762 lines
26 KiB
C
762 lines
26 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 Sean Maher <seanptmaher@gmail.com> (Copyright owned by Google, LLC)
|
|
*/
|
|
|
|
#if !defined(SIMDE_ARM_NEON_BSL_H)
|
|
#define SIMDE_ARM_NEON_BSL_H
|
|
|
|
#include "types.h"
|
|
#include "reinterpret.h"
|
|
#include "and.h"
|
|
#include "eor.h"
|
|
|
|
HEDLEY_DIAGNOSTIC_PUSH
|
|
SIMDE_DISABLE_UNWANTED_DIAGNOSTICS
|
|
SIMDE_BEGIN_DECLS_
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float16x4_t
|
|
simde_vbsl_f16(simde_uint16x4_t a, simde_float16x4_t b, simde_float16x4_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE) && defined(SIMDE_ARM_NEON_FP16)
|
|
return vbsl_f16(a, b, c);
|
|
#else
|
|
simde_uint16x4_private
|
|
r_,
|
|
a_ = simde_uint16x4_to_private(a),
|
|
b_ = simde_uint16x4_to_private(simde_vreinterpret_u16_f16(b)),
|
|
c_ = simde_uint16x4_to_private(simde_vreinterpret_u16_f16(c));
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpret_f16_u16(simde_uint16x4_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_f16
|
|
#define vbsl_f16(a, b, c) simde_vbsl_f16((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float32x2_t
|
|
simde_vbsl_f32(simde_uint32x2_t a, simde_float32x2_t b, simde_float32x2_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbsl_f32(a, b, c);
|
|
#else
|
|
simde_uint32x2_private
|
|
r_,
|
|
a_ = simde_uint32x2_to_private(a),
|
|
b_ = simde_uint32x2_to_private(simde_vreinterpret_u32_f32(b)),
|
|
c_ = simde_uint32x2_to_private(simde_vreinterpret_u32_f32(c));
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpret_f32_u32(simde_uint32x2_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_f32
|
|
#define vbsl_f32(a, b, c) simde_vbsl_f32((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float64x1_t
|
|
simde_vbsl_f64(simde_uint64x1_t a, simde_float64x1_t b, simde_float64x1_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A64V8_NATIVE)
|
|
return vbsl_f64(a, b, c);
|
|
#else
|
|
simde_uint64x1_private
|
|
r_,
|
|
a_ = simde_uint64x1_to_private(a),
|
|
b_ = simde_uint64x1_to_private(simde_vreinterpret_u64_f64(b)),
|
|
c_ = simde_uint64x1_to_private(simde_vreinterpret_u64_f64(c));
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpret_f64_u64(simde_uint64x1_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A64V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_f64
|
|
#define vbsl_f64(a, b, c) simde_vbsl_f64((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_int8x8_t
|
|
simde_vbsl_s8(simde_uint8x8_t a, simde_int8x8_t b, simde_int8x8_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbsl_s8(a, b, c);
|
|
#else
|
|
simde_uint8x8_private
|
|
r_,
|
|
a_ = simde_uint8x8_to_private(a),
|
|
b_ = simde_uint8x8_to_private(simde_vreinterpret_u8_s8(b)),
|
|
c_ = simde_uint8x8_to_private(simde_vreinterpret_u8_s8(c));
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpret_s8_u8(simde_uint8x8_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_s8
|
|
#define vbsl_s8(a, b, c) simde_vbsl_s8((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_int16x4_t
|
|
simde_vbsl_s16(simde_uint16x4_t a, simde_int16x4_t b, simde_int16x4_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbsl_s16(a, b, c);
|
|
#else
|
|
simde_uint16x4_private
|
|
r_,
|
|
a_ = simde_uint16x4_to_private(a),
|
|
b_ = simde_uint16x4_to_private(simde_vreinterpret_u16_s16(b)),
|
|
c_ = simde_uint16x4_to_private(simde_vreinterpret_u16_s16(c));
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpret_s16_u16(simde_uint16x4_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_s16
|
|
#define vbsl_s16(a, b, c) simde_vbsl_s16((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_int32x2_t
|
|
simde_vbsl_s32(simde_uint32x2_t a, simde_int32x2_t b, simde_int32x2_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbsl_s32(a, b, c);
|
|
#else
|
|
simde_uint32x2_private
|
|
r_,
|
|
a_ = simde_uint32x2_to_private(a),
|
|
b_ = simde_uint32x2_to_private(simde_vreinterpret_u32_s32(b)),
|
|
c_ = simde_uint32x2_to_private(simde_vreinterpret_u32_s32(c));
