283 lines
9.3 KiB
C
283 lines
9.3 KiB
C
/* SPDX-License-Identifier: MIT
|
|
*
|
|
* Permission is hereby granted, free of charge, to any person
|
|
* obtaining a copy of this software and associated documentation
|
|
* files (the "Software"), to deal in the Software without
|
|
* restriction, including without limitation the rights to use, copy,
|
|
* modify, merge, publish, distribute, sublicense, and/or sell copies
|
|
* of the Software, and to permit persons to whom the Software is
|
|
* furnished to do so, subject to the following conditions:
|
|
*
|
|
* The above copyright notice and this permission notice shall be
|
|
* included in all copies or substantial portions of the Software.
|
|
*
|
|
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
|
|
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
|
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
|
|
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
|
|
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
|
|
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
|
|
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
* SOFTWARE.
|
|
*
|
|
* Copyright:
|
|
* 2020 Evan Nemerson <evan@nemerson.com>
|
|
* 2020 Hidayat Khan <huk2209@gmail.com>
|
|
*/
|
|
|
|
#if !defined(SIMDE_X86_AVX512_SRLV_H)
|
|
#define SIMDE_X86_AVX512_SRLV_H
|
|
|
|
#include "types.h"
|
|
#include "../avx2.h"
|
|
#include "mov.h"
|
|
|
|
HEDLEY_DIAGNOSTIC_PUSH
|
|
SIMDE_DISABLE_UNWANTED_DIAGNOSTICS
|
|
SIMDE_BEGIN_DECLS_
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m128i
|
|
simde_mm_srlv_epi16 (simde__m128i a, simde__m128i b) {
|
|
#if defined(SIMDE_X86_AVX512VL_NATIVE) && defined(SIMDE_X86_AVX512BW_NATIVE)
|
|
return _mm_srlv_epi16(a, b);
|
|
#else
|
|
simde__m128i_private
|
|
a_ = simde__m128i_to_private(a),
|
|
b_ = simde__m128i_to_private(b),
|
|
r_;
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR)
|
|
r_.u16 = HEDLEY_REINTERPRET_CAST(__typeof__(r_.u16), (b_.u16 < 16)) & (a_.u16 >> b_.u16);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.u16) / sizeof(r_.u16[0])) ; i++) {
|
|
r_.u16[i] = (b_.u16[i] < 16) ? (a_.u16[i] >> b_.u16[i]) : 0;
|
|
}
|
|
#endif
|
|
|
|
return simde__m128i_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512BW_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm_srlv_epi16
|
|
#define _mm_srlv_epi16(a, b) simde_mm_srlv_epi16(a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m128i
|
|
simde_mm_mask_srlv_epi16(simde__m128i src, simde__mmask8 k, simde__m128i a, simde__m128i b) {
|
|
#if defined(SIMDE_X86_AVX512VL_NATIVE) && defined(SIMDE_X86_AVX512BW_NATIVE)
|
|
return _mm_mask_srlv_epi16(src, k, a, b);
|
|
#else
|
|
return simde_mm_mask_mov_epi16(src, k, simde_mm_srlv_epi16(a, b));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512VL_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm_mask_srlv_epi16
|
|
#define _mm_mask_srlv_epi16(src, k, a, b) simde_mm_mask_srlv_epi16(src, k, a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m128i
|
|
simde_mm_maskz_srlv_epi16(simde__mmask8 k, simde__m128i a, simde__m128i b) {
|
|
#if defined(SIMDE_X86_AVX512VL_NATIVE) && defined(SIMDE_X86_AVX512BW_NATIVE)
|
|
return _mm_maskz_srlv_epi16(k, a, b);
|
|
#else
|
|
return simde_mm_maskz_mov_epi16(k, simde_mm_srlv_epi16(a, b));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512VL_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm_maskz_srlv_epi16
