1 /* SPDX-License-Identifier: MIT
2 *
3 * Permission is hereby granted, free of charge, to any person
4 * obtaining a copy of this software and associated documentation
5 * files (the "Software"), to deal in the Software without
6 * restriction, including without limitation the rights to use, copy,
7 * modify, merge, publish, distribute, sublicense, and/or sell copies
8 * of the Software, and to permit persons to whom the Software is
9 * furnished to do so, subject to the following conditions:
10 *
11 * The above copyright notice and this permission notice shall be
12 * included in all copies or substantial portions of the Software.
13 *
14 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
15 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
16 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
17 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
18 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
19 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
20 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21 * SOFTWARE.
22 *
23 * Copyright:
24 * 2020 Evan Nemerson <evan@nemerson.com>
25 */
26
27 #if !defined(SIMDE_ARM_NEON_BIC_H)
28 #define SIMDE_ARM_NEON_BIC_H
29
30 #include "dup_n.h"
31 #include "types.h"
32
33 HEDLEY_DIAGNOSTIC_PUSH
34 SIMDE_DISABLE_UNWANTED_DIAGNOSTICS
35 SIMDE_BEGIN_DECLS_
36
37 SIMDE_FUNCTION_ATTRIBUTES
38 simde_int8x8_t
simde_vbic_s8(simde_int8x8_t a,simde_int8x8_t b)39 simde_vbic_s8(simde_int8x8_t a, simde_int8x8_t b) {
40 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
41 return vbic_s8(a, b);
42 #elif defined(SIMDE_X86_MMX_NATIVE)
43 return _mm_andnot_si64(b, a);
44 #else
45 simde_int8x8_private
46 a_ = simde_int8x8_to_private(a),
47 b_ = simde_int8x8_to_private(b),
48 r_;
49
50 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
51 r_.values[i] = a_.values[i] & ~b_.values[i];
52 }
53
54 return simde_int8x8_from_private(r_);
55 #endif
56 }
57 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
58 #undef vbic_s8
59 #define vbic_s8(a, b) simde_vbic_s8((a), (b))
60 #endif
61
62 SIMDE_FUNCTION_ATTRIBUTES
63 simde_int16x4_t
simde_vbic_s16(simde_int16x4_t a,simde_int16x4_t b)64 simde_vbic_s16(simde_int16x4_t a, simde_int16x4_t b) {
65 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
66 return vbic_s16(a, b);
67 #elif defined(SIMDE_X86_MMX_NATIVE)
68 return _mm_andnot_si64(b, a);
69 #else
70 simde_int16x4_private
71 a_ = simde_int16x4_to_private(a),
72 b_ = simde_int16x4_to_private(b),
73 r_;
74
75 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
76 r_.values[i] = a_.values[i] & ~b_.values[i];
77 }
78
79 return simde_int16x4_from_private(r_);
80 #endif
81 }
82 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
83 #undef vbic_s16
84 #define vbic_s16(a, b) simde_vbic_s16((a), (b))
85 #endif
86
87 SIMDE_FUNCTION_ATTRIBUTES
88 simde_int32x2_t
simde_vbic_s32(simde_int32x2_t a,simde_int32x2_t b)89 simde_vbic_s32(simde_int32x2_t a, simde_int32x2_t b) {
90 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
91 return vbic_s32(a, b);
92 #elif defined(SIMDE_X86_MMX_NATIVE)
93 return _mm_andnot_si64(b, a);
94 #else
95 simde_int32x2_private
96 a_ = simde_int32x2_to_private(a),
97 b_ = simde_int32x2_to_private(b),
98 r_;
99
100 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
101 r_.values[i] = a_.values[i] & ~b_.values[i];
102 }
103
104 return simde_int32x2_from_private(r_);
105 #endif
106 }
107 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
108 #undef vbic_s32
109 #define vbic_s32(a, b) simde_vbic_s32((a), (b))
110 #endif
111
112 SIMDE_FUNCTION_ATTRIBUTES
113 simde_int64x1_t
simde_vbic_s64(simde_int64x1_t a,simde_int64x1_t b)114 simde_vbic_s64(simde_int64x1_t a, simde_int64x1_t b) {
115 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