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpret_s32_u32(simde_uint32x2_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_s32
|
|
#define vbsl_s32(a, b, c) simde_vbsl_s32((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_int64x1_t
|
|
simde_vbsl_s64(simde_uint64x1_t a, simde_int64x1_t b, simde_int64x1_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbsl_s64(a, b, c);
|
|
#else
|
|
simde_uint64x1_private
|
|
r_,
|
|
a_ = simde_uint64x1_to_private(a),
|
|
b_ = simde_uint64x1_to_private(simde_vreinterpret_u64_s64(b)),
|
|
c_ = simde_uint64x1_to_private(simde_vreinterpret_u64_s64(c));
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpret_s64_u64(simde_uint64x1_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_s64
|
|
#define vbsl_s64(a, b, c) simde_vbsl_s64((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_uint8x8_t
|
|
simde_vbsl_u8(simde_uint8x8_t a, simde_uint8x8_t b, simde_uint8x8_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbsl_u8(a, b, c);
|
|
#else
|
|
simde_uint8x8_private
|
|
r_,
|
|
a_ = simde_uint8x8_to_private(a),
|
|
b_ = simde_uint8x8_to_private(b),
|
|
c_ = simde_uint8x8_to_private(c);
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_uint8x8_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_u8
|
|
#define vbsl_u8(a, b, c) simde_vbsl_u8((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_uint16x4_t
|
|
simde_vbsl_u16(simde_uint16x4_t a, simde_uint16x4_t b, simde_uint16x4_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbsl_u16(a, b, c);
|
|
#else
|
|
simde_uint16x4_private
|
|
r_,
|
|
a_ = simde_uint16x4_to_private(a),
|
|
b_ = simde_uint16x4_to_private(b),
|
|
c_ = simde_uint16x4_to_private(c);
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_uint16x4_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_u16
|
|
#define vbsl_u16(a, b, c) simde_vbsl_u16((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_uint32x2_t
|
|
simde_vbsl_u32(simde_uint32x2_t a, simde_uint32x2_t b, simde_uint32x2_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbsl_u32(a, b, c);
|
|
#else
|
|
simde_uint32x2_private
|
|
r_,
|
|
a_ = simde_uint32x2_to_private(a),
|
|
b_ = simde_uint32x2_to_private(b),
|
|
c_ = simde_uint32x2_to_private(c);
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_uint32x2_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_u32
|
|
#define vbsl_u32(a, b, c) simde_vbsl_u32((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_uint64x1_t
|
|
simde_vbsl_u64(simde_uint64x1_t a, simde_uint64x1_t b, simde_uint64x1_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbsl_u64(a, b, c);
|
|
#else
|
|
simde_uint64x1_private
|
|
r_,
|
|
a_ = simde_uint64x1_to_private(a),
|
|
b_ = simde_uint64x1_to_private(b),
|
|
c_ = simde_uint64x1_to_private(c);
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_uint64x1_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbsl_u64
|
|
#define vbsl_u64(a, b, c) simde_vbsl_u64((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float16x8_t
|
|
simde_vbslq_f16(simde_uint16x8_t a, simde_float16x8_t b, simde_float16x8_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE) && defined(SIMDE_ARM_NEON_FP16)
|
|
return vbslq_f16(a, b, c);
|
|
#else
|
|
simde_uint16x8_private
|
|
r_,
|
|
a_ = simde_uint16x8_to_private(a),
|
|
b_ = simde_uint16x8_to_private(simde_vreinterpretq_u16_f16(b)),
|
|
c_ = simde_uint16x8_to_private(simde_vreinterpretq_u16_f16(c));
|
|
|
|
#if defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpretq_f16_u16(simde_uint16x8_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_f16
|
|
#define vbslq_f16(a, b, c) simde_vbslq_f16((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float32x4_t
|
|
simde_vbslq_f32(simde_uint32x4_t a, simde_float32x4_t b, simde_float32x4_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbslq_f32(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_14_NATIVE)
|
|
return vec_sel(c, b, a);
|
|
#else
|
|
simde_uint32x4_private
|
|
r_,
|
|
a_ = simde_uint32x4_to_private(a),
|
|
b_ = simde_uint32x4_to_private(simde_vreinterpretq_u32_f32(b)),
|
|
c_ = simde_uint32x4_to_private(simde_vreinterpretq_u32_f32(c));
