5g-team
/
srsRAN_4G
mirror of https://github.com/pvnis/srsRAN_4G.git
2
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

adding neon support to new vector structure

Browse Source
master
yagoda 7 years ago
parent 2019ca31ef
commit f4b9e7311a
2 changed files with 116 additions and 44 deletions
Show Stats Download Patch File Download Diff File

123
lib/include/srslte/phy/utils/simd.h
Unescape Escape View File

@ -33,6 +33,11 @@
#endif #endif
#include <immintrin.h> #include <immintrin.h>
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#include <stdio.h>
#ifdef HAVE_NEON
#include <arm_neon.h>
#endif
/* /*
* SSE Macros * SSE Macros
@ -140,6 +145,7 @@
#define SRSLTE_SIMD_S_SIZE 0 #define SRSLTE_SIMD_S_SIZE 0
#define SRSLTE_SIMD_C16_SIZE 0 #define SRSLTE_SIMD_C16_SIZE 0
#endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
@ -159,7 +165,7 @@ typedef __m256 simd_f_t;
typedef __m128 simd_f_t; typedef __m128 simd_f_t;
#else /* HAVE_NEON */ #else /* HAVE_NEON */
#ifdef HAVE_NEON #ifdef HAVE_NEON
typedef float32x4 simd_f_t; typedef float32x4_t simd_f_t;
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -188,7 +194,7 @@ static inline simd_f_t srslte_simd_f_loadu(float *ptr) {
#ifdef LV_HAVE_AVX512 #ifdef LV_HAVE_AVX512
return _mm512_loadu_ps(ptr); return _mm512_loadu_ps(ptr);
#else /* LV_HAVE_AVX512 */ #else /* LV_HAVE_AVX512 */
#ifdef LV_HAVE_AVX2 #ifdef LV_HAVE_AVX2
return _mm256_loadu_ps(ptr); return _mm256_loadu_ps(ptr);
#else /* LV_HAVE_AVX2 */ #else /* LV_HAVE_AVX2 */
#ifdef LV_HAVE_SSE #ifdef LV_HAVE_SSE
@ -202,7 +208,7 @@ static inline simd_f_t srslte_simd_f_loadu(float *ptr) {
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
} }
static inline void srslte_simd_f_store(float *ptr, simd_f_t simdreg) {vst1q_f32 static inline void srslte_simd_f_store(float *ptr, simd_f_t simdreg) {
#ifdef LV_HAVE_AVX512 #ifdef LV_HAVE_AVX512
_mm512_store_ps(ptr, simdreg); _mm512_store_ps(ptr, simdreg);
#else /* LV_HAVE_AVX512 */ #else /* LV_HAVE_AVX512 */
@ -281,11 +287,11 @@ static inline simd_f_t srslte_simd_f_rcp(simd_f_t a) {
#ifdef LV_HAVE_AVX2 #ifdef LV_HAVE_AVX2
return _mm256_rcp_ps(a); return _mm256_rcp_ps(a);
#else /* LV_HAVE_AVX2 */ #else /* LV_HAVE_AVX2 */
#ifdef LV_HAVE_SSE #ifdef LV_HAVE_SSE
return _mm_rcp_ps(a); return _mm_rcp_ps(a);
#else /* LV_HAVE_SSE */ #else /* LV_HAVE_SSE */
#ifdef HAVE_NEON #ifdef HAVE_NEON
return vrecpeq_f32(a); return vmulq_f32(vrecpeq_f32(a), vrecpsq_f32(vrecpeq_f32(a), a));
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -302,6 +308,22 @@ static inline simd_f_t srslte_simd_f_addsub(simd_f_t a, simd_f_t b) {
#else /* LV_HAVE_AVX2 */ #else /* LV_HAVE_AVX2 */
#ifdef LV_HAVE_SSE #ifdef LV_HAVE_SSE
return _mm_addsub_ps(a, b); return _mm_addsub_ps(a, b);
#else /* LV_HAVE_SSE */
#ifdef HAVE_NEON // CURRENTLY USES GENERIC IMPLEMENTATION FOR NEON
float* a_ptr = &a;
float* b_ptr = &b;
simd_f_t ret;
float* c_ptr = &ret;
for(int i = 0; i<4;i++){
if(i%2==0){
c_ptr[i] = a_ptr[i] - b_ptr[i];
}else{
