1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
|
#ifndef INCLUDED_volk_32f_s32f_convert_32i_u_H
#define INCLUDED_volk_32f_s32f_convert_32i_u_H
#include <inttypes.h>
#include <stdio.h>
#ifdef LV_HAVE_SSE2
#include <emmintrin.h>
/*!
\brief Multiplies each point in the input buffer by the scalar value, then converts the result into a 32 bit integer value
\param inputVector The floating point input data buffer
\param outputVector The 32 bit output data buffer
\param scalar The value multiplied against each point in the input buffer
\param num_points The number of data values to be converted
\note Input buffer does NOT need to be properly aligned
*/
static inline void volk_32f_s32f_convert_32i_u_sse2(int32_t* outputVector, const float* inputVector, const float scalar, unsigned int num_points){
unsigned int number = 0;
const unsigned int quarterPoints = num_points / 4;
const float* inputVectorPtr = (const float*)inputVector;
int32_t* outputVectorPtr = outputVector;
float min_val = -2147483647;
float max_val = 2147483647;
float r;
__m128 vScalar = _mm_set_ps1(scalar);
__m128 inputVal1;
__m128i intInputVal1;
__m128 vmin_val = _mm_set_ps1(min_val);
__m128 vmax_val = _mm_set_ps1(max_val);
for(;number < quarterPoints; number++){
inputVal1 = _mm_loadu_ps(inputVectorPtr); inputVectorPtr += 4;
inputVal1 = _mm_max_ps(_mm_min_ps(_mm_mul_ps(inputVal1, vScalar), vmax_val), vmin_val);
intInputVal1 = _mm_cvtps_epi32(inputVal1);
_mm_storeu_si128((__m128i*)outputVectorPtr, intInputVal1);
outputVectorPtr += 4;
}
number = quarterPoints * 4;
for(; number < num_points; number++){
r = inputVector[number] * scalar;
if(r > max_val)
r = max_val;
else if(r < min_val)
r = min_val;
outputVector[number] = (int32_t)(r);
}
}
#endif /* LV_HAVE_SSE2 */
#ifdef LV_HAVE_SSE
#include <xmmintrin.h>
/*!
\brief Multiplies each point in the input buffer by the scalar value, then converts the result into a 32 bit integer value
\param inputVector The floating point input data buffer
\param outputVector The 32 bit output data buffer
\param scalar The value multiplied against each point in the input buffer
\param num_points The number of data values to be converted
\note Input buffer does NOT need to be properly aligned
*/
static inline void volk_32f_s32f_convert_32i_u_sse(int32_t* outputVector, const float* inputVector, const float scalar, unsigned int num_points){
unsigned int number = 0;
const unsigned int quarterPoints = num_points / 4;
const float* inputVectorPtr = (const float*)inputVector;
int32_t* outputVectorPtr = outputVector;
float min_val = -2147483647;
float max_val = 2147483647;
float r;
__m128 vScalar = _mm_set_ps1(scalar);
__m128 ret;
__m128 vmin_val = _mm_set_ps1(min_val);
__m128 vmax_val = _mm_set_ps1(max_val);
__VOLK_ATTR_ALIGNED(16) float outputFloatBuffer[4];
for(;number < quarterPoints; number++){
ret = _mm_loadu_ps(inputVectorPtr);
inputVectorPtr += 4;
ret = _mm_max_ps(_mm_min_ps(_mm_mul_ps(ret, vScalar), vmax_val), vmin_val);
_mm_store_ps(outputFloatBuffer, ret);
*outputVectorPtr++ = (int32_t)(outputFloatBuffer[0]);
*outputVectorPtr++ = (int32_t)(outputFloatBuffer[1]);
*outputVectorPtr++ = (int32_t)(outputFloatBuffer[2]);
*outputVectorPtr++ = (int32_t)(outputFloatBuffer[3]);
}
number = quarterPoints * 4;
for(; number < num_points; number++){
r = inputVector[number] * scalar;
if(r > max_val)
r = max_val;
else if(r < min_val)
r = min_val;
outputVector[number] = (int32_t)(r);
}
}
#endif /* LV_HAVE_SSE */
#ifdef LV_HAVE_GENERIC
/*!