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi32(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpretq_f32_u32(simde_uint32x4_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_f32
|
|
#define vbslq_f32(a, b, c) simde_vbslq_f32((a), (b), (c))
|
|
#endif
|
|
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float64x2_t
|
|
simde_vbslq_f64(simde_uint64x2_t a, simde_float64x2_t b, simde_float64x2_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A64V8_NATIVE)
|
|
return vbslq_f64(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P8_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return vec_sel(c, b, a);
|
|
#else
|
|
simde_uint64x2_private
|
|
r_,
|
|
a_ = simde_uint64x2_to_private(a),
|
|
b_ = simde_uint64x2_to_private(simde_vreinterpretq_u64_f64(b)),
|
|
c_ = simde_uint64x2_to_private(simde_vreinterpretq_u64_f64(c));
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi64(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpretq_f64_u64(simde_uint64x2_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A64V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_f64
|
|
#define vbslq_f64(a, b, c) simde_vbslq_f64((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_int8x16_t
|
|
simde_vbslq_s8(simde_uint8x16_t a, simde_int8x16_t b, simde_int8x16_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbslq_s8(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return vec_sel(c, b, a);
|
|
#else
|
|
simde_uint8x16_private
|
|
r_,
|
|
a_ = simde_uint8x16_to_private(a),
|
|
b_ = simde_uint8x16_to_private(simde_vreinterpretq_u8_s8(b)),
|
|
c_ = simde_uint8x16_to_private(simde_vreinterpretq_u8_s8(c));
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi32(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpretq_s8_u8(simde_uint8x16_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_s8
|
|
#define vbslq_s8(a, b, c) simde_vbslq_s8((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_int16x8_t
|
|
simde_vbslq_s16(simde_uint16x8_t a, simde_int16x8_t b, simde_int16x8_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbslq_s16(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return vec_sel(c, b, a);
|
|
#else
|
|
simde_uint16x8_private
|
|
r_,
|
|
a_ = simde_uint16x8_to_private(a),
|
|
b_ = simde_uint16x8_to_private(simde_vreinterpretq_u16_s16(b)),
|
|
c_ = simde_uint16x8_to_private(simde_vreinterpretq_u16_s16(c));
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi32(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpretq_s16_u16(simde_uint16x8_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_s16
|
|
#define vbslq_s16(a, b, c) simde_vbslq_s16((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_int32x4_t
|
|
simde_vbslq_s32(simde_uint32x4_t a, simde_int32x4_t b, simde_int32x4_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbslq_s32(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return vec_sel(c, b, a);
|
|
#else
|
|
simde_uint32x4_private
|
|
r_,
|
|
a_ = simde_uint32x4_to_private(a),
|
|
b_ = simde_uint32x4_to_private(simde_vreinterpretq_u32_s32(b)),
|
|
c_ = simde_uint32x4_to_private(simde_vreinterpretq_u32_s32(c));
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi32(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpretq_s32_u32(simde_uint32x4_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_s32
|
|
#define vbslq_s32(a, b, c) simde_vbslq_s32((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_int64x2_t
|
|
simde_vbslq_s64(simde_uint64x2_t a, simde_int64x2_t b, simde_int64x2_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbslq_s64(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return
|
|
simde_vreinterpretq_s64_s32(
|
|
simde_vbslq_s32(
|
|
simde_vreinterpretq_u32_u64(a),
|
|
simde_vreinterpretq_s32_s64(b),
|
|
simde_vreinterpretq_s32_s64(c)
|
|
)
|
|
);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P8_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return vec_sel(
|
|
HEDLEY_REINTERPRET_CAST(SIMDE_POWER_ALTIVEC_VECTOR(signed long long), c),
|
|