c_ptr[i] = a_ptr[i] + b_ptr[i];
}
}
return ret;
#endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
@ -373,6 +395,10 @@ static inline simd_f_t srslte_simd_f_swap(simd_f_t a) {
#else /* LV_HAVE_AVX2 */ #else /* LV_HAVE_AVX2 */
#ifdef LV_HAVE_SSE #ifdef LV_HAVE_SSE
return _mm_shuffle_ps(a, a, 0b10110001); return _mm_shuffle_ps(a, a, 0b10110001);
#else /* LV_HAVE_SSE */
#ifdef HAVE_NEON
return vcombine_f32(vrev64_f32(vget_low_f32(a)), vrev64_f32(vget_high_f32(a)));
#endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
@ -421,7 +447,9 @@ static inline simd_f_t srslte_simd_f_sqrt(simd_f_t a) {
return _mm_sqrt_ps(a); return _mm_sqrt_ps(a);
#else /* LV_HAVE_SSE */ #else /* LV_HAVE_SSE */
#ifdef HAVE_NEON #ifdef HAVE_NEON
return vrecpeq_f32(vrsqrteq_f32(a)); float32x4_t sqrt_reciprocal = vrsqrteq_f32(a);
sqrt_reciprocal = vmulq_f32(vrsqrtsq_f32(vmulq_f32(a,sqrt_reciprocal), sqrt_reciprocal),sqrt_reciprocal);
return vmulq_f32(a,sqrt_reciprocal);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -530,8 +558,8 @@ static inline simd_cf_t srslte_simd_cf_load(float *re, float *im) {
ret.im = _mm_load_ps(im); ret.im = _mm_load_ps(im);
#else /*HAVE_NEON*/ #else /*HAVE_NEON*/
#ifdef HAVE_NEON #ifdef HAVE_NEON
ret.val[0] = vld1q_f32(ptr); ret.val[0] = vld1q_f32(re);
ret.val[1] = vld1q_f32(ptr); ret.val[1] = vld1q_f32(im);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -554,8 +582,8 @@ static inline simd_cf_t srslte_simd_cf_loadu(float *re, float *im) {
ret.im = _mm_loadu_ps(im); ret.im = _mm_loadu_ps(im);
#else /*HAVE_NEON*/ #else /*HAVE_NEON*/
#ifdef HAVE_NEON #ifdef HAVE_NEON
ret.val[0] = vld1q_f32(ptr); ret.val[0] = vld1q_f32(re);
ret.val[1] = vld1q_f32(ptr); ret.val[1] = vld1q_f32(im);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -684,8 +712,8 @@ static inline simd_cf_t srslte_simd_cf_set1 (cf_t x) {
ret.im = _mm_set1_ps(__imag__ x); ret.im = _mm_set1_ps(__imag__ x);
#else /*HAVE_NEON*/ #else /*HAVE_NEON*/
#ifdef HAVE_NEON #ifdef HAVE_NEON
re.val[0] = vdupq_n_f32(__real__ x); ret.val[0] = vdupq_n_f32(__real__ x);
im.val[1] = vdupq_n_f32(__imag__ x); ret.val[1] = vdupq_n_f32(__imag__ x);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -772,8 +800,8 @@ static inline simd_cf_t srslte_simd_cf_add (simd_cf_t a, simd_cf_t b) {
ret.im = _mm_add_ps(a.im, b.im); ret.im = _mm_add_ps(a.im, b.im);
#else /* LV_HAVE_SSE */ #else /* LV_HAVE_SSE */
#ifdef HAVE_NEON #ifdef HAVE_NEON
ret.val[0] = vaddq_f32(a.val[0],a.val[0]); ret.val[0] = vaddq_f32(a.val[0],b.val[0]);
ret.val[1] = vaddq_f32(a.val[1],a.val[1]); ret.val[1] = vaddq_f32(a.val[1],b.val[1]);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -839,8 +867,8 @@ static inline simd_cf_t srslte_simd_cf_rcp (simd_cf_t a) {
simd_f_t a2re = vmulq_f32(a.val[0], a.val[0]); simd_f_t a2re = vmulq_f32(a.val[0], a.val[0]);
simd_f_t a2im = vmulq_f32(a.val[1], a.val[1]); simd_f_t a2im = vmulq_f32(a.val[1], a.val[1]);
simd_f_t mod2 = vaddq_f32(a2re, a2im); simd_f_t mod2 = vaddq_f32(a2re, a2im);
