53#ifndef INCLUDED_volk_32f_convert_64f_u_H
54#define INCLUDED_volk_32f_convert_64f_u_H
63 const float* inputVector,
64 unsigned int num_points)
66 unsigned int number = 0;
68 const unsigned int quarterPoints = num_points / 4;
70 const float* inputVectorPtr = (
const float*)inputVector;
71 double* outputVectorPtr = outputVector;
75 for (; number < quarterPoints; number++) {
76 inputVal = _mm_loadu_ps(inputVectorPtr);
79 ret = _mm256_cvtps_pd(inputVal);
80 _mm256_storeu_pd(outputVectorPtr, ret);
85 number = quarterPoints * 4;
86 for (; number < num_points; number++) {
87 outputVector[number] = (double)(inputVector[number]);
97 const float* inputVector,
98 unsigned int num_points)
100 unsigned int number = 0;
102 const unsigned int quarterPoints = num_points / 4;
104 const float* inputVectorPtr = (
const float*)inputVector;
105 double* outputVectorPtr = outputVector;
109 for (; number < quarterPoints; number++) {
110 inputVal = _mm_loadu_ps(inputVectorPtr);
113 ret = _mm_cvtps_pd(inputVal);
115 _mm_storeu_pd(outputVectorPtr, ret);
116 outputVectorPtr += 2;
118 inputVal = _mm_movehl_ps(inputVal, inputVal);
120 ret = _mm_cvtps_pd(inputVal);
122 _mm_storeu_pd(outputVectorPtr, ret);
123 outputVectorPtr += 2;
126 number = quarterPoints * 4;
127 for (; number < num_points; number++) {
128 outputVector[number] = (double)(inputVector[number]);
134#ifdef LV_HAVE_GENERIC
137 const float* inputVector,
138 unsigned int num_points)
140 double* outputVectorPtr = outputVector;
141 const float* inputVectorPtr = inputVector;
142 unsigned int number = 0;
144 for (number = 0; number < num_points; number++) {
145 *outputVectorPtr++ = ((double)(*inputVectorPtr++));
154#ifndef INCLUDED_volk_32f_convert_64f_a_H
155#define INCLUDED_volk_32f_convert_64f_a_H
161#include <immintrin.h>
164 const float* inputVector,
165 unsigned int num_points)
167 unsigned int number = 0;
169 const unsigned int quarterPoints = num_points / 4;
171 const float* inputVectorPtr = (
const float*)inputVector;
172 double* outputVectorPtr = outputVector;
176 for (; number < quarterPoints; number++) {
177 inputVal = _mm_load_ps(inputVectorPtr);
180 ret = _mm256_cvtps_pd(inputVal);
181 _mm256_store_pd(outputVectorPtr, ret);
183 outputVectorPtr += 4;
186 number = quarterPoints * 4;
187 for (; number < num_points; number++) {
188 outputVector[number] = (double)(inputVector[number]);
194#include <emmintrin.h>
197 const float* inputVector,
198 unsigned int num_points)
200 unsigned int number = 0;
202 const unsigned int quarterPoints = num_points / 4;
204 const float* inputVectorPtr = (
const float*)inputVector;
205 double* outputVectorPtr = outputVector;
209 for (; number < quarterPoints; number++) {
210 inputVal = _mm_load_ps(inputVectorPtr);
213 ret = _mm_cvtps_pd(inputVal);
215 _mm_store_pd(outputVectorPtr, ret);
216 outputVectorPtr += 2;
218 inputVal = _mm_movehl_ps(inputVal, inputVal);
220 ret = _mm_cvtps_pd(inputVal);
222 _mm_store_pd(outputVectorPtr, ret);
223 outputVectorPtr += 2;
226 number = quarterPoints * 4;
227 for (; number < num_points; number++) {
228 outputVector[number] = (double)(inputVector[number]);
234#include <riscv_vector.h>
236static inline void volk_32f_convert_64f_rvv(
double* outputVector,
237 const float* inputVector,
238 unsigned int num_points)
240 size_t n = num_points;
241 for (
size_t vl; n > 0; n -= vl, inputVector += vl, outputVector += vl) {
242 vl = __riscv_vsetvl_e32m4(n);
243 vfloat32m4_t v = __riscv_vle32_v_f32m4(inputVector, vl);
244 __riscv_vse64(outputVector, __riscv_vfwcvt_f(v, vl), vl);
static void volk_32f_convert_64f_u_sse2(double *outputVector, const float *inputVector, unsigned int num_points)
Definition volk_32f_convert_64f.h:96
static void volk_32f_convert_64f_a_avx(double *outputVector, const float *inputVector, unsigned int num_points)
Definition volk_32f_convert_64f.h:163
static void volk_32f_convert_64f_a_sse2(double *outputVector, const float *inputVector, unsigned int num_points)
Definition volk_32f_convert_64f.h:196
static void volk_32f_convert_64f_generic(double *outputVector, const float *inputVector, unsigned int num_points)
Definition volk_32f_convert_64f.h:136
static void volk_32f_convert_64f_u_avx(double *outputVector, const float *inputVector, unsigned int num_points)
Definition volk_32f_convert_64f.h:62