\brief Multiplies each point in the input buffer by the scalar value, then converts the result into a 32 bit integer value
\param inputVector The floating point input data buffer
\param outputVector The 32 bit output data buffer
\param scalar The value multiplied against each point in the input buffer
\param num_points The number of data values to be converted
\note Input buffer does NOT need to be properly aligned
*/
static inline void volk_32f_s32f_convert_32i_generic(int32_t* outputVector, const float* inputVector, const float scalar, unsigned int num_points){
int32_t* outputVectorPtr = outputVector;
const float* inputVectorPtr = inputVector;
unsigned int number = 0;
float min_val = -2147483647;
float max_val = 2147483647;
float r;
for(number = 0; number < num_points; number++){
r = *inputVectorPtr++ * scalar;
if(r > max_val)
r = max_val;
else if(r < min_val)
r = min_val;
*outputVectorPtr++ = (int32_t)(r);
}
}
#endif /* LV_HAVE_GENERIC */
#endif /* INCLUDED_volk_32f_s32f_convert_32i_u_H */
#ifndef INCLUDED_volk_32f_s32f_convert_32i_a_H
#define INCLUDED_volk_32f_s32f_convert_32i_a_H
#include <volk/volk_common.h>
#include <inttypes.h>
#include <stdio.h>
#ifdef LV_HAVE_AVX
#include <immintrin.h>
/*!
\brief Multiplies each point in the input buffer by the scalar value, then converts the result into a 32 bit integer value
\param inputVector The floating point input data buffer
\param outputVector The 32 bit output data buffer
\param scalar The value multiplied against each point in the input buffer
\param num_points The number of data values to be converted
*/
static inline void volk_32f_s32f_convert_32i_a_avx(int32_t* outputVector, const float* inputVector, const float scalar, unsigned int num_points){
unsigned int number = 0;
const unsigned int eighthPoints = num_points / 8;
const float* inputVectorPtr = (const float*)inputVector;
int32_t* outputVectorPtr = outputVector;
float min_val = -2147483647;
float max_val = 2147483647;
float r;
__m256 vScalar = _mm256_set1_ps(scalar);
__m256 inputVal1;
__m256i intInputVal1;
__m256 vmin_val = _mm256_set1_ps(min_val);
__m256 vmax_val = _mm256_set1_ps(max_val);
for(;number < eighthPoints; number++){
inputVal1 = _mm256_load_ps(inputVectorPtr); inputVectorPtr += 8;
inputVal1 = _mm256_max_ps(_mm256_min_ps(_mm256_mul_ps(inputVal1, vScalar), vmax_val), vmin_val);
intInputVal1 = _mm256_cvtps_epi32(inputVal1);
_mm256_store_si256((__m256i*)outputVectorPtr, intInputVal1);
outputVectorPtr += 8;
}
number = eighthPoints * 8;
for(; number < num_points; number++){
r = inputVector[number] * scalar;
if(r > max_val)
r = max_val;
else if(r < min_val)
r = min_val;
outputVector[number] = (int32_t)(r);
}
}
#endif /* LV_HAVE_AVX */
#ifdef LV_HAVE_SSE2
#include <emmintrin.h>
/*!
