Remove redundant a_generic kernels

Also, rename u_generic to generic. Signed-off-by: Clayton Smith <[email protected]>
gnuradio · Oct 24, 2023 · 0dc32c8 · 0dc32c8
1 parent e853e9b
commit 0dc32c8
Show file tree

Hide file tree

Showing 28 changed files with 6 additions and 558 deletions.
diff --git a/kernels/volk/volk_16i_convert_8i.h b/kernels/volk/volk_16i_convert_8i.h
@@ -276,20 +276,4 @@ static inline void volk_16i_convert_8i_neon(int8_t* outputVector,
 #endif /* LV_HAVE_NEON */
 
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_16i_convert_8i_a_generic(int8_t* outputVector,
- const int16_t* inputVector,
- unsigned int num_points)
-{
- int8_t* outputVectorPtr = outputVector;
- const int16_t* inputVectorPtr = inputVector;
- unsigned int number = 0;
-
- for (number = 0; number < num_points; number++) {
- *outputVectorPtr++ = ((int8_t)(*inputVectorPtr++ >> 8));
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
 #endif /* INCLUDED_volk_16i_convert_8i_a_H */
diff --git a/kernels/volk/volk_16i_s32f_convert_32f.h b/kernels/volk/volk_16i_s32f_convert_32f.h
@@ -483,21 +483,4 @@ static inline void volk_16i_s32f_convert_32f_a_sse(float* outputVector,
 }
 #endif /* LV_HAVE_SSE */
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_16i_s32f_convert_32f_a_generic(float* outputVector,
- const int16_t* inputVector,
- const float scalar,
- unsigned int num_points)
-{
- float* outputVectorPtr = outputVector;
- const int16_t* inputVectorPtr = inputVector;
- unsigned int number = 0;
-
- for (number = 0; number < num_points; number++) {
- *outputVectorPtr++ = ((float)(*inputVectorPtr++)) / scalar;
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
 #endif /* INCLUDED_volk_16i_s32f_convert_32f_a_H */
diff --git a/kernels/volk/volk_16u_byteswap.h b/kernels/volk/volk_16u_byteswap.h
@@ -271,21 +271,6 @@ static inline void volk_16u_byteswap_neon_table(uint16_t* intsToSwap,
 }
 #endif /* LV_HAVE_NEON */
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_16u_byteswap_a_generic(uint16_t* intsToSwap,
- unsigned int num_points)
-{
- uint16_t* inputPtr = intsToSwap;
- for (unsigned int point = 0; point < num_points; point++) {
- uint16_t output = *inputPtr;
- output = (((output >> 8) & 0xff) | ((output << 8) & 0xff00));
- *inputPtr = output;
- inputPtr++;
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
 #ifdef LV_HAVE_ORC
 
 extern void volk_16u_byteswap_a_orc_impl(uint16_t* intsToSwap, unsigned int num_points);

diff --git a/kernels/volk/volk_32f_asin_32f.h b/kernels/volk/volk_32f_asin_32f.h
@@ -474,7 +474,7 @@ volk_32f_asin_32f_u_sse4_1(float* bVector, const float* aVector, unsigned int nu
 #ifdef LV_HAVE_GENERIC
 
 static inline void
-volk_32f_asin_32f_u_generic(float* bVector, const float* aVector, unsigned int num_points)
+volk_32f_asin_32f_generic(float* bVector, const float* aVector, unsigned int num_points)
 {
  float* bPtr = bVector;
  const float* aPtr = aVector;

diff --git a/kernels/volk/volk_32f_convert_64f.h b/kernels/volk/volk_32f_convert_64f.h
@@ -231,21 +231,4 @@ static inline void volk_32f_convert_64f_a_sse2(double* outputVector,
 #endif /* LV_HAVE_SSE2 */
 
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_32f_convert_64f_a_generic(double* outputVector,
- const float* inputVector,
- unsigned int num_points)
-{
- double* outputVectorPtr = outputVector;
- const float* inputVectorPtr = inputVector;
- unsigned int number = 0;
-
- for (number = 0; number < num_points; number++) {
- *outputVectorPtr++ = ((double)(*inputVectorPtr++));
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
-
 #endif /* INCLUDED_volk_32f_convert_64f_a_H */
diff --git a/kernels/volk/volk_32f_exp_32f.h b/kernels/volk/volk_32f_exp_32f.h
@@ -165,22 +165,6 @@ volk_32f_exp_32f_a_sse2(float* bVector, const float* aVector, unsigned int num_p
 #endif /* LV_HAVE_SSE2 for aligned */
 
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void
-volk_32f_exp_32f_a_generic(float* bVector, const float* aVector, unsigned int num_points)
-{
- float* bPtr = bVector;
- const float* aPtr = aVector;
- unsigned int number = 0;
-
- for (number = 0; number < num_points; number++) {
- *bPtr++ = expf(*aPtr++);
- }
-}
-
-#endif /* LV_HAVE_GENERIC */
-
 #endif /* INCLUDED_volk_32f_exp_32f_a_H */
 
