8329254: optimize integral reverse operations on x86 GFNI target.

Reviewed-by: sviswanathan
SAP · Apr 11, 2024 · b04b304 · b04b304
1 parent d9c84e7
commit b04b304
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 4 deletions.
diff --git a/src/hotspot/cpu/x86/vm_version_x86.cpp b/src/hotspot/cpu/x86/vm_version_x86.cpp
@@ -2951,6 +2951,8 @@ uint64_t VM_Version::CpuidInfo::feature_flags() const {
  if (sef_cpuid7_ecx1_eax.bits.avx_ifma != 0)
  result |= CPU_AVX_IFMA;
  }
+ if (sef_cpuid7_ecx.bits.gfni != 0)
+ result |= CPU_GFNI;
  if (sef_cpuid7_ebx.bits.avx512f != 0 &&
  xem_xcr0_eax.bits.opmask != 0 &&
  xem_xcr0_eax.bits.zmm512 != 0 &&
@@ -2976,8 +2978,6 @@ uint64_t VM_Version::CpuidInfo::feature_flags() const {
  result |= CPU_AVX512_VPCLMULQDQ;
  if (sef_cpuid7_ecx.bits.vaes != 0)
  result |= CPU_AVX512_VAES;
- if (sef_cpuid7_ecx.bits.gfni != 0)
- result |= CPU_GFNI;
  if (sef_cpuid7_ecx.bits.avx512_vnni != 0)
  result |= CPU_AVX512_VNNI;
  if (sef_cpuid7_ecx.bits.avx512_bitalg != 0)

diff --git a/src/hotspot/cpu/x86/x86_64.ad b/src/hotspot/cpu/x86/x86_64.ad
@@ -5662,7 +5662,7 @@ instruct bytes_reversebit_int(rRegI dst, rRegI src, rRegI rtmp, rFlagsReg cr) %{
  ins_pipe( ialu_reg );
 %}
 
-instruct bytes_reversebit_int_gfni(rRegI dst, rRegI src, regF xtmp1, regF xtmp2, rRegL rtmp, rFlagsReg cr) %{
+instruct bytes_reversebit_int_gfni(rRegI dst, rRegI src, vlRegF xtmp1, vlRegF xtmp2, rRegL rtmp, rFlagsReg cr) %{
  predicate(VM_Version::supports_gfni());
  match(Set dst (ReverseI src));
  effect(TEMP dst, TEMP xtmp1, TEMP xtmp2, TEMP rtmp, KILL cr);
@@ -5684,7 +5684,7 @@ instruct bytes_reversebit_long(rRegL dst, rRegL src, rRegL rtmp1, rRegL rtmp2, r
  ins_pipe( ialu_reg );
 %}
 
-instruct bytes_reversebit_long_gfni(rRegL dst, rRegL src, regD xtmp1, regD xtmp2, rRegL rtmp, rFlagsReg cr) %{
+instruct bytes_reversebit_long_gfni(rRegL dst, rRegL src, vlRegD xtmp1, vlRegD xtmp2, rRegL rtmp, rFlagsReg cr) %{
  predicate(VM_Version::supports_gfni());
  match(Set dst (ReverseL src));
  effect(TEMP dst, TEMP xtmp1, TEMP xtmp2, TEMP rtmp, KILL cr);