\brief Multiplies each point in the input buffer by the scalar value, then converts the result into a 32 bit integer value
\param inputVector The floating point input data buffer
\param outputVector The 32 bit output data buffer
\param scalar The value multiplied against each point in the input buffer
\param num_points The number of data values to be converted
*/
static inline void volk_32f_s32f_convert_32i_a_sse2(int32_t* outputVector, const float* inputVector, const float scalar, unsigned int num_points){
unsigned int number = 0;
const unsigned int quarterPoints = num_points / 4;
const float* inputVectorPtr = (const float*)inputVector;
int32_t* outputVectorPtr = outputVector;
float min_val = -2147483647;
float max_val = 2147483647;
float r;
__m128 vScalar = _mm_set_ps1(scalar);
__m128 inputVal1;
__m128i intInputVal1;
__m128 vmin_val = _mm_set_ps1(min_val);
__m128 vmax_val = _mm_set_ps1(max_val);
for(;number < quarterPoints; number++){
inputVal1 = _mm_load_ps(inputVectorPtr); inputVectorPtr += 4;
inputVal1 = _mm_max_ps(_mm_min_ps(_mm_mul_ps(inputVal1, vScalar), vmax_val), vmin_val);
intInputVal1 = _mm_cvtps_epi32(inputVal1);
_mm_store_si128((__m128i*)outputVectorPtr, intInputVal1);
outputVectorPtr += 4;
}
number = quarterPoints * 4;
for(; number < num_points; number++){
r = inputVector[number] * scalar;
if(r > max_val)
r = max_val;
else if(r < min_val)
r = min_val;
outputVector[number] = (int32_t)(r);
}
}
#endif /* LV_HAVE_SSE2 */
#ifdef LV_HAVE_SSE
#include <xmmintrin.h>
/*!
\brief Multiplies each point in the input buffer by the scalar value, then converts the result into a 32 bit integer value
\param inputVector The floating point input data buffer
\param outputVector The 32 bit output data buffer
\param scalar The value multiplied against each point in the input buffer
\param num_points The number of data values to be converted
*/
static inline void volk_32f_s32f_convert_32i_a_sse(int32_t* outputVector, const float* inputVector, const float scalar, unsigned int num_points){
unsigned int number = 0;
const unsigned int quarterPoints = num_points / 4;
const float* inputVectorPtr = (const float*)inputVector;
int32_t* outputVectorPtr = outputVector;
float min_val = -2147483647;
float max_val = 2147483647;
float r;
__m128 vScalar = _mm_set_ps1(scalar);
__m128 ret;
__m128 vmin_val = _mm_set_ps1(min_val);
__m128 vmax_val = _mm_set_ps1(max_val);
__VOLK_ATTR_ALIGNED(16) float outputFloatBuffer[4];
for(;number < quarterPoints; number++){
ret = _mm_load_ps(inputVectorPtr);
inputVectorPtr += 4;
ret = _mm_max_ps(_mm_min_ps(_mm_mul_ps(ret, vScalar), vmax_val), vmin_val);
_mm_store_ps(outputFloatBuffer, ret);
*outputVectorPtr++ = (int32_t)(outputFloatBuffer[0]);
*outputVectorPtr++ = (int32_t)(outputFloatBuffer[1]);
*outputVectorPtr++ = (int32_t)(outputFloatBuffer[2]);
*outputVectorPtr++ = (int32_t)(outputFloatBuffer[3]);
}
number = quarterPoints * 4;
for(; number < num_points; number++){
r = inputVector[number] * scalar;
if(r > max_val)
r = max_val;
else if(r < min_val)
r = min_val;
outputVector[number] = (int32_t)(r);
}
}
#endif /* LV_HAVE_SSE */
#ifdef LV_HAVE_GENERIC
/*!
\brief Multiplies each point in the input buffer by the scalar value, then converts the result into a 32 bit integer value
\param inputVector The floating point input data buffer
\param outputVector The 32 bit output data buffer
\param scalar The value multiplied against each point in the input buffer
\param num_points The number of data values to be converted
*/
static inline void volk_32f_s32f_convert_32i_a_generic(int32_t* outputVector, const float* inputVector, const float scalar, unsigned int num_points){
int32_t* outputVectorPtr = outputVector;
const float* inputVectorPtr = inputVector;
unsigned int number = 0;
float min_val = -2147483647;
float max_val = 2147483647;
float r;
for(number = 0; number < num_points; number++){
r = *inputVectorPtr++ * scalar;
if(r > max_val)
r = max_val;
else if(r < min_val)
r = min_val;
*outputVectorPtr++ = (int32_t)(r);
}
}
#endif /* LV_HAVE_GENERIC */
#endif /* INCLUDED_volk_32f_s32f_convert_32i_a_H */
|
