57#ifndef INCLUDED_volk_32f_binary_slicer_32i_H
58#define INCLUDED_volk_32f_binary_slicer_32i_H
65 unsigned int num_points)
68 const float* aPtr = aVector;
69 unsigned int number = 0;
71 for (number = 0; number < num_points; number++) {
86 unsigned int num_points)
89 const float* aPtr = aVector;
90 unsigned int number = 0;
92 for (number = 0; number < num_points; number++) {
93 *cPtr++ = (*aPtr++ >= 0);
100#include <emmintrin.h>
103 const float* aVector,
104 unsigned int num_points)
107 const float* aPtr = aVector;
108 unsigned int number = 0;
110 unsigned int quarter_points = num_points / 4;
116 for (number = 0; number < quarter_points; number++) {
129 for (number = quarter_points * 4; number < num_points; number++) {
141#include <immintrin.h>
144 const float* aVector,
145 unsigned int num_points)
148 const float* aPtr = aVector;
149 unsigned int number = 0;
151 unsigned int quarter_points = num_points / 8;
152 __m256 a_val, res_f, binary_f;
154 __m256 zero_val, one_val;
155 zero_val = _mm256_set1_ps(0.0f);
156 one_val = _mm256_set1_ps(1.0f);
158 for (number = 0; number < quarter_points; number++) {
159 a_val = _mm256_load_ps(aPtr);
161 res_f = _mm256_cmp_ps(a_val, zero_val, _CMP_GE_OS);
162 binary_f = _mm256_and_ps(res_f, one_val);
163 binary_i = _mm256_cvtps_epi32(binary_f);
165 _mm256_store_si256((__m256i*)cPtr, binary_i);
171 for (number = quarter_points * 8; number < num_points; number++) {
183#include <emmintrin.h>
186 const float* aVector,
187 unsigned int num_points)
190 const float* aPtr = aVector;
191 unsigned int number = 0;
193 unsigned int quarter_points = num_points / 4;
199 for (number = 0; number < quarter_points; number++) {
212 for (number = quarter_points * 4; number < num_points; number++) {
224#include <immintrin.h>
227 const float* aVector,
228 unsigned int num_points)
231 const float* aPtr = aVector;
232 unsigned int number = 0;
234 unsigned int quarter_points = num_points / 8;
235 __m256 a_val, res_f, binary_f;
237 __m256 zero_val, one_val;
238 zero_val = _mm256_set1_ps(0.0f);
239 one_val = _mm256_set1_ps(1.0f);
241 for (number = 0; number < quarter_points; number++) {
242 a_val = _mm256_loadu_ps(aPtr);
244 res_f = _mm256_cmp_ps(a_val, zero_val, _CMP_GE_OS);
245 binary_f = _mm256_and_ps(res_f, one_val);
246 binary_i = _mm256_cvtps_epi32(binary_f);
248 _mm256_storeu_si256((__m256i*)cPtr, binary_i);
254 for (number = quarter_points * 8; number < num_points; number++) {