00[3] * g[99];
+        g[116] = g[112] * (zizj + c0z[0]) + b10[0] * g[96];
+        g[117] = g[113] * (zizj + c0z[1]) + b10[1] * g[97];
+        g[118] = g[114] * (zizj + c0z[2]) + b10[2] * g[98];
+        g[119] = g[115] * (zizj + c0z[3]) + b10[3] * g[99];
+        g[100] = g[96] * (zizj + c0z[0]);
+        g[101] = g[97] * (zizj + c0z[1]);
+        g[102] = g[98] * (zizj + c0z[2]);
+        g[103] = g[99] * (zizj + c0z[3]);
+        g[124] = g[120] * (zizj + c0z[0]) + b10[0] * g[104] + b00[0] * g[112];
+        g[125] = g[121] * (zizj + c0z[1]) + b10[1] * g[105] + b00[1] * g[113];
+        g[126] = g[122] * (zizj + c0z[2]) + b10[2] * g[106] + b00[2] * g[114];
+        g[127] = g[123] * (zizj + c0z[3]) + b10[3] * g[107] + b00[3] * g[115];
+        g[108] = zizj * g[104] + cpz[0] * g[112] + b00[0] * g[96];
+        g[109] = zizj * g[105] + cpz[1] * g[113] + b00[1] * g[97];
+        g[110] = zizj * g[106] + cpz[2] * g[114] + b00[2] * g[98];
+        g[111] = zizj * g[107] + cpz[3] * g[115] + b00[3] * g[99];
+}
+
+static inline void _srg0_2d4d_1200(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *b10 = bc->b10;
+        double xixj = envs->rirj[0];
+        double yiyj = envs->rirj[1];
+        double zizj = envs->rirj[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[8] = c0x[0];
+        g[9] = c0x[1];
+        g[10] = c0x[2];
+        g[11] = c0x[3];
+        g[16] = c0x[0] * c0x[0] + b10[0];
+        g[17] = c0x[1] * c0x[1] + b10[1];
+        g[18] = c0x[2] * c0x[2] + b10[2];
+        g[19] = c0x[3] * c0x[3] + b10[3];
+        g[20] = g[16] * (xixj + c0x[0]) + c0x[0] * 2 * b10[0];
+        g[21] = g[17] * (xixj + c0x[1]) + c0x[1] * 2 * b10[1];
+        g[22] = g[18] * (xixj + c0x[2]) + c0x[2] * 2 * b10[2];
+        g[23] = g[19] * (xixj + c0x[3]) + c0x[3] * 2 * b10[3];
+        g[12] = c0x[0] * (xixj + c0x[0]) + b10[0];
+        g[13] = c0x[1] * (xixj + c0x[1]) + b10[1];
+        g[14] = c0x[2] * (xixj + c0x[2]) + b10[2];
+        g[15] = c0x[3] * (xixj + c0x[3]) + b10[3];
+        g[4] = xixj + c0x[0];
+        g[5] = xixj + c0x[1];
+        g[6] = xixj + c0x[2];
+        g[7] = xixj + c0x[3];
+        g[32] = 1;
+        g[33] = 1;
+        g[34] = 1;
+        g[35] = 1;
+        g[40] = c0y[0];
+        g[41] = c0y[1];
+        g[42] = c0y[2];
+        g[43] = c0y[3];
+        g[48] = c0y[0] * c0y[0] + b10[0];
+        g[49] = c0y[1] * c0y[1] + b10[1];
+        g[50] = c0y[2] * c0y[2] + b10[2];
+        g[51] = c0y[3] * c0y[3] + b10[3];
+        g[52] = g[48] * (yiyj + c0y[0]) + c0y[0] * 2 * b10[0];
+        g[53] = g[49] * (yiyj + c0y[1]) + c0y[1] * 2 * b10[1];
+        g[54] = g[50] * (yiyj + c0y[2]) + c0y[2] * 2 * b10[2];
+        g[55] = g[51] * (yiyj + c0y[3]) + c0y[3] * 2 * b10[3];
+        g[44] = c0y[0] * (yiyj + c0y[0]) + b10[0];
+        g[45] = c0y[1] * (yiyj + c0y[1]) + b10[1];
+        g[46] = c0y[2] * (yiyj + c0y[2]) + b10[2];
+        g[47] = c0y[3] * (yiyj + c0y[3]) + b10[3];
+        g[36] = yiyj + c0y[0];
+        g[37] = yiyj + c0y[1];
+        g[38] = yiyj + c0y[2];
+        g[39] = yiyj + c0y[3];
+        //g[64] = w[0];
+        //g[65] = w[0];
+        //g[66] = w[1];
+        //g[67] = w[1];
+        g[72] = c0z[0] * g[64];
+        g[73] = c0z[1] * g[65];
+        g[74] = c0z[2] * g[66];
+        g[75] = c0z[3] * g[67];
+        g[80] = c0z[0] * g[72] + b10[0] * g[64];
+        g[81] = c0z[1] * g[73] + b10[1] * g[65];
+        g[82] = c0z[2] * g[74] + b10[2] * g[66];
+        g[83] = c0z[3] * g[75] + b10[3] * g[67];
+        g[84] = g[80] * (zizj + c0z[0]) + 2 * b10[0] * g[72];
+        g[85] = g[81] * (zizj + c0z[1]) + 2 * b10[1] * g[73];
+        g[86] = g[82] * (zizj + c0z[2]) + 2 * b10[2] * g[74];
+        g[87] = g[83] * (zizj + c0z[3]) + 2 * b10[3] * g[75];
+        g[76] = g[72] * (zizj + c0z[0]) + b10[0] * g[64];
+        g[77] = g[73] * (zizj + c0z[1]) + b10[1] * g[65];
+        g[78] = g[74] * (zizj + c0z[2]) + b10[2] * g[66];
+        g[79] = g[75] * (zizj + c0z[3]) + b10[3] * g[67];
+        g[68] = g[64] * (zizj + c0z[0]);
+        g[69] = g[65] * (zizj + c0z[1]);
+        g[70] = g[66] * (zizj + c0z[2]);
+        g[71] = g[67] * (zizj + c0z[3]);
+}
+
+static inline void _srg0_2d4d_2000(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *b10 = bc->b10;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = c0x[0] * c0x[0] + b10[0];
+        g[9] = c0x[1] * c0x[1] + b10[1];
+        g[10] = c0x[2] * c0x[2] + b10[2];
+        g[11] = c0x[3] * c0x[3] + b10[3];
+        g[12] = 1;
+        g[13] = 1;
+        g[14] = 1;
+        g[15] = 1;
+        g[16] = c0y[0];
+        g[17] = c0y[1];
+        g[18] = c0y[2];
+        g[19] = c0y[3];
+        g[20] = c0y[0] * c0y[0] + b10[0];
+        g[21] = c0y[1] * c0y[1] + b10[1];
+        g[22] = c0y[2] * c0y[2] + b10[2];
+        g[23] = c0y[3] * c0y[3] + b10[3];
+        //g[24] = w[0];
+        //g[25] = w[0];
+        //g[26] = w[1];
+        //g[27] = w[1];
+        g[28] = c0z[0] * g[24];
+        g[29] = c0z[1] * g[25];
+        g[30] = c0z[2] * g[26];
+        g[31] = c0z[3] * g[27];
+        g[32] = c0z[0] * g[28] + b10[0] * g[24];