HEDLEY_REINTERPRET_CAST(SIMDE_POWER_ALTIVEC_VECTOR(signed long long), b),
|
|
HEDLEY_REINTERPRET_CAST(SIMDE_POWER_ALTIVEC_VECTOR(unsigned long long), a));
|
|
#else
|
|
simde_uint64x2_private
|
|
r_,
|
|
a_ = simde_uint64x2_to_private(a),
|
|
b_ = simde_uint64x2_to_private(simde_vreinterpretq_u64_s64(b)),
|
|
c_ = simde_uint64x2_to_private(simde_vreinterpretq_u64_s64(c));
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi32(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_vreinterpretq_s64_u64(simde_uint64x2_from_private(r_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_s64
|
|
#define vbslq_s64(a, b, c) simde_vbslq_s64((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_uint8x16_t
|
|
simde_vbslq_u8(simde_uint8x16_t a, simde_uint8x16_t b, simde_uint8x16_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbslq_u8(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return vec_sel(c, b, a);
|
|
#else
|
|
simde_uint8x16_private
|
|
r_,
|
|
a_ = simde_uint8x16_to_private(a),
|
|
b_ = simde_uint8x16_to_private(b),
|
|
c_ = simde_uint8x16_to_private(c);
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi32(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_uint8x16_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_u8
|
|
#define vbslq_u8(a, b, c) simde_vbslq_u8((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_uint16x8_t
|
|
simde_vbslq_u16(simde_uint16x8_t a, simde_uint16x8_t b, simde_uint16x8_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbslq_u16(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return vec_sel(c, b, a);
|
|
#else
|
|
simde_uint16x8_private
|
|
r_,
|
|
a_ = simde_uint16x8_to_private(a),
|
|
b_ = simde_uint16x8_to_private(b),
|
|
c_ = simde_uint16x8_to_private(c);
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi32(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_uint16x8_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_u16
|
|
#define vbslq_u16(a, b, c) simde_vbslq_u16((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_uint32x4_t
|
|
simde_vbslq_u32(simde_uint32x4_t a, simde_uint32x4_t b, simde_uint32x4_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbslq_u32(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return vec_sel(c, b, a);
|
|
#else
|
|
simde_uint32x4_private
|
|
r_,
|
|
a_ = simde_uint32x4_to_private(a),
|
|
b_ = simde_uint32x4_to_private(b),
|
|
c_ = simde_uint32x4_to_private(c);
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi32(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_uint32x4_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_u32
|
|
#define vbslq_u32(a, b, c) simde_vbslq_u32((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_uint64x2_t
|
|
simde_vbslq_u64(simde_uint64x2_t a, simde_uint64x2_t b, simde_uint64x2_t c) {
|
|
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
|
|
return vbslq_u64(a, b, c);
|
|
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE) || defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
|
|
return
|
|
simde_vreinterpretq_u64_u32(
|
|
simde_vbslq_u32(
|
|
simde_vreinterpretq_u32_u64(a),
|
|
simde_vreinterpretq_u32_u64(b),
|
|
simde_vreinterpretq_u32_u64(c)
|
|
)
|
|
);
|
|
#else
|
|
simde_uint64x2_private
|
|
r_,
|
|
a_ = simde_uint64x2_to_private(a),
|
|
b_ = simde_uint64x2_to_private(b),
|
|
c_ = simde_uint64x2_to_private(c);
|
|
|
|
#if defined(SIMDE_WASM_SIMD128_NATIVE)
|
|
r_.v128 = wasm_v128_bitselect(b_.v128, c_.v128, a_.v128);
|
|
#elif defined(SIMDE_X86_AVX512VL_NATIVE)
|
|
r_.m128i = _mm_ternarylogic_epi32(a_.m128i, b_.m128i, c_.m128i, 0xca);
|
|
#elif defined(SIMDE_VECTOR_SUBSCRIPT_OPS)
|
|
r_.values = c_.values ^ ((b_.values ^ c_.values) & a_.values);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = (b_.values[i] & a_.values[i]) | (c_.values[i] & ~a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_uint64x2_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES)
|
|
#undef vbslq_u64
|
|
#define vbslq_u64(a, b, c) simde_vbslq_u64((a), (b), (c))
|
|
#endif
|
|
|
|
SIMDE_END_DECLS_
|
|
HEDLEY_DIAGNOSTIC_POP
|
|
|
|
#endif /* !defined(SIMDE_ARM_NEON_BSL_H) */
|