 #ifndef INCLUDED_volk_32f_exp_32f_u_H
@@ -269,7 +253,7 @@ volk_32f_exp_32f_u_sse2(float* bVector, const float* aVector, unsigned int num_p
 #ifdef LV_HAVE_GENERIC
 
 static inline void
-volk_32f_exp_32f_u_generic(float* bVector, const float* aVector, unsigned int num_points)
+volk_32f_exp_32f_generic(float* bVector, const float* aVector, unsigned int num_points)
 {
  float* bPtr = bVector;
  const float* aPtr = aVector;

diff --git a/kernels/volk/volk_32f_log2_32f.h b/kernels/volk/volk_32f_log2_32f.h
@@ -459,24 +459,6 @@ volk_32f_log2_32f_neon(float* bVector, const float* aVector, unsigned int num_po
 #define INCLUDED_volk_32f_log2_32f_u_H
 
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void
-volk_32f_log2_32f_u_generic(float* bVector, const float* aVector, unsigned int num_points)
-{
- float* bPtr = bVector;
- const float* aPtr = aVector;
- unsigned int number = 0;
-
- for (number = 0; number < num_points; number++) {
- float const result = log2f(*aPtr++);
- *bPtr++ = isinf(result) ? -127.0f : result;
- }
-}
-
-#endif /* LV_HAVE_GENERIC */
-
-
 #ifdef LV_HAVE_SSE4_1
 #include <smmintrin.h>
 
@@ -555,7 +537,7 @@ volk_32f_log2_32f_u_sse4_1(float* bVector, const float* aVector, unsigned int nu
  }
 
  number = quarterPoints * 4;
- volk_32f_log2_32f_u_generic(bPtr, aPtr, num_points - number);
+ volk_32f_log2_32f_generic(bPtr, aPtr, num_points - number);
 }
 
 #endif /* LV_HAVE_SSE4_1 for unaligned */
@@ -643,7 +625,7 @@ static inline void volk_32f_log2_32f_u_avx2_fma(float* bVector,
  }
 
  number = eighthPoints * 8;
- volk_32f_log2_32f_u_generic(bPtr, aPtr, num_points - number);
+ volk_32f_log2_32f_generic(bPtr, aPtr, num_points - number);
 }
 
 #endif /* LV_HAVE_AVX2 && LV_HAVE_FMA for unaligned */
@@ -731,7 +713,7 @@ volk_32f_log2_32f_u_avx2(float* bVector, const float* aVector, unsigned int num_
  }
 
  number = eighthPoints * 8;
- volk_32f_log2_32f_u_generic(bPtr, aPtr, num_points - number);
+ volk_32f_log2_32f_generic(bPtr, aPtr, num_points - number);
 }
 
 #endif /* LV_HAVE_AVX2 for unaligned */

diff --git a/kernels/volk/volk_32f_s32f_convert_16i.h b/kernels/volk/volk_32f_s32f_convert_16i.h
@@ -553,29 +553,4 @@ static inline void volk_32f_s32f_convert_16i_a_sse(int16_t* outputVector,
 #endif /* LV_HAVE_SSE */
 
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_32f_s32f_convert_16i_a_generic(int16_t* outputVector,
- const float* inputVector,
- const float scalar,
- unsigned int num_points)
-{
- int16_t* outputVectorPtr = outputVector;
- const float* inputVectorPtr = inputVector;
- unsigned int number = 0;
- float min_val = SHRT_MIN;
- float max_val = SHRT_MAX;
- float r;
-
- for (number = 0; number < num_points; number++) {
- r = *inputVectorPtr++ * scalar;
- if (r < min_val)
- r = min_val;
- else if (r > max_val)
- r = max_val;
- *outputVectorPtr++ = (int16_t)rintf(r);
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
 #endif /* INCLUDED_volk_32f_s32f_convert_16i_a_H */
diff --git a/kernels/volk/volk_32f_s32f_convert_32i.h b/kernels/volk/volk_32f_s32f_convert_32i.h
@@ -406,16 +406,4 @@ static inline void volk_32f_s32f_convert_32i_a_sse(int32_t* outputVector,
 #endif /* LV_HAVE_SSE */
 
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_32f_s32f_convert_32i_a_generic(int32_t* outputVector,
- const float* inputVector,
- const float scalar,
- unsigned int num_points)
-{
- volk_32f_s32f_convert_32i_generic(outputVector, inputVector, scalar, num_points);
-}
-
-#endif /* LV_HAVE_GENERIC */
-
 #endif /* INCLUDED_volk_32f_s32f_convert_32i_a_H */
diff --git a/kernels/volk/volk_32f_s32f_multiply_32f.h b/kernels/volk/volk_32f_s32f_multiply_32f.h
@@ -258,25 +258,6 @@ static inline void volk_32f_s32f_multiply_32f_u_neon(float* cVector,
 #endif /* LV_HAVE_NEON */
 