+        g[33] = c0z[1] * g[29] + b10[1] * g[25];
+        g[34] = c0z[2] * g[30] + b10[2] * g[26];
+        g[35] = c0z[3] * g[31] + b10[3] * g[27];
+}
+
+static inline void _srg0_2d4d_2001(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b10 = bc->b10;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = c0x[0] * c0x[0] + b10[0];
+        g[9] = c0x[1] * c0x[1] + b10[1];
+        g[10] = c0x[2] * c0x[2] + b10[2];
+        g[11] = c0x[3] * c0x[3] + b10[3];
+        g[12] = cpx[0];
+        g[13] = cpx[1];
+        g[14] = cpx[2];
+        g[15] = cpx[3];
+        g[16] = cpx[0] * c0x[0] + b00[0];
+        g[17] = cpx[1] * c0x[1] + b00[1];
+        g[18] = cpx[2] * c0x[2] + b00[2];
+        g[19] = cpx[3] * c0x[3] + b00[3];
+        g[20] = c0x[0] * (g[16] + b00[0]) + b10[0] * cpx[0];
+        g[21] = c0x[1] * (g[17] + b00[1]) + b10[1] * cpx[1];
+        g[22] = c0x[2] * (g[18] + b00[2]) + b10[2] * cpx[2];
+        g[23] = c0x[3] * (g[19] + b00[3]) + b10[3] * cpx[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[28] = c0y[0];
+        g[29] = c0y[1];
+        g[30] = c0y[2];
+        g[31] = c0y[3];
+        g[32] = c0y[0] * c0y[0] + b10[0];
+        g[33] = c0y[1] * c0y[1] + b10[1];
+        g[34] = c0y[2] * c0y[2] + b10[2];
+        g[35] = c0y[3] * c0y[3] + b10[3];
+        g[36] = cpy[0];
+        g[37] = cpy[1];
+        g[38] = cpy[2];
+        g[39] = cpy[3];
+        g[40] = cpy[0] * c0y[0] + b00[0];
+        g[41] = cpy[1] * c0y[1] + b00[1];
+        g[42] = cpy[2] * c0y[2] + b00[2];
+        g[43] = cpy[3] * c0y[3] + b00[3];
+        g[44] = c0y[0] * (g[40] + b00[0]) + b10[0] * cpy[0];
+        g[45] = c0y[1] * (g[41] + b00[1]) + b10[1] * cpy[1];
+        g[46] = c0y[2] * (g[42] + b00[2]) + b10[2] * cpy[2];
+        g[47] = c0y[3] * (g[43] + b00[3]) + b10[3] * cpy[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[52] = c0z[0] * g[48];
+        g[53] = c0z[1] * g[49];
+        g[54] = c0z[2] * g[50];
+        g[55] = c0z[3] * g[51];
+        g[56] = c0z[0] * g[52] + b10[0] * g[48];
+        g[57] = c0z[1] * g[53] + b10[1] * g[49];
+        g[58] = c0z[2] * g[54] + b10[2] * g[50];
+        g[59] = c0z[3] * g[55] + b10[3] * g[51];
+        g[60] = cpz[0] * g[48];
+        g[61] = cpz[1] * g[49];
+        g[62] = cpz[2] * g[50];
+        g[63] = cpz[3] * g[51];
+        g[64] = cpz[0] * g[52] + b00[0] * g[48];
+        g[65] = cpz[1] * g[53] + b00[1] * g[49];
+        g[66] = cpz[2] * g[54] + b00[2] * g[50];
+        g[67] = cpz[3] * g[55] + b00[3] * g[51];
+        g[68] = c0z[0] * g[64] + b10[0] * g[60] + b00[0] * g[52];
+        g[69] = c0z[1] * g[65] + b10[1] * g[61] + b00[1] * g[53];
+        g[70] = c0z[2] * g[66] + b10[2] * g[62] + b00[2] * g[54];
+        g[71] = c0z[3] * g[67] + b10[3] * g[63] + b00[3] * g[55];
+}
+
+static inline void _srg0_2d4d_2010(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *cpx = bc->c0px;
+        double *cpy = bc->c0py;
+        double *cpz = bc->c0pz;
+        double *b00 = bc->b00;
+        double *b10 = bc->b10;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = c0x[0] * c0x[0] + b10[0];
+        g[9] = c0x[1] * c0x[1] + b10[1];
+        g[10] = c0x[2] * c0x[2] + b10[2];
+        g[11] = c0x[3] * c0x[3] + b10[3];
+        g[12] = cpx[0];
+        g[13] = cpx[1];
+        g[14] = cpx[2];
+        g[15] = cpx[3];
+        g[16] = cpx[0] * c0x[0] + b00[0];
+        g[17] = cpx[1] * c0x[1] + b00[1];
+        g[18] = cpx[2] * c0x[2] + b00[2];
+        g[19] = cpx[3] * c0x[3] + b00[3];
+        g[20] = c0x[0] * (g[16] + b00[0]) + b10[0] * cpx[0];
+        g[21] = c0x[1] * (g[17] + b00[1]) + b10[1] * cpx[1];
+        g[22] = c0x[2] * (g[18] + b00[2]) + b10[2] * cpx[2];
+        g[23] = c0x[3] * (g[19] + b00[3]) + b10[3] * cpx[3];
+        g[24] = 1;
+        g[25] = 1;
+        g[26] = 1;
+        g[27] = 1;
+        g[28] = c0y[0];
+        g[29] = c0y[1];
+        g[30] = c0y[2];
+        g[31] = c0y[3];
+        g[32] = c0y[0] * c0y[0] + b10[0];
+        g[33] = c0y[1] * c0y[1] + b10[1];
+        g[34] = c0y[2] * c0y[2] + b10[2];
+        g[35] = c0y[3] * c0y[3] + b10[3];
+        g[36] = cpy[0];
+        g[37] = cpy[1];
+        g[38] = cpy[2];
+        g[39] = cpy[3];
+        g[40] = cpy[0] * c0y[0] + b00[0];
+        g[41] = cpy[1] * c0y[1] + b00[1];
+        g[42] = cpy[2] * c0y[2] + b00[2];
+        g[43] = cpy[3] * c0y[3] + b00[3];
+        g[44] = c0y[0] * (g[40] + b00[0]) + b10[0] * cpy[0];
+        g[45] = c0y[1] * (g[41] + b00[1]) + b10[1] * cpy[1];
+        g[46] = c0y[2] * (g[42] + b00[2]) + b10[2] * cpy[2];
+        g[47] = c0y[3] * (g[43] + b00[3]) + b10[3] * cpy[3];
+        //g[48] = w[0];
+        //g[49] = w[0];
+        //g[50] = w[1];
+        //g[51] = w[1];
+        g[52] = c0z[0] * g[48];
+        g[53] = c0z[1] * g[49];
+        g[54] = c0z[2] * g[50];
+        g[55] = c0z[3] * g[51];
+        g[56] = c0z[0] * g[52] + b10[0] * g[48];
+        g[57] = c0z[1] * g[53] + b10[1] * g[49];
+        g[58] = c0z[2] * g[54] + b10[2] * g[50];
+        g[59] = c0z[3] * g[55] + b10[3] * g[51];
+        g[60] = cpz[0] * g[48];