simd_f_t rcp = vrecpeq_f32(mod2); simd_f_t rcp = vmulq_f32(vrecpeq_f32(mod2), vrecpsq_f32(vrecpeq_f32(mod2), mod2));
simd_f_t neg_a_im = vnegq_f32(vdupq_n_f32(-0.0f), a.val[1]); simd_f_t neg_a_im = vnegq_f32(a.val[1]);
ret.val[0] = vmulq_f32(a.val[0], rcp); ret.val[0] = vmulq_f32(a.val[0], rcp);
ret.val[1] = vmulq_f32(neg_a_im, rcp); ret.val[1] = vmulq_f32(neg_a_im, rcp);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
@ -868,6 +896,7 @@ static inline simd_cf_t srslte_simd_cf_zero (void) {
ret.val[0] = vdupq_n_f32(0); ret.val[0] = vdupq_n_f32(0);
ret.val[1] = vdupq_n_f32(0); ret.val[1] = vdupq_n_f32(0);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
return ret; return ret;
@ -889,9 +918,9 @@ typedef __m256 simd_sel_t;
typedef __m128i simd_i_t; typedef __m128i simd_i_t;
typedef __m128 simd_sel_t; typedef __m128 simd_sel_t;
#else /* LV_HAVE_AVX2 */ #else /* LV_HAVE_AVX2 */
#ifdef LV_HAVE_SSE #ifdef HAVE_NEON
typedef int32x4_t simd_i_t; typedef int32x4_t simd_i_t;
typedef __m128 simd_sel_t; typedef int32x4_t simd_sel_t;
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -908,7 +937,7 @@ static inline simd_i_t srslte_simd_i_load(int *x) {
return _mm_load_si128((__m128i*)x); return _mm_load_si128((__m128i*)x);
#else #else
#ifdef HAVE_NEON #ifdef HAVE_NEON
return vld1_s32((int32x4_t*)x); return vld1q_s32((int*)x);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -924,6 +953,10 @@ static inline void srslte_simd_i_store(int *x, simd_i_t reg) {
#else #else
#ifdef LV_HAVE_SSE #ifdef LV_HAVE_SSE
_mm_store_si128((__m128i*)x, reg); _mm_store_si128((__m128i*)x, reg);
#else
#ifdef HAVE_NEON
vst1q_s32((int*)x, reg);
#endif /*HAVE_NEON*/
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
@ -963,7 +996,7 @@ static inline simd_i_t srslte_simd_i_add(simd_i_t a, simd_i_t b) {
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
}vcgtq_f32 }
static inline simd_sel_t srslte_simd_f_max(simd_f_t a, simd_f_t b) { static inline simd_sel_t srslte_simd_f_max(simd_f_t a, simd_f_t b) {
#ifdef LV_HAVE_AVX512 #ifdef LV_HAVE_AVX512
@ -992,6 +1025,25 @@ static inline simd_i_t srslte_simd_i_select(simd_i_t a, simd_i_t b, simd_sel_t s
#else #else
#ifdef LV_HAVE_SSE #ifdef LV_HAVE_SSE
return (__m128i) _mm_blendv_ps((__m128)a, (__m128)b, selector); return (__m128i) _mm_blendv_ps((__m128)a, (__m128)b, selector);
#else /* LV_HAVE_SSE */
#ifdef HAVE_NEON // CURRENTLY USES GENERIC IMPLEMENTATION FOR NEON
int* a_ptr = &a;
int* b_ptr = &b;
simd_i_t ret;
int* sel = &selector;
int* c_ptr = &ret;
for(int i = 0;i<4;i++)
{
if(sel[i] == -1){
c_ptr[i] = b_ptr[i];
}else{
c_ptr[i] = a_ptr[i];
}
}
return ret;
#endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
@ -1030,7 +1082,7 @@ static inline simd_s_t srslte_simd_s_load(int16_t *ptr) {
return _mm_load_si128((__m128i*) ptr); return _mm_load_si128((__m128i*) ptr);
#else /* LV_HAVE_SSE */ #else /* LV_HAVE_SSE */
#ifdef HAVE_NEON #ifdef HAVE_NEON
