70 #ifndef INCLUDED_volk_32f_s32f_normalize_a_H
71 #define INCLUDED_volk_32f_s32f_normalize_a_H
77 #include <xmmintrin.h>
79 static inline void volk_32f_s32f_normalize_a_sse(
float* vecBuffer,
const float scalar,
unsigned int num_points){
80 unsigned int number = 0;
81 float* inputPtr = vecBuffer;
83 const float invScalar = 1.0 / scalar;
84 __m128 vecScalar = _mm_set_ps1(invScalar);
88 const uint64_t quarterPoints = num_points / 4;
89 for(;number < quarterPoints; number++){
91 input1 = _mm_load_ps(inputPtr);
93 input1 = _mm_mul_ps(input1, vecScalar);
95 _mm_store_ps(inputPtr, input1);
100 number = quarterPoints*4;
101 for(; number < num_points; number++){
102 *inputPtr *= invScalar;
108 #ifdef LV_HAVE_GENERIC
110 static inline void volk_32f_s32f_normalize_generic(
float* vecBuffer,
const float scalar,
unsigned int num_points){
111 unsigned int number = 0;
112 float* inputPtr = vecBuffer;
113 const float invScalar = 1.0 / scalar;
114 for(number = 0; number < num_points; number++){
115 *inputPtr *= invScalar;
123 extern void volk_32f_s32f_normalize_a_orc_impl(
float* dst,
float* src,
const float scalar,
unsigned int num_points);
124 static inline void volk_32f_s32f_normalize_u_orc(
float* vecBuffer,
const float scalar,
unsigned int num_points){
125 float invscalar = 1.0 / scalar;
126 volk_32f_s32f_normalize_a_orc_impl(vecBuffer, vecBuffer, invscalar, num_points);
unsigned __int64 uint64_t
Definition: stdint.h:90