+        g[61] = cpz[1] * g[49];
+        g[62] = cpz[2] * g[50];
+        g[63] = cpz[3] * g[51];
+        g[64] = cpz[0] * g[52] + b00[0] * g[48];
+        g[65] = cpz[1] * g[53] + b00[1] * g[49];
+        g[66] = cpz[2] * g[54] + b00[2] * g[50];
+        g[67] = cpz[3] * g[55] + b00[3] * g[51];
+        g[68] = c0z[0] * g[64] + b10[0] * g[60] + b00[0] * g[52];
+        g[69] = c0z[1] * g[65] + b10[1] * g[61] + b00[1] * g[53];
+        g[70] = c0z[2] * g[66] + b10[2] * g[62] + b00[2] * g[54];
+        g[71] = c0z[3] * g[67] + b10[3] * g[63] + b00[3] * g[55];
+}
+
+static inline void _srg0_2d4d_2100(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *b10 = bc->b10;
+        double xixj = envs->rirj[0];
+        double yiyj = envs->rirj[1];
+        double zizj = envs->rirj[2];
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = c0x[0] * c0x[0] + b10[0];
+        g[9] = c0x[1] * c0x[1] + b10[1];
+        g[10] = c0x[2] * c0x[2] + b10[2];
+        g[11] = c0x[3] * c0x[3] + b10[3];
+        g[24] = g[8] * (xixj + c0x[0]) + c0x[0] * 2 * b10[0];
+        g[25] = g[9] * (xixj + c0x[1]) + c0x[1] * 2 * b10[1];
+        g[26] = g[10] * (xixj + c0x[2]) + c0x[2] * 2 * b10[2];
+        g[27] = g[11] * (xixj + c0x[3]) + c0x[3] * 2 * b10[3];
+        g[20] = c0x[0] * (xixj + c0x[0]) + b10[0];
+        g[21] = c0x[1] * (xixj + c0x[1]) + b10[1];
+        g[22] = c0x[2] * (xixj + c0x[2]) + b10[2];
+        g[23] = c0x[3] * (xixj + c0x[3]) + b10[3];
+        g[16] = xixj + c0x[0];
+        g[17] = xixj + c0x[1];
+        g[18] = xixj + c0x[2];
+        g[19] = xixj + c0x[3];
+        g[32] = 1;
+        g[33] = 1;
+        g[34] = 1;
+        g[35] = 1;
+        g[36] = c0y[0];
+        g[37] = c0y[1];
+        g[38] = c0y[2];
+        g[39] = c0y[3];
+        g[40] = c0y[0] * c0y[0] + b10[0];
+        g[41] = c0y[1] * c0y[1] + b10[1];
+        g[42] = c0y[2] * c0y[2] + b10[2];
+        g[43] = c0y[3] * c0y[3] + b10[3];
+        g[56] = g[40] * (yiyj + c0y[0]) + c0y[0] * 2 * b10[0];
+        g[57] = g[41] * (yiyj + c0y[1]) + c0y[1] * 2 * b10[1];
+        g[58] = g[42] * (yiyj + c0y[2]) + c0y[2] * 2 * b10[2];
+        g[59] = g[43] * (yiyj + c0y[3]) + c0y[3] * 2 * b10[3];
+        g[52] = c0y[0] * (yiyj + c0y[0]) + b10[0];
+        g[53] = c0y[1] * (yiyj + c0y[1]) + b10[1];
+        g[54] = c0y[2] * (yiyj + c0y[2]) + b10[2];
+        g[55] = c0y[3] * (yiyj + c0y[3]) + b10[3];
+        g[48] = yiyj + c0y[0];
+        g[49] = yiyj + c0y[1];
+        g[50] = yiyj + c0y[2];
+        g[51] = yiyj + c0y[3];
+        //g[64] = w[0];
+        //g[65] = w[0];
+        //g[66] = w[1];
+        //g[67] = w[1];
+        g[68] = c0z[0] * g[64];
+        g[69] = c0z[1] * g[65];
+        g[70] = c0z[2] * g[66];
+        g[71] = c0z[3] * g[67];
+        g[72] = c0z[0] * g[68] + b10[0] * g[64];
+        g[73] = c0z[1] * g[69] + b10[1] * g[65];
+        g[74] = c0z[2] * g[70] + b10[2] * g[66];
+        g[75] = c0z[3] * g[71] + b10[3] * g[67];
+        g[88] = g[72] * (zizj + c0z[0]) + 2 * b10[0] * g[68];
+        g[89] = g[73] * (zizj + c0z[1]) + 2 * b10[1] * g[69];
+        g[90] = g[74] * (zizj + c0z[2]) + 2 * b10[2] * g[70];
+        g[91] = g[75] * (zizj + c0z[3]) + 2 * b10[3] * g[71];
+        g[84] = g[68] * (zizj + c0z[0]) + b10[0] * g[64];
+        g[85] = g[69] * (zizj + c0z[1]) + b10[1] * g[65];
+        g[86] = g[70] * (zizj + c0z[2]) + b10[2] * g[66];
+        g[87] = g[71] * (zizj + c0z[3]) + b10[3] * g[67];
+        g[80] = g[64] * (zizj + c0z[0]);
+        g[81] = g[65] * (zizj + c0z[1]);
+        g[82] = g[66] * (zizj + c0z[2]);
+        g[83] = g[67] * (zizj + c0z[3]);
+}
+
+static inline void _srg0_2d4d_3000(double *restrict g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        double *c0x = bc->c00x;
+        double *c0y = bc->c00y;
+        double *c0z = bc->c00z;
+        double *b10 = bc->b10;
+        g[0] = 1;
+        g[1] = 1;
+        g[2] = 1;
+        g[3] = 1;
+        g[4] = c0x[0];
+        g[5] = c0x[1];
+        g[6] = c0x[2];
+        g[7] = c0x[3];
+        g[8] = c0x[0] * c0x[0] + b10[0];
+        g[9] = c0x[1] * c0x[1] + b10[1];
+        g[10] = c0x[2] * c0x[2] + b10[2];
+        g[11] = c0x[3] * c0x[3] + b10[3];
+        g[12] = c0x[0] * (g[8] + 2 * b10[0]);
+        g[13] = c0x[1] * (g[9] + 2 * b10[1]);
+        g[14] = c0x[2] * (g[10] + 2 * b10[2]);
+        g[15] = c0x[3] * (g[11] + 2 * b10[3]);
+        g[16] = 1;
+        g[17] = 1;
+        g[18] = 1;
+        g[19] = 1;
+        g[20] = c0y[0];
+        g[21] = c0y[1];
+        g[22] = c0y[2];
+        g[23] = c0y[3];
+        g[24] = c0y[0] * c0y[0] + b10[0];
+        g[25] = c0y[1] * c0y[1] + b10[1];
+        g[26] = c0y[2] * c0y[2] + b10[2];
+        g[27] = c0y[3] * c0y[3] + b10[3];
+        g[28] = c0y[0] * (g[24] + 2 * b10[0]);
+        g[29] = c0y[1] * (g[25] + 2 * b10[1]);
+        g[30] = c0y[2] * (g[26] + 2 * b10[2]);
+        g[31] = c0y[3] * (g[27] + 2 * b10[3]);
+        //g[32] = w[0];
+        //g[33] = w[0];
+        //g[34] = w[1];
+        //g[35] = w[1];
+        g[36] = c0z[0] * g[32];
+        g[37] = c0z[1] * g[33];