116 return vbic_s64(a, b);
117 #elif defined(SIMDE_X86_MMX_NATIVE)
118 return _mm_andnot_si64(b, a);
119 #else
120 simde_int64x1_private
121 a_ = simde_int64x1_to_private(a),
122 b_ = simde_int64x1_to_private(b),
123 r_;
124
125 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
126 r_.values[i] = a_.values[i] & ~b_.values[i];
127 }
128
129 return simde_int64x1_from_private(r_);
130 #endif
131 }
132 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
133 #undef vbic_s64
134 #define vbic_s64(a, b) simde_vbic_s64((a), (b))
135 #endif
136
137 SIMDE_FUNCTION_ATTRIBUTES
138 simde_uint8x8_t
simde_vbic_u8(simde_uint8x8_t a,simde_uint8x8_t b)139 simde_vbic_u8(simde_uint8x8_t a, simde_uint8x8_t b) {
140 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
141 return vbic_u8(a, b);
142 #elif defined(SIMDE_X86_MMX_NATIVE)
143 return _mm_andnot_si64(b, a);
144 #else
145 simde_uint8x8_private
146 a_ = simde_uint8x8_to_private(a),
147 b_ = simde_uint8x8_to_private(b),
148 r_;
149
150 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
151 r_.values[i] = a_.values[i] & ~b_.values[i];
152 }
153
154 return simde_uint8x8_from_private(r_);
155 #endif
156 }
157 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
158 #undef vbic_u8
159 #define vbic_u8(a, b) simde_vbic_u8((a), (b))
160 #endif
161
162 SIMDE_FUNCTION_ATTRIBUTES
163 simde_uint16x4_t
simde_vbic_u16(simde_uint16x4_t a,simde_uint16x4_t b)164 simde_vbic_u16(simde_uint16x4_t a, simde_uint16x4_t b) {
165 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
166 return vbic_u16(a, b);
167 #elif defined(SIMDE_X86_MMX_NATIVE)
168 return _mm_andnot_si64(b, a);
169 #else
170 simde_uint16x4_private
171 a_ = simde_uint16x4_to_private(a),
172 b_ = simde_uint16x4_to_private(b),
173 r_;
174
175 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
176 r_.values[i] = a_.values[i] & ~b_.values[i];
177 }
178
179 return simde_uint16x4_from_private(r_);
180 #endif
181 }
182 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
183 #undef vbic_u16
184 #define vbic_u16(a, b) simde_vbic_u16((a), (b))
185 #endif
186
187 SIMDE_FUNCTION_ATTRIBUTES
188 simde_uint32x2_t
simde_vbic_u32(simde_uint32x2_t a,simde_uint32x2_t b)189 simde_vbic_u32(simde_uint32x2_t a, simde_uint32x2_t b) {
190 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
191 return vbic_u32(a, b);
192 #elif defined(SIMDE_X86_MMX_NATIVE)
193 return _mm_andnot_si64(b, a);
194 #else
195 simde_uint32x2_private
196 a_ = simde_uint32x2_to_private(a),
197 b_ = simde_uint32x2_to_private(b),
198 r_;
199
200 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
201 r_.values[i] = a_.values[i] & ~b_.values[i];
202 }
203
204 return simde_uint32x2_from_private(r_);
205 #endif
206 }
207 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
208 #undef vbic_u32
209 #define vbic_u32(a, b) simde_vbic_u32((a), (b))
210 #endif
211
212 SIMDE_FUNCTION_ATTRIBUTES
213 simde_uint64x1_t
simde_vbic_u64(simde_uint64x1_t a,simde_uint64x1_t b)214 simde_vbic_u64(simde_uint64x1_t a, simde_uint64x1_t b) {
215 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
216 return vbic_u64(a, b);
217 #elif defined(SIMDE_X86_MMX_NATIVE)
218 return _mm_andnot_si64(b, a);
219 #else
220 simde_uint64x1_private
221 a_ = simde_uint64x1_to_private(a),
222 b_ = simde_uint64x1_to_private(b),
223 r_;
224
225 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
226 r_.values[i] = a_.values[i] & ~b_.values[i];
227 }
228
229 return simde_uint64x1_from_private(r_);
230 #endif
231 }
232 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
233 #undef vbic_u64
234 #define vbic_u64(a, b) simde_vbic_u64((a), (b))
235 #endif
236
237 SIMDE_FUNCTION_ATTRIBUTES
238 simde_int8x16_t
simde_vbicq_s8(simde_int8x16_t a,simde_int8x16_t b)239 simde_vbicq_s8(simde_int8x16_t a, simde_int8x16_t b) {