return vld1q_s16((int16x8_t*) ptr); return vld1q_s16(ptr);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -1048,7 +1100,7 @@ static inline simd_s_t srslte_simd_s_loadu(int16_t *ptr) {
return _mm_loadu_si128((__m128i*) ptr) return _mm_loadu_si128((__m128i*) ptr)
#else /* LV_HAVE_SSE */ #else /* LV_HAVE_SSE */
#ifdef HAVE_NEON #ifdef HAVE_NEON
return vld1q_s16((int16x8_t*) ptr); return vld1q_s16(ptr);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -1066,7 +1118,7 @@ static inline void srslte_simd_s_store(int16_t *ptr, simd_s_t simdreg) {
_mm_store_si128((__m128i*) ptr, simdreg); _mm_store_si128((__m128i*) ptr, simdreg);
#else /* LV_HAVE_SSE */ #else /* LV_HAVE_SSE */
#ifdef HAVE_NEON #ifdef HAVE_NEON
vst1q_s16((int16x8_t*) ptr, simdreg); vst1q_s16( ptr, simdreg);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -1084,13 +1136,12 @@ static inline void srslte_simd_s_storeu(int16_t *ptr, simd_s_t simdreg) {
_mm_storeu_si128((__m128i*) ptr, simdreg); _mm_storeu_si128((__m128i*) ptr, simdreg);
#else /* LV_HAVE_SSE */ #else /* LV_HAVE_SSE */
#ifdef HAVE_NEON #ifdef HAVE_NEON
vst1q_s16((int16x8_t*) ptr, simdreg); vst1q_s16(ptr, simdreg);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
} }
vdupq_n_s16
static inline simd_s_t srslte_simd_s_zero(void) { static inline simd_s_t srslte_simd_s_zero(void) {
#ifdef LV_HAVE_AVX512 #ifdef LV_HAVE_AVX512
return _mm512_setzero_si512(); return _mm512_setzero_si512();
@ -1107,7 +1158,7 @@ static inline simd_s_t srslte_simd_s_zero(void) {
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
#endif /* LV_HAVE_AVX512 */ #endif /* LV_HAVE_AVX512 */
}vmulq_s16 }
static inline simd_s_t srslte_simd_s_mul(simd_s_t a, simd_s_t b) { static inline simd_s_t srslte_simd_s_mul(simd_s_t a, simd_s_t b) {
#ifdef LV_HAVE_AVX512 #ifdef LV_HAVE_AVX512
@ -1294,7 +1345,7 @@ static inline void srslte_simd_c16i_store(c16_t *ptr, simd_c16_t simdreg) {
_mm_store_si128((__m128i *) (ptr + 8), _mm_blend_epi16(re_sw, simdreg.im.m128, 0b10101010)); _mm_store_si128((__m128i *) (ptr + 8), _mm_blend_epi16(re_sw, simdreg.im.m128, 0b10101010));
#else /*HAVE_NEON*/ #else /*HAVE_NEON*/
#ifdef HAVE_NEON #ifdef HAVE_NEON
vst2q_f32((float*)(ptr) ,simdreg); vst2q_s16((int16_t*)(ptr) ,simdreg.m128);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -1314,7 +1365,7 @@ static inline void srslte_simd_c16i_storeu(c16_t *ptr, simd_c16_t simdreg) {
_mm_storeu_si128((__m128i *) (ptr + 8), _mm_blend_epi16(re_sw, simdreg.im.m128, 0b10101010)); _mm_storeu_si128((__m128i *) (ptr + 8), _mm_blend_epi16(re_sw, simdreg.im.m128, 0b10101010));
#else /*HAVE_NEON*/ #else /*HAVE_NEON*/
#ifdef HAVE_NEON #ifdef HAVE_NEON
vst2q_f32((float*)(ptr) ,simdreg); vst2q_s16((int16_t*)(ptr) ,simdreg.m128);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -1330,8 +1381,8 @@ static inline void srslte_simd_c16_store(int16_t *re, int16_t *im, simd_c16_t si
_mm_store_si128((__m128i *) im, simdreg.im.m128); _mm_store_si128((__m128i *) im, simdreg.im.m128);
#else #else
#ifdef HAVE_NEON #ifdef HAVE_NEON
vst1q_f32((int16_t *) re, simdreg.m128.val[0]); vst1q_s16((int16_t *) re, simdreg.m128.val[0]);
