224 lines
6.4 KiB
C
224 lines
6.4 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:
|
|
* 2020-2021 Evan Nemerson <evan@nemerson.com>
|
|
* 2023 Yi-Yen Chung <eric681@andestech.com> (Copyright owned by Andes Technology)
|
|
*/
|
|
|
|
#if !defined(SIMDE_ARM_NEON_RNDN_H)
|
|
#define SIMDE_ARM_NEON_RNDN_H
|
|
|
|
#include "types.h"
|
|
|
|
HEDLEY_DIAGNOSTIC_PUSH
|
|
SIMDE_DISABLE_UNWANTED_DIAGNOSTICS
|
|
SIMDE_BEGIN_DECLS_
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float16_t
|
|
simde_vrndnh_f16(simde_float16_t a) {
|
|
#if \
|
|
defined(SIMDE_ARM_NEON_A32V8_NATIVE) && \
|
|
(!defined(__clang__) || SIMDE_DETECT_CLANG_VERSION_CHECK(7,0,0)) && \
|
|
(!defined(HEDLEY_GCC_VERSION) || (defined(SIMDE_ARM_NEON_A64V8_NATIVE) && HEDLEY_GCC_VERSION_CHECK(8,0,0))) && defined(SIMDE_ARM_NEON_FP16)
|
|
return vrndnh_f16(a);
|
|
#else
|
|
simde_float32_t a_ = simde_float16_to_float32(a);
|
|
return simde_float16_from_float32(simde_math_roundevenf(a_));
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vrndnh_f16
|
|
#define vrndnh_f16(a) simde_vrndnh_f16(a)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float32_t
|
|
simde_vrndns_f32(simde_float32_t a) {
|
|
#if \
|
|
defined(SIMDE_ARM_NEON_A32V8_NATIVE) && \
|
|
(!defined(__clang__) || SIMDE_DETECT_CLANG_VERSION_CHECK(7,0,0)) && \
|
|
(!defined(HEDLEY_GCC_VERSION) || (defined(SIMDE_ARM_NEON_A64V8_NATIVE) && HEDLEY_GCC_VERSION_CHECK(8,0,0)))
|
|
return vrndns_f32(a);
|
|
#else
|
|
return simde_math_roundevenf(a);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vrndns_f32
|
|
#define vrndns_f32(a) simde_vrndns_f32(a)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float16x4_t
|
|
simde_vrndn_f16(simde_float16x4_t a) {
|
|
#if defined(SIMDE_ARM_NEON_A32V8_NATIVE) && defined(SIMDE_ARM_NEON_FP16)
|
|
return vrndn_f16(a);
|
|
#else
|
|
simde_float16x4_private
|
|
r_,
|
|
a_ = simde_float16x4_to_private(a);
|
|
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = simde_vrndnh_f16(a_.values[i]);
|
|
}
|
|
|
|
return simde_float16x4_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vrndn_f16
|
|
#define vrndn_f16(a) simde_vrndn_f16(a)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float32x2_t
|
|
simde_vrndn_f32(simde_float32x2_t a) {
|
|
#if defined(SIMDE_ARM_NEON_A32V8_NATIVE)
|
|
return vrndn_f32(a);
|
|
#else
|
|
simde_float32x2_private
|
|
r_,
|
|
a_ = simde_float32x2_to_private(a);
|
|
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = simde_vrndns_f32(a_.values[i]);
|
|
}
|
|
|
|
return simde_float32x2_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vrndn_f32
|
|
#define vrndn_f32(a) simde_vrndn_f32(a)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float64x1_t
|
|
simde_vrndn_f64(simde_float64x1_t a) {
|
|
#if \
|
|
defined(SIMDE_ARM_NEON_A64V8_NATIVE)
|
|
return vrndn_f64(a);
|
|
#else
|
|
simde_float64x1_private
|
|
r_,
|
|
a_ = simde_float64x1_to_private(a);
|
|
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = simde_math_roundeven(a_.values[i]);
|
|
}
|
|
|
|
return simde_float64x1_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vrndn_f64
|
|
#define vrndn_f64(a) simde_vrndn_f64(a)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float16x8_t
|
|
simde_vrndnq_f16(simde_float16x8_t a) {
|
|
#if defined(SIMDE_ARM_NEON_A32V8_NATIVE) && defined(SIMDE_ARM_NEON_FP16)
|
|
return vrndnq_f16(a);
|
|
#else
|
|
simde_float16x8_private
|
|
r_,
|
|
a_ = simde_float16x8_to_private(a);
|
|
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = simde_vrndnh_f16(a_.values[i]);
|
|
}
|
|
|
|
return simde_float16x8_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vrndnq_f16
|
|
#define vrndnq_f16(a) simde_vrndnq_f16(a)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float32x4_t
|
|
simde_vrndnq_f32(simde_float32x4_t a) {
|
|
#if defined(SIMDE_ARM_NEON_A32V8_NATIVE)
|
|
return vrndnq_f32(a);
|
|
#else
|
|
simde_float32x4_private
|
|
r_,
|
|
a_ = simde_float32x4_to_private(a);
|
|
|
|
#if defined(SIMDE_X86_SSE4_1_NATIVE)
|
|
r_.m128 = _mm_round_ps(a_.m128, _MM_FROUND_TO_NEAREST_INT);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = simde_vrndns_f32(a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_float32x4_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vrndnq_f32
|
|
#define vrndnq_f32(a) simde_vrndnq_f32(a)
|
|
#endif
|
|
|
|
SIMDE_FUNCTION_ATTRIBUTES
|
|
simde_float64x2_t
|
|
simde_vrndnq_f64(simde_float64x2_t a) {
|
|
#if \
|
|
defined(SIMDE_ARM_NEON_A64V8_NATIVE)
|
|
return vrndnq_f64(a);
|
|
#else
|
|
simde_float64x2_private
|
|
r_,
|
|
a_ = simde_float64x2_to_private(a);
|
|
|
|
#if defined(SIMDE_X86_SSE4_1_NATIVE)
|
|
r_.m128d = _mm_round_pd(a_.m128d, _MM_FROUND_TO_NEAREST_INT);
|
|
#else
|
|
SIMDE_VECTORIZE
|
|
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
|
|
r_.values[i] = simde_math_roundeven(a_.values[i]);
|
|
}
|
|
#endif
|
|
|
|
return simde_float64x2_from_private(r_);
|
|
#endif
|
|
}
|
|
#if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES)
|
|
#undef vrndnq_f64
|
|
#define vrndnq_f64(a) simde_vrndnq_f64(a)
|
|
#endif
|
|
|
|
SIMDE_END_DECLS_
|
|
HEDLEY_DIAGNOSTIC_POP
|
|
|
|
#endif /* !defined(SIMDE_ARM_NEON_RNDN_H) */
|