240 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
241 return vbicq_s8(a, b);
242 #elif defined(SIMDE_X86_SSE2_NATIVE)
243 return _mm_andnot_si128(b, a);
244 #elif defined(SIMDE_WASM_SIMD128_NATIVE)
245 return wasm_v128_andnot(a, b);
246 #elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
247 return vec_andc(a, b);
248 #else
249 simde_int8x16_private
250 a_ = simde_int8x16_to_private(a),
251 b_ = simde_int8x16_to_private(b),
252 r_;
253
254 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
255 r_.values[i] = a_.values[i] & ~b_.values[i];
256 }
257
258 return simde_int8x16_from_private(r_);
259 #endif
260 }
261 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
262 #undef vbicq_s8
263 #define vbicq_s8(a, b) simde_vbicq_s8((a), (b))
264 #endif
265
266 SIMDE_FUNCTION_ATTRIBUTES
267 simde_int16x8_t
simde_vbicq_s16(simde_int16x8_t a,simde_int16x8_t b)268 simde_vbicq_s16(simde_int16x8_t a, simde_int16x8_t b) {
269 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
270 return vbicq_s16(a, b);
271 #elif defined(SIMDE_X86_SSE2_NATIVE)
272 return _mm_andnot_si128(b, a);
273 #elif defined(SIMDE_WASM_SIMD128_NATIVE)
274 return wasm_v128_andnot(a, b);
275 #elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
276 return vec_andc(a, b);
277 #else
278 simde_int16x8_private
279 a_ = simde_int16x8_to_private(a),
280 b_ = simde_int16x8_to_private(b),
281 r_;
282
283 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
284 r_.values[i] = a_.values[i] & ~b_.values[i];
285 }
286
287 return simde_int16x8_from_private(r_);
288 #endif
289 }
290 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
291 #undef vbicq_s16
292 #define vbicq_s16(a, b) simde_vbicq_s16((a), (b))
293 #endif
294
295 SIMDE_FUNCTION_ATTRIBUTES
296 simde_int32x4_t
simde_vbicq_s32(simde_int32x4_t a,simde_int32x4_t b)297 simde_vbicq_s32(simde_int32x4_t a, simde_int32x4_t b) {
298 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
299 return vbicq_s32(a, b);
300 #elif defined(SIMDE_X86_SSE2_NATIVE)
301 return _mm_andnot_si128(b, a);
302 #elif defined(SIMDE_WASM_SIMD128_NATIVE)
303 return wasm_v128_andnot(a, b);
304 #elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
305 return vec_andc(a, b);
306 #else
307 simde_int32x4_private
308 a_ = simde_int32x4_to_private(a),
309 b_ = simde_int32x4_to_private(b),
310 r_;
311
312 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
313 r_.values[i] = a_.values[i] & ~b_.values[i];
314 }
315
316 return simde_int32x4_from_private(r_);
317 #endif
318 }
319 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
320 #undef vbicq_s32
321 #define vbicq_s32(a, b) simde_vbicq_s32((a), (b))
322 #endif
323
324 SIMDE_FUNCTION_ATTRIBUTES
325 simde_int64x2_t
simde_vbicq_s64(simde_int64x2_t a,simde_int64x2_t b)326 simde_vbicq_s64(simde_int64x2_t a, simde_int64x2_t b) {
327 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
328 return vbicq_s64(a, b);
329 #elif defined(SIMDE_X86_SSE2_NATIVE)
330 return _mm_andnot_si128(b, a);
331 #elif defined(SIMDE_WASM_SIMD128_NATIVE)
332 return wasm_v128_andnot(a, b);
333 #elif defined(SIMDE_POWER_ALTIVEC_P7_NATIVE)
334 return vec_andc(a, b);
335 #else
336 simde_int64x2_private
337 a_ = simde_int64x2_to_private(a),
338 b_ = simde_int64x2_to_private(b),
339 r_;
340
341 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
342 r_.values[i] = a_.values[i] & ~b_.values[i];
343 }
344
345 return simde_int64x2_from_private(r_);
346 #endif
347 }
348 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
349 #undef vbicq_s64
350 #define vbicq_s64(a, b) simde_vbicq_s64((a), (b))
351 #endif
352
353 SIMDE_FUNCTION_ATTRIBUTES
354 simde_uint8x16_t
simde_vbicq_u8(simde_uint8x16_t a,simde_uint8x16_t b)355 simde_vbicq_u8(simde_uint8x16_t a, simde_uint8x16_t b) {