vst1q_f32((int16_t *) im, simdreg.m128.val[1]); vst1q_s16((int16_t *) im, simdreg.m128.val[1]);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
@ -1347,15 +1398,13 @@ static inline void srslte_simd_c16_storeu(int16_t *re, int16_t *im, simd_c16_t s
_mm_storeu_si128((__m128i *) im, simdreg.im.m128); _mm_storeu_si128((__m128i *) im, simdreg.im.m128);
#else #else
#ifdef HAVE_NEON #ifdef HAVE_NEON
vst1q_f32((int16_t *) re, simdreg.m128.val[0]); vst1q_s16((int16_t *) re, simdreg.m128.val[0]);
vst1q_f32((int16_t *) im, simdreg.m128.val[1]); vst1q_s16((int16_t *) im, simdreg.m128.val[1]);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
} }
//TODO
static inline simd_c16_t srslte_simd_c16_prod (simd_c16_t a, simd_c16_t b) { static inline simd_c16_t srslte_simd_c16_prod (simd_c16_t a, simd_c16_t b) {
simd_c16_t ret; simd_c16_t ret;
#ifdef LV_HAVE_AVX2 #ifdef LV_HAVE_AVX2
@ -1385,14 +1434,14 @@ static inline simd_c16_t srslte_simd_c16_add (simd_c16_t a, simd_c16_t b) {
ret.im.m128 = _mm_add_epi16(a.im.m128, b.im.m128); ret.im.m128 = _mm_add_epi16(a.im.m128, b.im.m128);
#else #else
#ifdef HAVE_NEON #ifdef HAVE_NEON
ret.m128.val[0] = vaddq_s32(a.m127.val[0],a.m127.val[0]); ret.m128.val[0] = vaddq_s16(a.m128.val[0],a.m128.val[0]);
ret.m128.val[1] = vaddq_s32(a.m127.val[1],a.m127.val[1]); ret.m128.val[1] = vaddq_s16(a.m128.val[1],a.m128.val[1]);
#endif /* HAVE_NEON */ #endif /* HAVE_NEON */
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
#endif /* LV_HAVE_AVX2 */ #endif /* LV_HAVE_AVX2 */
return ret; return ret;
} }
vdupq_n_s16
static inline simd_c16_t srslte_simd_c16_zero (void) { static inline simd_c16_t srslte_simd_c16_zero (void) {
simd_c16_t ret; simd_c16_t ret;
#ifdef LV_HAVE_AVX2 #ifdef LV_HAVE_AVX2

37
lib/src/phy/utils/vector_simd.c
Unescape Escape View File

@ -751,10 +751,37 @@ void srslte_vec_div_fff_simd(float *x, float *y, float *z, int len) {
} }
} }
int srslte_vec_sc_prod_ccc_simd2(cf_t *x, cf_t h, cf_t *z, int len)
{
int i = 0;
const unsigned int loops = len / 4;
#ifdef HAVE_NEON
simd_cf_t h_vec;
h_vec.val[0] = srslte_simd_f_set1(__real__ h);
h_vec.val[1] = srslte_simd_f_set1(__imag__ h);
for (; i < loops; i++) {
simd_cf_t in = srslte_simd_cfi_load(&x[i*4]);
simd_cf_t temp = srslte_simd_cf_prod(in, h_vec);
srslte_simd_cfi_store(&z[i*4], temp);
}
#endif
i = loops * 4;
return i;
}
void srslte_vec_sc_prod_ccc_simd(cf_t *x, cf_t h, cf_t *z, int len) { void srslte_vec_sc_prod_ccc_simd(cf_t *x, cf_t h, cf_t *z, int len) {
int i = 0; int i = 0;
#if SRSLTE_SIMD_F_SIZE #if SRSLTE_SIMD_F_SIZE
#ifdef HAVE_NEON
i = srslte_vec_sc_prod_ccc_simd2(x, h, z, len);
#else
const simd_f_t hre = srslte_simd_f_set1(__real__ h); const simd_f_t hre = srslte_simd_f_set1(__real__ h);
const simd_f_t him = srslte_simd_f_set1(__imag__ h); const simd_f_t him = srslte_simd_f_set1(__imag__ h);
@ -766,8 +793,8 @@ void srslte_vec_sc_prod_ccc_simd(cf_t *x, cf_t h, cf_t *z, int len) {
