23 #ifndef INCLUDED_volk_32f_x2_divide_32f_a_H
24 #define INCLUDED_volk_32f_x2_divide_32f_a_H
30 #include <xmmintrin.h>
38 static inline void volk_32f_x2_divide_32f_a_sse(
float* cVector,
const float* aVector,
const float* bVector,
unsigned int num_points){
39 unsigned int number = 0;
40 const unsigned int quarterPoints = num_points / 4;
42 float* cPtr = cVector;
43 const float* aPtr = aVector;
44 const float* bPtr= bVector;
46 __m128 aVal, bVal, cVal;
47 for(;number < quarterPoints; number++){
49 aVal = _mm_load_ps(aPtr);
50 bVal = _mm_load_ps(bPtr);
52 cVal = _mm_div_ps(aVal, bVal);
54 _mm_store_ps(cPtr,cVal);
61 number = quarterPoints * 4;
62 for(;number < num_points; number++){
63 *cPtr++ = (*aPtr++) / (*bPtr++);
68 #ifdef LV_HAVE_GENERIC
76 static inline void volk_32f_x2_divide_32f_generic(
float* cVector,
const float* aVector,
const float* bVector,
unsigned int num_points){
77 float* cPtr = cVector;
78 const float* aPtr = aVector;
79 const float* bPtr= bVector;
80 unsigned int number = 0;
82 for(number = 0; number < num_points; number++){
83 *cPtr++ = (*aPtr++) / (*bPtr++);
96 extern void volk_32f_x2_divide_32f_a_orc_impl(
float* cVector,
const float* aVector,
const float* bVector,
unsigned int num_points);
97 static inline void volk_32f_x2_divide_32f_u_orc(
float* cVector,
const float* aVector,
const float* bVector,
unsigned int num_points){
98 volk_32f_x2_divide_32f_a_orc_impl(cVector, aVector, bVector, num_points);