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_32f_s32f_multiply_32f_a_generic(float* cVector,
- const float* aVector,
- const float scalar,
- unsigned int num_points)
-{
- unsigned int number = 0;
- const float* inputPtr = aVector;
- float* outputPtr = cVector;
- for (number = 0; number < num_points; number++) {
- *outputPtr = (*inputPtr) * scalar;
- inputPtr++;
- outputPtr++;
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
-
 #ifdef LV_HAVE_ORC
 
 extern void volk_32f_s32f_multiply_32f_a_orc_impl(float* dst,

diff --git a/kernels/volk/volk_32f_x2_add_32f.h b/kernels/volk/volk_32f_x2_add_32f.h
@@ -373,24 +373,6 @@ extern void volk_32f_x2_add_32f_a_neonpipeline(float* cVector,
  unsigned int num_points);
 #endif /* LV_HAVE_NEONV7 */
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_32f_x2_add_32f_a_generic(float* cVector,
- const float* aVector,
- const float* bVector,
- unsigned int num_points)
-{
- float* cPtr = cVector;
- const float* aPtr = aVector;
- const float* bPtr = bVector;
- unsigned int number = 0;
-
- for (number = 0; number < num_points; number++) {
- *cPtr++ = (*aPtr++) + (*bPtr++);
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
 
 #ifdef LV_HAVE_ORC
 

diff --git a/kernels/volk/volk_32f_x2_dot_prod_32f.h b/kernels/volk/volk_32f_x2_dot_prod_32f.h
@@ -470,30 +470,6 @@ static inline void volk_32f_x2_dot_prod_32f_u_avx512f(float* result,
 #include <volk/volk_common.h>
 
 
-#ifdef LV_HAVE_GENERIC
-
-
-static inline void volk_32f_x2_dot_prod_32f_a_generic(float* result,
- const float* input,
- const float* taps,
- unsigned int num_points)
-{
-
- float dotProduct = 0;
- const float* aPtr = input;
- const float* bPtr = taps;
- unsigned int number = 0;
-
- for (number = 0; number < num_points; number++) {
- dotProduct += ((*aPtr++) * (*bPtr++));
- }
-
- *result = dotProduct;
-}
-
-#endif /*LV_HAVE_GENERIC*/
-
-
 #ifdef LV_HAVE_SSE
 
 

diff --git a/kernels/volk/volk_32f_x2_multiply_32f.h b/kernels/volk/volk_32f_x2_multiply_32f.h
@@ -341,25 +341,6 @@ static inline void volk_32f_x2_multiply_32f_neon(float* cVector,
 #endif /* LV_HAVE_NEON */
 
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_32f_x2_multiply_32f_a_generic(float* cVector,
- const float* aVector,
- const float* bVector,
- unsigned int num_points)
-{
- float* cPtr = cVector;
- const float* aPtr = aVector;
- const float* bPtr = bVector;
- unsigned int number = 0;
-
- for (number = 0; number < num_points; number++) {
- *cPtr++ = (*aPtr++) * (*bPtr++);
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
-
 #ifdef LV_HAVE_ORC
 extern void volk_32f_x2_multiply_32f_a_orc_impl(float* cVector,
  const float* aVector,

diff --git a/kernels/volk/volk_32fc_conjugate_32fc.h b/kernels/volk/volk_32fc_conjugate_32fc.h
@@ -260,21 +260,4 @@ static inline void volk_32fc_conjugate_32fc_a_neon(lv_32fc_t* cVector,
 #endif /* LV_HAVE_NEON */
 
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_32fc_conjugate_32fc_a_generic(lv_32fc_t* cVector,
- const lv_32fc_t* aVector,
- unsigned int num_points)
-{
- lv_32fc_t* cPtr = cVector;
- const lv_32fc_t* aPtr = aVector;
- unsigned int number = 0;
-
- for (number = 0; number < num_points; number++) {
- *cPtr++ = lv_conj(*aPtr++);
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
-
 #endif /* INCLUDED_volk_32fc_conjugate_32fc_a_H */
diff --git a/kernels/volk/volk_32fc_deinterleave_64f_x2.h b/kernels/volk/volk_32fc_deinterleave_64f_x2.h
@@ -277,25 +277,6 @@ static inline void volk_32fc_deinterleave_64f_x2_a_sse2(double* iBuffer,
 }
 #endif /* LV_HAVE_SSE */
 
-#ifdef LV_HAVE_GENERIC
-
-static inline void volk_32fc_deinterleave_64f_x2_a_generic(double* iBuffer,
- double* qBuffer,
- const lv_32fc_t* complexVector,
- unsigned int num_points)
-{
- unsigned int number = 0;
- const float* complexVectorPtr = (float*)complexVector;
- double* iBufferPtr = iBuffer;
- double* qBufferPtr = qBuffer;
-
- for (number = 0; number < num_points; number++) {
- *iBufferPtr++ = (double)*complexVectorPtr++;
- *qBufferPtr++ = (double)*complexVectorPtr++;
- }
-}
-#endif /* LV_HAVE_GENERIC */
-
 #ifdef LV_HAVE_NEONV8
 #include <arm_neon.h>