simd_f_t sw = srslte_simd_f_swap(temp); simd_f_t sw = srslte_simd_f_swap(temp);
simd_f_t m2 = srslte_simd_f_mul(him, sw); simd_f_t m2 = srslte_simd_f_mul(him, sw);
simd_f_t r = srslte_simd_f_addsub(m1, m2); simd_f_t r = srslte_simd_f_addsub(m1, m2);
srslte_simd_f_store((float *) &z[i], r); srslte_simd_f_store((float *) &z[i], r);
} }
} else { } else {
for (; i < len - SRSLTE_SIMD_F_SIZE / 2 + 1; i += SRSLTE_SIMD_F_SIZE / 2) { for (; i < len - SRSLTE_SIMD_F_SIZE / 2 + 1; i += SRSLTE_SIMD_F_SIZE / 2) {
@ -782,10 +809,11 @@ void srslte_vec_sc_prod_ccc_simd(cf_t *x, cf_t h, cf_t *z, int len) {
} }
} }
#endif #endif
#endif
for (; i < len; i++) { for (; i < len; i++) {
z[i] = x[i] * h; z[i] = x[i] * h;
} }
} }
void srslte_vec_sc_prod_fff_simd(float *x, float h, float *z, int len) { void srslte_vec_sc_prod_fff_simd(float *x, float h, float *z, int len) {
@ -831,7 +859,6 @@ void srslte_vec_abs_cf_simd(cf_t *x, float *z, int len) {
simd_f_t z1 = srslte_simd_f_hadd(mul1, mul2); simd_f_t z1 = srslte_simd_f_hadd(mul1, mul2);
z1 = srslte_simd_f_sqrt(z1); z1 = srslte_simd_f_sqrt(z1);
srslte_simd_f_store(&z[i], z1); srslte_simd_f_store(&z[i], z1);
} }
} else { } else {
@ -966,9 +993,7 @@ uint32_t srslte_vec_max_fi_simd(float *x, int len) {
if (SRSLTE_IS_ALIGNED(x)) { if (SRSLTE_IS_ALIGNED(x)) {
for (; i < len - SRSLTE_SIMD_I_SIZE + 1; i += SRSLTE_SIMD_I_SIZE) { for (; i < len - SRSLTE_SIMD_I_SIZE + 1; i += SRSLTE_SIMD_I_SIZE) {
simd_f_t a = srslte_simd_f_load(&x[i]); simd_f_t a = srslte_simd_f_load(&x[i]);
simd_sel_t res = srslte_simd_f_max(a, simd_max_values); simd_sel_t res = srslte_simd_f_max(a, simd_max_values);
simd_max_indexes = srslte_simd_i_select(simd_max_indexes, simd_indexes, res); simd_max_indexes = srslte_simd_i_select(simd_max_indexes, simd_indexes, res);
simd_max_values = (simd_f_t) srslte_simd_i_select((simd_i_t) simd_max_values, (simd_i_t) a, res); simd_max_values = (simd_f_t) srslte_simd_i_select((simd_i_t) simd_max_values, (simd_i_t) a, res);
simd_indexes = srslte_simd_i_add(simd_indexes, simd_inc); simd_indexes = srslte_simd_i_add(simd_indexes, simd_inc);
@ -976,9 +1001,7 @@ uint32_t srslte_vec_max_fi_simd(float *x, int len) {
} else { } else {
for (; i < len - SRSLTE_SIMD_I_SIZE + 1; i += SRSLTE_SIMD_I_SIZE) { for (; i < len - SRSLTE_SIMD_I_SIZE + 1; i += SRSLTE_SIMD_I_SIZE) {
simd_f_t a = srslte_simd_f_loadu(&x[i]); simd_f_t a = srslte_simd_f_loadu(&x[i]);
simd_sel_t res = srslte_simd_f_max(a, simd_max_values); simd_sel_t res = srslte_simd_f_max(a, simd_max_values);
simd_max_indexes = srslte_simd_i_select(simd_max_indexes, simd_indexes, res); simd_max_indexes = srslte_simd_i_select(simd_max_indexes, simd_indexes, res);
simd_max_values = (simd_f_t) srslte_simd_i_select((simd_i_t) simd_max_values, (simd_i_t) a, res); simd_max_values = (simd_f_t) srslte_simd_i_select((simd_i_t) simd_max_values, (simd_i_t) a, res);
simd_indexes = srslte_simd_i_add(simd_indexes, simd_inc); simd_indexes = srslte_simd_i_add(simd_indexes, simd_inc);

Write Preview
Loading…
Cancel
Save