356 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
357 return vbicq_u8(a, b);
358 #elif defined(SIMDE_X86_SSE2_NATIVE)
359 return _mm_andnot_si128(b, a);
360 #elif defined(SIMDE_WASM_SIMD128_NATIVE)
361 return wasm_v128_andnot(a, b);
362 #elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
363 return vec_andc(a, b);
364 #else
365 simde_uint8x16_private
366 a_ = simde_uint8x16_to_private(a),
367 b_ = simde_uint8x16_to_private(b),
368 r_;
369
370 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
371 r_.values[i] = a_.values[i] & ~b_.values[i];
372 }
373
374 return simde_uint8x16_from_private(r_);
375 #endif
376 }
377 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
378 #undef vbicq_u8
379 #define vbicq_u8(a, b) simde_vbicq_u8((a), (b))
380 #endif
381
382 SIMDE_FUNCTION_ATTRIBUTES
383 simde_uint16x8_t
simde_vbicq_u16(simde_uint16x8_t a,simde_uint16x8_t b)384 simde_vbicq_u16(simde_uint16x8_t a, simde_uint16x8_t b) {
385 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
386 return vbicq_u16(a, b);
387 #elif defined(SIMDE_X86_SSE2_NATIVE)
388 return _mm_andnot_si128(b, a);
389 #elif defined(SIMDE_WASM_SIMD128_NATIVE)
390 return wasm_v128_andnot(a, b);
391 #elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
392 return vec_andc(a, b);
393 #else
394 simde_uint16x8_private
395 a_ = simde_uint16x8_to_private(a),
396 b_ = simde_uint16x8_to_private(b),
397 r_;
398
399 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
400 r_.values[i] = a_.values[i] & ~b_.values[i];
401 }
402
403 return simde_uint16x8_from_private(r_);
404 #endif
405 }
406 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
407 #undef vbicq_u16
408 #define vbicq_u16(a, b) simde_vbicq_u16((a), (b))
409 #endif
410
411 SIMDE_FUNCTION_ATTRIBUTES
412 simde_uint32x4_t
simde_vbicq_u32(simde_uint32x4_t a,simde_uint32x4_t b)413 simde_vbicq_u32(simde_uint32x4_t a, simde_uint32x4_t b) {
414 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
415 return vbicq_u32(a, b);
416 #elif defined(SIMDE_X86_SSE2_NATIVE)
417 return _mm_andnot_si128(b, a);
418 #elif defined(SIMDE_WASM_SIMD128_NATIVE)
419 return wasm_v128_andnot(a, b);
420 #elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
421 return vec_andc(a, b);
422 #else
423 simde_uint32x4_private
424 a_ = simde_uint32x4_to_private(a),
425 b_ = simde_uint32x4_to_private(b),
426 r_;
427
428 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
429 r_.values[i] = a_.values[i] & ~b_.values[i];
430 }
431
432 return simde_uint32x4_from_private(r_);
433 #endif
434 }
435 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
436 #undef vbicq_u32
437 #define vbicq_u32(a, b) simde_vbicq_u32((a), (b))
438 #endif
439
440 SIMDE_FUNCTION_ATTRIBUTES
441 simde_uint64x2_t
simde_vbicq_u64(simde_uint64x2_t a,simde_uint64x2_t b)442 simde_vbicq_u64(simde_uint64x2_t a, simde_uint64x2_t b) {
443 #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
444 return vbicq_u64(a, b);
445 #elif defined(SIMDE_X86_SSE2_NATIVE)
446 return _mm_andnot_si128(b, a);
447 #elif defined(SIMDE_WASM_SIMD128_NATIVE)
448 return wasm_v128_andnot(a, b);
449 #elif defined(SIMDE_POWER_ALTIVEC_P7_NATIVE)
450 return vec_andc(a, b);
451 #else
452 simde_uint64x2_private
453 a_ = simde_uint64x2_to_private(a),
454 b_ = simde_uint64x2_to_private(b),
455 r_;
456
457 for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
458 r_.values[i] = a_.values[i] & ~b_.values[i];
459 }
460
461 return simde_uint64x2_from_private(r_);
462 #endif
463 }
464 #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
465 #undef vbicq_u64
466 #define vbicq_u64(a, b) simde_vbicq_u64((a), (b))
467 #endif
468
469 SIMDE_END_DECLS_
470 HEDLEY_DIAGNOSTIC_POP
471
472 #endif /* !defined(SIMDE_ARM_NEON_BIC_H) */
473