|
|
#define _mm_maskz_srlv_epi16(k, a, b) simde_mm_maskz_srlv_epi16(k, a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m128i
|
|
simde_mm_mask_srlv_epi32(simde__m128i src, simde__mmask8 k, simde__m128i a, simde__m128i b) {
|
|
#if defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
return _mm_mask_srlv_epi32(src, k, a, b);
|
|
#else
|
|
return simde_mm_mask_mov_epi32(src, k, simde_mm_srlv_epi32(a, b));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512VL_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm_mask_srlv_epi32
|
|
#define _mm_mask_srlv_epi32(src, k, a, b) simde_mm_mask_srlv_epi32(src, k, a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m128i
|
|
simde_mm_maskz_srlv_epi32(simde__mmask8 k, simde__m128i a, simde__m128i b) {
|
|
#if defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
return _mm_maskz_srlv_epi32(k, a, b);
|
|
#else
|
|
return simde_mm_maskz_mov_epi32(k, simde_mm_srlv_epi32(a, b));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512VL_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm_maskz_srlv_epi32
|
|
#define _mm_maskz_srlv_epi32(k, a, b) simde_mm_maskz_srlv_epi32(k, a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m128i
|
|
simde_mm_mask_srlv_epi64(simde__m128i src, simde__mmask8 k, simde__m128i a, simde__m128i b) {
|
|
#if defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
return _mm_mask_srlv_epi64(src, k, a, b);
|
|
#else
|
|
return simde_mm_mask_mov_epi64(src, k, simde_mm_srlv_epi64(a, b));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512VL_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm_mask_srlv_epi64
|
|
#define _mm_mask_srlv_epi64(src, k, a, b) simde_mm_mask_srlv_epi64(src, k, a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m128i
|
|
simde_mm_maskz_srlv_epi64(simde__mmask8 k, simde__m128i a, simde__m128i b) {
|
|
#if defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
return _mm_maskz_srlv_epi64(k, a, b);
|
|
#else
|
|
return simde_mm_maskz_mov_epi64(k, simde_mm_srlv_epi64(a, b));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512VL_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm_maskz_srlv_epi64
|
|
#define _mm_maskz_srlv_epi64(k, a, b) simde_mm_maskz_srlv_epi64(k, a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m256i
|
|
simde_mm256_srlv_epi16 (simde__m256i a, simde__m256i b) {
|
|
#if defined(SIMDE_X86_AVX512VL_NATIVE) && defined(SIMDE_X86_AVX512BW_NATIVE)
|
|
return _mm256_srlv_epi16(a, b);
|
|
#else
|
|
simde__m256i_private
|
|
a_ = simde__m256i_to_private(a),
|
|
b_ = simde__m256i_to_private(b),
|
|
r_;
|
|
|
|
#if SIMDE_NATURAL_VECTOR_SIZE_LE(128)
|
|
for (size_t i = 0 ; i < (sizeof(r_.m128i) / sizeof(r_.m128i[0])) ; i++) {
|
|
r_.m128i[i] = simde_mm_srlv_epi16(a_.m128i[i], b_.m128i[i]);
|
|
}
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR)
|
|
r_.u16 = HEDLEY_REINTERPRET_CAST(__typeof__(r_.u16), (b_.u16 < 16)) & (a_.u16 >> b_.u16);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.u16) / sizeof(r_.u16[0])) ; i++) {
|
|
r_.u16[i] = (b_.u16[i] < 16) ? (a_.u16[i] >> b_.u16[i]) : 0;
|
|
}
|
|
#endif
|
|
|
|
return simde__m256i_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512BW_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm256_srlv_epi16
|
|