+        g[38] = c0z[2] * g[34];
+        g[39] = c0z[3] * g[35];
+        g[40] = c0z[0] * g[36] + b10[0] * g[32];
+        g[41] = c0z[1] * g[37] + b10[1] * g[33];
+        g[42] = c0z[2] * g[38] + b10[2] * g[34];
+        g[43] = c0z[3] * g[39] + b10[3] * g[35];
+        g[44] = c0z[0] * g[40] + 2 * b10[0] * g[36];
+        g[45] = c0z[1] * g[41] + 2 * b10[1] * g[37];
+        g[46] = c0z[2] * g[42] + 2 * b10[2] * g[38];
+        g[47] = c0z[3] * g[43] + 2 * b10[3] * g[39];
+}
+
+void CINTsrg0_2e_2d4d_unrolled_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        int type_ijkl = ((envs->li_ceil << 6) | (envs->lj_ceil << 4) |
+                         (envs->lk_ceil << 2) | (envs->ll_ceil));
+        switch (type_ijkl) {
+                case 0b00000000: _srg0_2d4d_0000(g, bc, envs); return;
+                case 0b00000001: _srg0_2d4d_0001(g, bc, envs); return;
+                case 0b00000010: _srg0_2d4d_0002(g, bc, envs); return;
+                case 0b00000011: _srg0_2d4d_0003(g, bc, envs); return;
+                case 0b00000100: _srg0_2d4d_0010(g, bc, envs); return;
+                case 0b00000101: _srg0_2d4d_0011(g, bc, envs); return;
+                case 0b00000110: _srg0_2d4d_0012(g, bc, envs); return;
+                case 0b00001000: _srg0_2d4d_0020(g, bc, envs); return;
+                case 0b00001001: _srg0_2d4d_0021(g, bc, envs); return;
+                case 0b00001100: _srg0_2d4d_0030(g, bc, envs); return;
+                case 0b00010000: _srg0_2d4d_0100(g, bc, envs); return;
+                case 0b00010001: _srg0_2d4d_0101(g, bc, envs); return;
+                case 0b00010010: _srg0_2d4d_0102(g, bc, envs); return;
+                case 0b00010100: _srg0_2d4d_0110(g, bc, envs); return;
+                case 0b00010101: _srg0_2d4d_0111(g, bc, envs); return;
+                case 0b00011000: _srg0_2d4d_0120(g, bc, envs); return;
+                case 0b00100000: _srg0_2d4d_0200(g, bc, envs); return;
+                case 0b00100001: _srg0_2d4d_0201(g, bc, envs); return;
+                case 0b00100100: _srg0_2d4d_0210(g, bc, envs); return;
+                case 0b00110000: _srg0_2d4d_0300(g, bc, envs); return;
+                case 0b01000000: _srg0_2d4d_1000(g, bc, envs); return;
+                case 0b01000001: _srg0_2d4d_1001(g, bc, envs); return;
+                case 0b01000010: _srg0_2d4d_1002(g, bc, envs); return;
+                case 0b01000100: _srg0_2d4d_1010(g, bc, envs); return;
+                case 0b01000101: _srg0_2d4d_1011(g, bc, envs); return;
+                case 0b01001000: _srg0_2d4d_1020(g, bc, envs); return;
+                case 0b01010000: _srg0_2d4d_1100(g, bc, envs); return;
+                case 0b01010001: _srg0_2d4d_1101(g, bc, envs); return;
+                case 0b01010100: _srg0_2d4d_1110(g, bc, envs); return;
+                case 0b01100000: _srg0_2d4d_1200(g, bc, envs); return;
+                case 0b10000000: _srg0_2d4d_2000(g, bc, envs); return;
+                case 0b10000001: _srg0_2d4d_2001(g, bc, envs); return;
+                case 0b10000100: _srg0_2d4d_2010(g, bc, envs); return;
+                case 0b10010000: _srg0_2d4d_2100(g, bc, envs); return;
+                case 0b11000000: _srg0_2d4d_3000(g, bc, envs); return;
+        }
+        fprintf(stderr, "Dimension error for CINTg0_2e_lj2d4d: iklj = %d %d %d %d",
+               (int)envs->li_ceil, (int)envs->lk_ceil,
+               (int)envs->ll_ceil, (int)envs->lj_ceil);
+}
+
+void CINTg0_2e_lj2d4d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        CINTg0_2e_2d_simd1(g, bc, envs);
+        CINTg0_lj_4d_simd1(g, envs);
+}
+
+void CINTg0_2e_kj2d4d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        CINTg0_2e_2d_simd1(g, bc, envs);
+        CINTg0_kj_4d_simd1(g, envs);
+}
+void CINTg0_2e_ik2d4d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        CINTg0_2e_2d_simd1(g, bc, envs);
+        CINTg0_ik_4d_simd1(g, envs);
+}
+void CINTg0_2e_il2d4d_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs)
+{
+        CINTg0_2e_2d_simd1(g, bc, envs);
+        CINTg0_il_4d_simd1(g, envs);
+}
+
+/*
+ * g[i,k,l,j] = < ik | lj > = ( i j | k l )
+ */
+int CINTg0_2e_simd1(double *g, double *cutoff,
+                    Rys2eT *bc, CINTEnvVars *envs, int idsimd)
+{
+        ALIGNMM double u[MXRYSROOTS];
+        DEF_GXYZ(double, g, gx, gy, gz);
+        int irys;
+        int nroots = envs->nrys_roots;
+        double aij = envs->ai[idsimd] + envs->aj[idsimd];
+        double akl = envs->ak[idsimd] + envs->al[idsimd];
+        double a0, a1, fac1, x;
+        double *rij = envs->rij;
+        double *rkl = envs->rkl;
+        double *w = g + envs->g_size * 2; // ~ gz
+        double xij_kl = rij[0*SIMDD+idsimd] - rkl[0*SIMDD+idsimd];
+        double yij_kl = rij[1*SIMDD+idsimd] - rkl[1*SIMDD+idsimd];
+        double zij_kl = rij[2*SIMDD+idsimd] - rkl[2*SIMDD+idsimd];
+        double rr = xij_kl * xij_kl + yij_kl * yij_kl + zij_kl * zij_kl;
+
+        a1 = aij * akl;
+        a0 = a1 / (aij + akl);
+        fac1 = sqrt(a0 / (a1 * a1 * a1)) * envs->fac[idsimd];
