AnimTestbed/3rdparty/ozz-animation/include/ozz/base/maths/internal/simd_math_config.h

160 lines
5.4 KiB
C
Raw Normal View History

2021-11-11 21:22:24 +01:00
//----------------------------------------------------------------------------//
// //
// ozz-animation is hosted at http://github.com/guillaumeblanc/ozz-animation //
// and distributed under the MIT License (MIT). //
// //
// Copyright (c) Guillaume Blanc //
// //
// 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. //
// //
//----------------------------------------------------------------------------//
#ifndef OZZ_OZZ_BASE_MATHS_INTERNAL_SIMD_MATH_CONFIG_H_
#define OZZ_OZZ_BASE_MATHS_INTERNAL_SIMD_MATH_CONFIG_H_
#include "ozz/base/platform.h"
// Avoid SIMD instruction detection if reference (aka scalar) implementation is
// forced.
#if !defined(OZZ_BUILD_SIMD_REF)
// Try to match a SSE2+ version.
#if defined(__AVX2__) || defined(OZZ_SIMD_AVX2)
#include <immintrin.h>
#define OZZ_SIMD_AVX2
#define OZZ_SIMD_AVX // avx is available if avx2 is.
#endif
#if defined(__FMA__) || defined(OZZ_SIMD_FMA)
#include <immintrin.h>
#define OZZ_SIMD_FMA
#endif
#if defined(__AVX__) || defined(OZZ_SIMD_AVX)
#include <immintrin.h>
#define OZZ_SIMD_AVX
#define OZZ_SIMD_SSE4_2 // SSE4.2 is available if avx is.
#endif
#if defined(__SSE4_2__) || defined(OZZ_SIMD_SSE4_2)
#include <nmmintrin.h>
#define OZZ_SIMD_SSE4_2
#define OZZ_SIMD_SSE4_1 // SSE4.1 is available if SSE4.2 is.
#endif
#if defined(__SSE4_1__) || defined(OZZ_SIMD_SSE4_1)
#include <smmintrin.h>
#define OZZ_SIMD_SSE4_1
#define OZZ_SIMD_SSSE3 // SSSE3 is available if SSE4.1 is.
#endif
#if defined(__SSSE3__) || defined(OZZ_SIMD_SSSE3)
#include <tmmintrin.h>
#define OZZ_SIMD_SSSE3
#define OZZ_SIMD_SSE3 // SSE3 is available if SSSE3 is.
#endif
#if defined(__SSE3__) || defined(OZZ_SIMD_SSE3)
#include <pmmintrin.h>
#define OZZ_SIMD_SSE3
#define OZZ_SIMD_SSE2 // SSE2 is available if SSE3 is.
#endif
// x64/amd64 have SSE2 instructions
// _M_IX86_FP is 2 if /arch:SSE2, /arch:AVX or /arch:AVX2 was used.
#if defined(__SSE2__) || defined(_M_AMD64) || defined(_M_X64) || \
(_M_IX86_FP >= 2) || defined(OZZ_SIMD_SSE2)
#include <emmintrin.h>
#define OZZ_SIMD_SSE2
#define OZZ_SIMD_SSEx // OZZ_SIMD_SSEx is the generic flag for SSE support
#endif
// End of SIMD instruction detection
#endif // !OZZ_BUILD_SIMD_REF
// SEE* intrinsics available
#if defined(OZZ_SIMD_SSEx)
namespace ozz {
namespace math {
// Vector of four floating point values.
typedef __m128 SimdFloat4;
// Argument type for Float4.
typedef const __m128 _SimdFloat4;
// Vector of four integer values.
typedef __m128i SimdInt4;
// Argument type for Int4.
typedef const __m128i _SimdInt4;
} // namespace math
} // namespace ozz
#else // No builtin simd available
// No simd instruction set detected, switch back to reference implementation.
// OZZ_SIMD_REF is the generic flag for SIMD reference implementation.
#define OZZ_SIMD_REF
// Declares reference simd float and integer vectors outside of ozz::math, in
// order to match non-reference implementation details.
// Vector of four floating point values.
struct SimdFloat4Def {
alignas(16) float x;
float y;
float z;
float w;
};
// Vector of four integer values.
struct SimdInt4Def {
alignas(16) int x;
int y;
int z;
int w;
};
namespace ozz {
namespace math {
// Vector of four floating point values.
typedef SimdFloat4Def SimdFloat4;
// Argument type for SimdFloat4
typedef const SimdFloat4& _SimdFloat4;
// Vector of four integer values.
typedef SimdInt4Def SimdInt4;
// Argument type for SimdInt4.
typedef const SimdInt4& _SimdInt4;
} // namespace math
} // namespace ozz
#endif // OZZ_SIMD_x
// Native SIMD operator already exist on some compilers, so they have to be disable from ozz implementation
#if !defined(OZZ_SIMD_REF) && (defined(__GNUC__) || defined(__llvm__))
#define OZZ_DISABLE_SSE_NATIVE_OPERATORS
#endif
#endif // OZZ_OZZ_BASE_MATHS_INTERNAL_SIMD_MATH_CONFIG_H_