#define _mm256_srlv_epi16(a, b) simde_mm256_srlv_epi16(a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m512i
|
|
simde_mm512_srlv_epi16 (simde__m512i a, simde__m512i b) {
|
|
#if defined(SIMDE_X86_AVX512BW_NATIVE)
|
|
return _mm512_srlv_epi16(a, b);
|
|
#else
|
|
simde__m512i_private
|
|
a_ = simde__m512i_to_private(a),
|
|
b_ = simde__m512i_to_private(b),
|
|
r_;
|
|
|
|
#if SIMDE_NATURAL_VECTOR_SIZE_LE(256)
|
|
for (size_t i = 0 ; i < (sizeof(r_.m256i) / sizeof(r_.m256i[0])) ; i++) {
|
|
r_.m256i[i] = simde_mm256_srlv_epi16(a_.m256i[i], b_.m256i[i]);
|
|
}
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR)
|
|
r_.u16 = HEDLEY_REINTERPRET_CAST(__typeof__(r_.u16), (b_.u16 < 16)) & (a_.u16 >> b_.u16);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.u16) / sizeof(r_.u16[0])) ; i++) {
|
|
r_.u16[i] = (b_.u16[i] < 16) ? (a_.u16[i] >> b_.u16[i]) : 0;
|
|
}
|
|
#endif
|
|
|
|
return simde__m512i_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512BW_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm512_srlv_epi16
|
|
#define _mm512_srlv_epi16(a, b) simde_mm512_srlv_epi16(a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m512i
|
|
simde_mm512_srlv_epi32 (simde__m512i a, simde__m512i b) {
|
|
#if defined(SIMDE_X86_AVX512F_NATIVE)
|
|
return _mm512_srlv_epi32(a, b);
|
|
#else
|
|
simde__m512i_private
|
|
a_ = simde__m512i_to_private(a),
|
|
b_ = simde__m512i_to_private(b),
|
|
r_;
|
|
|
|
#if SIMDE_NATURAL_VECTOR_SIZE_LE(256)
|
|
for (size_t i = 0 ; i < (sizeof(r_.m256i) / sizeof(r_.m256i[0])) ; i++) {
|
|
r_.m256i[i] = simde_mm256_srlv_epi32(a_.m256i[i], b_.m256i[i]);
|
|
}
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR)
|
|
r_.u32 = HEDLEY_REINTERPRET_CAST(__typeof__(r_.u32), (b_.u32 < 32)) & (a_.u32 >> b_.u32);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.u32) / sizeof(r_.u32[0])) ; i++) {
|
|
r_.u32[i] = (b_.u32[i] < 32) ? (a_.u32[i] >> b_.u32[i]) : 0;
|
|
}
|
|
#endif
|
|
|
|
return simde__m512i_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512F_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm512_srlv_epi32
|
|
#define _mm512_srlv_epi32(a, b) simde_mm512_srlv_epi32(a, b)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde__m512i
|
|
simde_mm512_srlv_epi64 (simde__m512i a, simde__m512i b) {
|
|
#if defined(SIMDE_X86_AVX512F_NATIVE)
|
|
return _mm512_srlv_epi64(a, b);
|
|
#else
|
|
simde__m512i_private
|
|
a_ = simde__m512i_to_private(a),
|
|
b_ = simde__m512i_to_private(b),
|
|
r_;
|
|
|
|
#if SIMDE_NATURAL_VECTOR_SIZE_LE(256)
|
|
for (size_t i = 0 ; i < (sizeof(r_.m256i) / sizeof(r_.m256i[0])) ; i++) {
|
|
r_.m256i[i] = simde_mm256_srlv_epi64(a_.m256i[i], b_.m256i[i]);
|
|
}
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_SCALAR)
|
|
r_.u64 = HEDLEY_REINTERPRET_CAST(__typeof__(r_.u64), (b_.u64 < 64)) & (a_.u64 >> b_.u64);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.u64) / sizeof(r_.u64[0])) ; i++) {
|
|
r_.u64[i] = (b_.u64[i] < 64) ? (a_.u64[i] >> b_.u64[i]) : 0;
|
|
}
|
|
#endif
|
|
|
|
return simde__m512i_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_X86_AVX512F_ENABLE_NATIVE_ALIASES)
|
|
#undef _mm512_srlv_epi64
|
|
#define _mm512_srlv_epi64(a, b) simde_mm512_srlv_epi64(a, b)
|
|
#endif
|
|
|
|
SIMDE_END_DECLS_
|
|
HEDLEY_DIAGNOSTIC_POP
|
|
|
|
#endif /* !defined(SIMDE_X86_AVX512_SRLV_H) */
|