+        x = a0 * rr;
+        const double omega = envs->env[PTR_RANGE_OMEGA];
+        double theta = 0;
+        if (omega == 0.) {
+                CINTrys_roots(nroots, x, u, w);
+        } else if (omega < 0.) {
+                // short-range part of range-separated Coulomb
+                theta = omega * omega / (omega * omega + a0);
+                // very small erfc() leads to ~0 weights. They can cause
+                // numerical issue in sr_rys_roots.
+                if (theta * x > cutoff[idsimd] || theta * x > EXPCUTOFF_SR) {
+                        return 0;
+                }
+                int rorder = envs->rys_order;
+                if (rorder == nroots) {
+                        CINTsr_rys_roots(nroots, x, sqrt(theta), u, w);
+                } else {
+                        double sqrt_theta = -sqrt(theta);
+                        CINTrys_roots(rorder, x, u, w);
+                        CINTrys_roots(rorder, theta*x, u+rorder, w+rorder);
+                        if (envs->g_size == 2) {
+                                g[0] = 1;
+                                g[1] = 1;
+                                g[2] = 1;
+                                g[3] = 1;
+                                g[4] *= fac1;
+                                g[5] *= fac1 * sqrt_theta;
+                                return 1;
+                        }
+                        for (irys = rorder; irys < nroots; irys++) {
+                                double ut = u[irys] * theta;
+                                u[irys] = ut / (u[irys]+1.-ut);
+                                w[irys] *= sqrt_theta;
+                        }
+                }
+        } else { // omega > 0.
+                // long-range part of range-separated Coulomb
+                theta = omega * omega / (omega * omega + a0);
+                x *= theta;
+                fac1 *= sqrt(theta);
+                CINTrys_roots(nroots, x, u, w);
+                for (irys = 0; irys < nroots; irys++) {
+                        double ut = u[irys] * theta;
+                        u[irys] = ut / (u[irys]+1.-ut);
+                }
+        }
+        if (envs->g_size == 1) {
+                g[0] = 1;
+                g[1] = 1;
+                g[2] *= fac1;
+                return 1;
+        }
+
+        double u2, tmp1, tmp2, tmp3, tmp4, tmp5;
+        double rijrx = rij[0*SIMDD+idsimd] - envs->rx_in_rijrx[0];
+        double rijry = rij[1*SIMDD+idsimd] - envs->rx_in_rijrx[1];
+        double rijrz = rij[2*SIMDD+idsimd] - envs->rx_in_rijrx[2];
+        double rklrx = rkl[0*SIMDD+idsimd] - envs->rx_in_rklrx[0];
+        double rklry = rkl[1*SIMDD+idsimd] - envs->rx_in_rklrx[1];
+        double rklrz = rkl[2*SIMDD+idsimd] - envs->rx_in_rklrx[2];
+        double *b00 = bc->b00;
+        double *b10 = bc->b10;
+        double *b01 = bc->b01;
+        double *c00x = bc->c00x;
+        double *c00y = bc->c00y;
+        double *c00z = bc->c00z;
+        double *c0px = bc->c0px;
+        double *c0py = bc->c0py;
+        double *c0pz = bc->c0pz;
+
+        for (irys = 0; irys < nroots; irys++) {
                 /*
-                 *t2 = u(i)/(1+u(i))
+                 *u(irys) = t2/(1-t2)
+                 *t2 = u(irys)/(1+u(irys))
                  *u2 = aij*akl/(aij+akl)*t2/(1-t2)
                  */
-                u2 = a0 * u[i];
-                div = 1 / (u2 * (aij + akl) + a1);
-                tmp1 = u2 * div;
-                tmp4 = .5 * div;
-                b00[i] = 0.5 * tmp1;
+                u2 = a0 * u[irys];
+                tmp4 = .5 / (u2 * (aij + akl) + a1);
+                tmp5 = u2 * tmp4;
+                tmp1 = 2. * tmp5;
                 tmp2 = tmp1 * akl;
                 tmp3 = tmp1 * aij;
-                b10[i] = b00[i] + tmp4 * akl;
-                b01[i] = b00[i] + tmp4 * aij;
-                c00x[i] = rijrx[0] - tmp2 * rijrkl[0];
-                c00y[i] = rijrx[1] - tmp2 * rijrkl[1];
-                c00z[i] = rijrx[2] - tmp2 * rijrkl[2];
-                c0px[i] = rklrx[0] + tmp3 * rijrkl[0];
-                c0py[i] = rklrx[1] + tmp3 * rijrkl[1];
-                c0pz[i] = rklrx[2] + tmp3 * rijrkl[2];
+                b00[irys] = tmp5;
+                b10[irys] = tmp5 + tmp4 * akl;
+                b01[irys] = tmp5 + tmp4 * aij;
+                c00x[irys] = rijrx - tmp2 * xij_kl;
+                c00y[irys] = rijry - tmp2 * yij_kl;
+                c00z[irys] = rijrz - tmp2 * zij_kl;
+                c0px[irys] = rklrx + tmp3 * xij_kl;
+                c0py[irys] = rklry + tmp3 * yij_kl;
+                c0pz[irys] = rklrz + tmp3 * zij_kl;
+                w[irys] *= fac1;
         }
 
         (*envs->f_g0_2d4d_simd1)(g, bc, envs);
diff --git a/src/g3c2e.c b/src/g3c2e.c
index 8e99004..f2221a2 100644
--- a/src/g3c2e.c
+++ b/src/g3c2e.c
@@ -54,7 +54,14 @@ void CINTinit_int3c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
         envs->nfk = (envs->k_l+1)*(envs->k_l+2)/2;
         envs->nfl = 1;
         envs->nf = envs->nfi * envs->nfk * envs->nfj;
-        envs->common_factor = 1;
+
+        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
+        envs->rj = env + atm(PTR_COORD, bas(ATOM_OF, j_sh));
+        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
+
+        envs->common_factor = (M_PI*M_PI*M_PI)*2/SQRTPI
+                * CINTcommon_fac_sp(envs->i_l) * CINTcommon_fac_sp(envs->j_l)
+                * CINTcommon_fac_sp(envs->k_l);
         if (env[PTR_EXPCUTOFF] == 0) {
                 envs->expcutoff = EXPCUTOFF;
         } else {
@@ -70,14 +77,14 @@ void CINTinit_int3c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
         envs->lj_ceil = envs->j_l + ng[JINC];
         envs->lk_ceil = 0;
         envs->ll_ceil = envs->k_l + ng[KINC];
-
-        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
-        envs->rj = env + atm(PTR_COORD, bas(ATOM_OF, j_sh));
-        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
-
-        int nroots = (envs->li_ceil + envs->lj_ceil + envs->ll_ceil)/2 + 1;
-        envs->nrys_roots = nroots;
-        assert(nroots < MXRYSROOTS);
+        int rys_order =(envs->li_ceil + envs->lj_ceil + envs->ll_ceil)/2 + 1;
+        int nrys_roots = rys_order;
+        double omega = env[PTR_RANGE_OMEGA];
+        if (omega < 0 && rys_order <= 3) {
+                nrys_roots *= 2;
+        }
+        envs->rys_order = rys_order;
+        envs->nrys_roots = nrys_roots;
 
         int dli, dlj, dlk;
         int ibase = envs->li_ceil > envs->lj_ceil;
@@ -91,11 +98,11 @@ void CINTinit_int3c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
         }
         dlk = envs->ll_ceil + 1;
 
-        envs->g_stride_i = nroots;
-        envs->g_stride_k = nroots * dli;
-        envs->g_stride_l = nroots * dli;
-        envs->g_stride_j = nroots * dli * dlk;
-        envs->g_size     = nroots * dli * dlk * dlj;
+        envs->g_stride_i = nrys_roots;
+        envs->g_stride_k = nrys_roots * dli;
+        envs->g_stride_l = nrys_roots * dli;
+        envs->g_stride_j = nrys_roots * dli * dlk;
+        envs->g_size     = nrys_roots * dli * dlk * dlj;
 
         MM_STORE(envs->al, MM_SET1(0.));
         MM_STORE(envs->rkl+0*SIMDD, MM_SET1(envs->rk[0]));
@@ -122,9 +129,13 @@ void CINTinit_int3c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
                 envs->rirj[2] = envs->rj[2] - envs->ri[2];
         }
 
-        if (nroots <= 2) {
+        if (rys_order <= 2) {
                 envs->f_g0_2d4d = &CINTg0_2e_2d4d_unrolled;
                 envs->f_g0_2d4d_simd1 = &CINTg0_2e_2d4d_unrolled_simd1;
+                if (rys_order != nrys_roots) {
+                        envs->f_g0_2d4d = &CINTsrg0_2e_2d4d_unrolled;
+                        envs->f_g0_2d4d_simd1 = &CINTsrg0_2e_2d4d_unrolled_simd1;
+                }
         } else if (ibase) {
                 envs->f_g0_2d4d = &CINTg0_2e_il2d4d;
                 envs->f_g0_2d4d_simd1 = &CINTg0_2e_il2d4d_simd1;
@@ -136,113 +147,3 @@ void CINTinit_int3c2e_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
         envs->f_g0_2e_simd1 = &CINTg0_2e_simd1;
 }
 
-
-#ifdef WITH_GTG
-void CINTg0_2e_lj2d4d_regular(double *g, Rys2eT *bc, CINTEnvVars *envs);
-void CINTg0_2e_lj2d4d_simd1_regular(double *g, Rys2eT *bc, CINTEnvVars *envs);
-int CINTg0_2e_gtg(double *g, Rys2eT *bc, CINTEnvVars *envs, int count);
-int CINTg0_2e_gtg_simd1(double *g, Rys2eT *bc, CINTEnvVars *envs, int idsimd);
-
-void CINTinit_int3c2e_gtg_EnvVars(CINTEnvVars *envs, int *ng, int *shls,
-                                  int *atm, int natm, int *bas, int nbas, double *env)
-{
-        envs->natm = natm;
-        envs->nbas = nbas;
-        envs->atm = atm;
-        envs->bas = bas;
-        envs->env = env;
-        envs->shls = shls;
-
-        int i_sh = shls[0];
-        int j_sh = shls[1];
-        int k_sh = shls[2];
-        envs->i_l = bas(ANG_OF, i_sh);
-        envs->j_l = bas(ANG_OF, j_sh);
-        envs->k_l = bas(ANG_OF, k_sh);
-        envs->l_l = 0;
-        envs->x_ctr[0] = bas(NCTR_OF, i_sh);
-        envs->x_ctr[1] = bas(NCTR_OF, j_sh);
-        envs->x_ctr[2] = bas(NCTR_OF, k_sh);
-        envs->x_ctr[3] = 1;
-        envs->nfi = (envs->i_l+1)*(envs->i_l+2)/2;
-        envs->nfj = (envs->j_l+1)*(envs->j_l+2)/2;
-        envs->nfk = (envs->k_l+1)*(envs->k_l+2)/2;
-        envs->nfl = 1;
-        envs->nf = envs->nfi * envs->nfk * envs->nfj;
-        envs->common_factor = SQRTPI * .5;
-        if (env[PTR_EXPCUTOFF] == 0) {
-                envs->expcutoff = EXPCUTOFF;
-        } else {
-                // +1 to ensure accuracy. See comments in libcint/cint2e.c
-                envs->expcutoff = MAX(MIN_EXPCUTOFF, env[PTR_EXPCUTOFF]);
-        }
-
-        envs->gbits = ng[GSHIFT];
-        envs->ncomp_e1 = ng[POS_E1];
-        envs->ncomp_e2 = ng[POS_E2];
-        envs->ncomp_tensor = ng[TENSOR];
-
-        envs->li_ceil = envs->i_l + ng[IINC];
-        envs->lj_ceil = envs->j_l + ng[JINC];
-        envs->lk_ceil = 0;
-        envs->ll_ceil = envs->k_l + ng[KINC];
-
-        envs->ri = env + atm(PTR_COORD, bas(ATOM_OF, i_sh));
-        envs->rj = env + atm(PTR_COORD, bas(ATOM_OF, j_sh));
-        envs->rk = env + atm(PTR_COORD, bas(ATOM_OF, k_sh));
-        envs->nrys_roots = 1;
-
-        int dli, dlj, dlk;
-        int ibase = envs->li_ceil > envs->lj_ceil;
-
-        if (ibase) {
-                dli = envs->li_ceil + envs->lj_ceil + 1;
-                dlj = envs->lj_ceil + 1;
-        } else {
-                dli = envs->li_ceil + 1;
-                dlj = envs->li_ceil + envs->lj_ceil + 1;
-        }
-        dlk = envs->ll_ceil + 1;
-
-        envs->g_stride_i = 1;
-        envs->g_stride_k = dli;
-        envs->g_stride_l = dli;
-        envs->g_stride_j = dli * dlk;
-        envs->g_size     = dli * dlk * dlj;
-
-        MM_STORE(envs->al, MM_SET1(0.));
-        MM_STORE(envs->rkl+0*SIMDD, MM_SET1(envs->rk[0]));
-        MM_STORE(envs->rkl+1*SIMDD, MM_SET1(envs->rk[1]));
-        MM_STORE(envs->rkl+2*SIMDD, MM_SET1(envs->rk[2]));
-        envs->g2d_klmax = envs->g_stride_k;
-        envs->rkrl[0] = envs->rk[0];
-        envs->rkrl[1] = envs->rk[1];
-        envs->rkrl[2] = envs->rk[2];
-        // in g0_2d rklrx = rkl - rx = 0 => rkl = rx
-        envs->rx_in_rklrx = envs->rk;
-
-        if (ibase) {
-                envs->g2d_ijmax = envs->g_stride_i;
-                envs->rx_in_rijrx = envs->ri;
-                envs->rirj[0] = envs->ri[0] - envs->rj[0];
-                envs->rirj[1] = envs->ri[1] - envs->rj[1];
-                envs->rirj[2] = envs->ri[2] - envs->rj[2];
-        } else {
-                envs->g2d_ijmax = envs->g_stride_j;
-                envs->rx_in_rijrx = envs->rj;
-                envs->rirj[0] = envs->rj[0] - envs->ri[0];
-                envs->rirj[1] = envs->rj[1] - envs->ri[1];
-                envs->rirj[2] = envs->rj[2] - envs->ri[2];
-        }
-
-        if (ibase) {
-                envs->f_g0_2d4d = &CINTg0_2e_il2d4d;
-                envs->f_g0_2d4d_simd1 = &CINTg0_2e_il2d4d_simd1;
-        } else {
-                envs->f_g0_2d4d = &CINTg0_2e_lj2d4d_regular;
-                envs->f_g0_2d4d_simd1 = &CINTg0_2e_lj2d4d_simd1_regular;
-        }
-        envs->f_g0_2e = &CINTg0_2e_gtg;
-        envs->f_g0_2e_simd1 = &CINTg0_2e_gtg_simd1;
-}
-#endif
diff --git a/src/gout2e.c b/src/gout2e.c
index f2e2950..c981e7f 100644
--- a/src/gout2e.c
+++ b/src/gout2e.c
@@ -27,12 +27,12 @@
 #if (SIMDD == 8)
 void CINTgout2e(double *gout, double *g, int *idx, CINTEnvVars *envs)
 {
+        int nrys_roots = envs->nrys_roots;
         int nf = envs->nf;
-        if (nf == 1) {
+        if (nf == 1 && nrys_roots == 1) {
                 double *gz = g + envs->g_size * 2 * SIMDD;
                 MM_STORE(gout, MM_LOAD(gz));
         } else {
-                int nrys_roots = envs->nrys_roots;
                 int i, n;
                 double *gx, *gy, *gz;
                 double *hx, *hy, *hz;
@@ -224,12 +224,12 @@ void CINTgout2e(double *gout, double *g, int *idx, CINTEnvVars *envs)
 
 void CINTgout2e(double *gout, double *g, int *idx, CINTEnvVars *envs)
 {
+        int nrys_roots = envs->nrys_roots;
         int nf = envs->nf;
-        if (nf == 1) {
+        if (nf == 1 && nrys_roots == 1) {
                 double *gz = g + envs->g_size * 2 * SIMDD;
                 MM_STORE(gout, MM_LOAD(gz));
         } else {
-                int nrys_roots = envs->nrys_roots;
                 int i, n;
                 double *gx, *gy, *gz;
                 double *hx, *hy, *hz;
@@ -528,12 +528,12 @@ void CINTgout2e(double *gout, double *g, int *idx, CINTEnvVars *envs)
 
 void CINTgout2e(double *gout, double *g, int *idx, CINTEnvVars *envs)
 {
+        int nrys_roots = envs->nrys_roots;
         int nf = envs->nf;
-        if (nf == 1) {
+        if (nf == 1 && nrys_roots == 1) {
                 double *gz = g + envs->g_size * 2 * SIMDD;
                 MM_STORE(gout, MM_LOAD(gz));
         } else {
-                int nrys_roots = envs->nrys_roots;
                 int i, n;
                 double *gx, *gy, *gz;
                 double *hx, *hy, *hz;
diff --git a/src/optimizer.c b/src/optimizer.c
index 9d85ccd..30fef5f 100644
--- a/src/optimizer.c
+++ b/src/optimizer.c
@@ -260,38 +260,6 @@ void CINTall_2e_stg_optimizer(CINTOpt **opt, int *ng,
 }
 #endif
 
-#ifdef WITH_GTG
-void CINTall_2e_gtg_optimizer(CINTOpt **opt, int *ng,
-                              int *atm, int natm, int *bas, int nbas, double *env)
-{
-        CINTinit_2e_optimizer(opt, atm, natm, bas, nbas, env);
-        CINTOpt_setij(*opt, ng, atm, natm, bas, nbas, env);
-        CINTOpt_set_non0coeff(*opt, atm, natm, bas, nbas, env);
-        gen_idx(*opt, &CINTinit_int2e_gtg_EnvVars, &CINTg4c_index_xyz,
-                4, 6, ng, atm, natm, bas, nbas, env);
-}
-
-void CINTall_3c2e_gtg_optimizer(CINTOpt **opt, int *ng,
-                                int *atm, int natm, int *bas, int nbas, double *env)
-{
-        CINTinit_2e_optimizer(opt, atm, natm, bas, nbas, env);
-        CINTOpt_setij(*opt, ng, atm, natm, bas, nbas, env);
-        CINTOpt_set_non0coeff(*opt, atm, natm, bas, nbas, env);
-        gen_idx(*opt, &CINTinit_int3c2e_gtg_EnvVars, &CINTg4c_index_xyz,
-                3, 12, ng, atm, natm, bas, nbas, env);
-}
-
-void CINTall_2c2e_gtg_optimizer(CINTOpt **opt, int *ng,
-                                int *atm, int natm, int *bas, int nbas, double *env)
-{
-        CINTinit_2e_optimizer(opt, atm, natm, bas, nbas, env);
-        CINTOpt_set_log_maxc(*opt, atm, natm, bas, nbas, env);
-        CINTOpt_set_non0coeff(*opt, atm, natm, bas, nbas, env);
-        gen_idx(*opt, &CINTinit_int2c2e_gtg_EnvVars, &CINTg2c_index_xyz,
-                2, ANG_MAX, ng, atm, natm, bas, nbas, env);
-}
-#endif
-
 void CINTOpt_log_max_pgto_coeff(double *log_maxc, double *coeff, int nprim, int nctr)
 {
         int i, ip;
@@ -350,7 +318,6 @@ int CINTset_pairdata(PairData *pairdata, double *ai, double *aj, double *ri, dou
         int lij = li_ceil + lj_ceil;
         if (lij > 0) {
                 double dist_ij = sqrt(rr_ij);
-#ifdef WITH_RANGE_COULOMB
                 double omega = env[PTR_RANGE_OMEGA];
                 if (omega < 0) {
                         double r_guess = 8.;
@@ -360,9 +327,6 @@ int CINTset_pairdata(PairData *pairdata, double *ai, double *aj, double *ri, dou
                 } else {
                         log_rr_ij += lij * approx_log(dist_ij + 1.);
                 }
-#else
-                log_rr_ij += lij * approx_log(dist_ij + 1.);
-#endif
         }
         PairData *pdata;
 
diff --git a/src/optimizer.h b/src/optimizer.h
index 2cd1087..e15802a 100644
--- a/src/optimizer.h
+++ b/src/optimizer.h
@@ -65,15 +65,6 @@ void CINTall_2e_stg_optimizer(CINTOpt **opt, int *ng,
                               int *atm, int natm, int *bas, int nbas, double *env);
 #endif
 
-#ifdef WITH_GTG
-void CINTall_2e_gtg_optimizer(CINTOpt **opt, int *ng,
-                              int *atm, int natm, int *bas, int nbas, double *env);
-void CINTall_3c2e_gtg_optimizer(CINTOpt **opt, int *ng,
-                                int *atm, int natm, int *bas, int nbas, double *env);
-void CINTall_2c2e_gtg_optimizer(CINTOpt **opt, int *ng,
-                                int *atm, int natm, int *bas, int nbas, double *env);
-#endif
-
 #ifndef HAVE_DEFINED_APPROX_LOG
 #define HAVE_DEFINED_APPROX_LOG
 #ifdef __X86__
diff --git a/src/rys_roots.c b/src/rys_roots.c
index a676afd..d189c7d 100644
--- a/src/rys_roots.c
+++ b/src/rys_roots.c
@@ -77,6 +77,13 @@ void _CINTrys_roots_batch(int nroots, double *x, double *u, double *w, int count
         double roots[MXRYSROOTS * 2];
         double *weights = roots + nroots;
         int i, k;
+        if (count < SIMDD) {
+                __MD r0 = MM_SET1(0.);
+                for (i = 0; i < nroots; i++) {
+                        MM_STORE(u+i*SIMDD, r0);
+                        MM_STORE(w+i*SIMDD, r0);
+                }
+        }
         for (i = 0; i < count; i++) {
                 CINTrys_roots(nroots, x[i], roots, weights);
                 for (k = 0; k < nroots; k++) {
@@ -96,7 +103,7 @@ void CINTrys_roots(int nroots, double x, double *u, double *w)
                         w[i] = POLY_SMALLX_W0[off+i] + POLY_SMALLX_W1[off+i] * x;
                 }
                 return;
-        } else if (x >= 50.) {
+        } else if (x >= 35+nroots*5) {
                 int off = nroots * (nroots - 1) / 2;
                 int i;
                 double rt;
@@ -141,7 +148,6 @@ void CINTrys_roots(int nroots, double x, double *u, double *w)
         }
 }
 
-#ifdef WITH_RANGE_COULOMB
 /*
  * lower is the lower bound of the sr integral
  */
@@ -295,28 +301,29 @@ int _CINTsr_rys_roots_batch(CINTEnvVars *envs, double *x, double *theta,
         double *weights = roots + nroots;
         int i, k;
         int all_negligible = 1;
-        double xt;
-
+        ALIGNMM double xt[SIMDD];
+        ALIGNMM double sqrt_theta[SIMDD];
+        __MD rtheta = MM_LOAD(theta);
+        __MD r0 = MM_SET1(0.);
+        MM_STORE(sqrt_theta, MM_SQRT(rtheta));
+        MM_STORE(xt, MM_MUL(MM_LOAD(x), rtheta));
+        for (i = 0; i < nroots; i++) {
+                MM_STORE(u+i*SIMDD, r0);
+                MM_STORE(w+i*SIMDD, r0);
+        }
         for (i = 0; i < count; i++) {
                 // very small erfc() leads to ~0 weights
-                xt = x[i] * theta[i]; 
-                if (xt < cutoff[i] && xt < EXPCUTOFF_SR) {
-                        all_negligible = 0;
-                        CINTsr_rys_roots(nroots, x[i], sqrt(theta[i]), roots, weights);
+                if (xt[i] < cutoff[i] && xt[i] < EXPCUTOFF_SR) {
+                        CINTsr_rys_roots(nroots, x[i], sqrt_theta[i], roots, weights);
                         for (k = 0; k < nroots; k++) {
                                 u[k*SIMDD+i] = roots[k];
                                 w[k*SIMDD+i] = weights[k];
                         }
-                } else {
-                        for (k = 0; k < nroots; k++) {
-                                u[k*SIMDD+i] = 0;
-                                w[k*SIMDD+i] = 0;
-                        }
+                        all_negligible = 0;
                 }
         }
         return all_negligible;
 }
-#endif  // WITH_RANGE_COULOMB
 
 static int rys_root1(double X, double *roots, double *weights)
 {