/* 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 * 2020 Himanshi Mathur * 2020 Hidayat Khan * 2021 Andrew Rodriguez */ #if !defined(SIMDE_X86_AVX512_CVT_H) #define SIMDE_X86_AVX512_CVT_H #include "types.h" #include "mov.h" #include "../f16c.h" HEDLEY_DIAGNOSTIC_PUSH SIMDE_DISABLE_UNWANTED_DIAGNOSTICS SIMDE_BEGIN_DECLS_ SIMDE_FUNCTION_ATTRIBUTES simde__m128d simde_mm_cvtepi64_pd (simde__m128i a) { #if defined(SIMDE_X86_AVX512VL_NATIVE) && defined(SIMDE_X86_AVX512DQ_NATIVE) return _mm_cvtepi64_pd(a); #else simde__m128d_private r_; simde__m128i_private a_ = simde__m128i_to_private(a); #if defined(SIMDE_X86_SSE2_NATIVE) /* https://stackoverflow.com/questions/41144668/how-to-efficiently-perform-double-int64-conversions-with-sse-avx */ __m128i xH = _mm_srai_epi32(a_.n, 16); #if defined(SIMDE_X86_SSE4_2_NATIVE) xH = _mm_blend_epi16(xH, _mm_setzero_si128(), 0x33); #else xH = _mm_and_si128(xH, _mm_set_epi16(~INT16_C(0), ~INT16_C(0), INT16_C(0), INT16_C(0), ~INT16_C(0), ~INT16_C(0), INT16_C(0), INT16_C(0))); #endif xH = _mm_add_epi64(xH, _mm_castpd_si128(_mm_set1_pd(442721857769029238784.0))); const __m128i e = _mm_castpd_si128(_mm_set1_pd(0x0010000000000000)); #if defined(SIMDE_X86_SSE4_2_NATIVE) __m128i xL = _mm_blend_epi16(a_.n, e, 0x88); #else __m128i m = _mm_set_epi16(INT16_C(0), ~INT16_C(0), ~INT16_C(0), ~INT16_C(0), INT16_C(0), ~INT16_C(0), ~INT16_C(0), ~INT16_C(0)); __m128i xL = _mm_or_si128(_mm_and_si128(m, a_.n), _mm_andnot_si128(m, e)); #endif __m128d f = _mm_sub_pd(_mm_castsi128_pd(xH), _mm_set1_pd(442726361368656609280.0)); return _mm_add_pd(f, _mm_castsi128_pd(xL)); #elif defined(SIMDE_CONVERT_VECTOR_) SIMDE_CONVERT_VECTOR_(r_.f64, a_.i64); #else SIMDE_VECTORIZE for (size_t i = 0 ; i < (sizeof(r_.f64) / sizeof(r_.f64[0])) ; i++) { r_.f64[i] = HEDLEY_STATIC_CAST(simde_float64, a_.i64[i]); } #endif return simde__m128d_from_private(r_); #endif } #if defined(SIMDE_X86_AVX512BW_ENABLE_NATIVE_ALIASES) #undef _mm_cvtepi64_pd #define _mm_cvtepi64_pd(a) simde_mm_cvtepi64_pd(a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m128d simde_mm_mask_cvtepi64_pd(simde__m128d src, simde__mmask8 k, simde__m128i a) { #if defined(SIMDE_X86_AVX512VL_NATIVE) && defined(SIMDE_X86_AVX512DQ_NATIVE) return _mm_mask_cvtepi64_pd(src, k, a); #else return simde_mm_mask_mov_pd(src, k, simde_mm_cvtepi64_pd(a)); #endif } #if defined(SIMDE_X86_AVX512VL_ENABLE_NATIVE_ALIASES) #undef _mm_mask_cvtepi64_pd #define _mm_mask_cvtepi64_pd(src, k, a) simde_mm_mask_cvtepi64_pd(src, k, a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m128d simde_mm_maskz_cvtepi64_pd(simde__mmask8 k, simde__m128i a) { #if defined(SIMDE_X86_AVX512VL_NATIVE) && defined(SIMDE_X86_AVX512DQ_NATIVE) return _mm_maskz_cvtepi64_pd(k, a); #else return simde_mm_maskz_mov_pd(k, simde_mm_cvtepi64_pd(a)); #endif } #if defined(SIMDE_X86_AVX512VL_ENABLE_NATIVE_ALIASES) || defined(SIMDE_X86_AVX512DQ_ENABLE_NATIVE_ALIASES) #undef _mm_maskz_cvtepi64_pd #define _mm_maskz_cvtepi64_pd(k, a) simde_mm_maskz_cvtepi64_pd(k, a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m256i simde_mm512_cvtepi16_epi8 (simde__m512i a) { #if defined(SIMDE_X86_AVX512BW_NATIVE) return _mm512_cvtepi16_epi8(a); #else simde__m256i_private r_; simde__m512i_private a_ = simde__m512i_to_private(a); #if defined(SIMDE_CONVERT_VECTOR_) SIMDE_CONVERT_VECTOR_(r_.i8, a_.i16); #else SIMDE_VECTORIZE for (size_t i = 0 ; i < (sizeof(r_.i8) / sizeof(r_.i8[0])) ; i++) { r_.i8[i] = HEDLEY_STATIC_CAST(int8_t, a_.i16[i]); } #endif return simde__m256i_from_private(r_); #endif } #if defined(SIMDE_X86_AVX512BW_ENABLE_NATIVE_ALIASES) || defined(SIMDE_X86_AVX512DQ_ENABLE_NATIVE_ALIASES) #undef _mm512_cvtepi16_epi8 #define _mm512_cvtepi16_epi8(a) simde_mm512_cvtepi16_epi8(a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m256i simde_mm512_mask_cvtepi16_epi8 (simde__m256i src, simde__mmask32 k, simde__m512i a) { #if defined(SIMDE_X86_AVX512BW_NATIVE) return _mm512_mask_cvtepi16_epi8(src, k, a); #else return simde_mm256_mask_mov_epi8(src, k, simde_mm512_cvtepi16_epi8(a)); #endif } #if defined(SIMDE_X86_AVX512BW_ENABLE_NATIVE_ALIASES) #undef _mm512_mask_cvtepi16_epi8 #define _mm512_mask_cvtepi16_epi8(src, k, a) simde_mm512_mask_cvtepi16_epi8(src, k, a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m256i simde_mm512_maskz_cvtepi16_epi8 (simde__mmask32 k, simde__m512i a) { #if defined(SIMDE_X86_AVX512BW_NATIVE) return _mm512_maskz_cvtepi16_epi8(k, a); #else return simde_mm256_maskz_mov_epi8(k, simde_mm512_cvtepi16_epi8(a)); #endif } #if defined(SIMDE_X86_AVX512BW_ENABLE_NATIVE_ALIASES) #undef _mm512_maskz_cvtepi16_epi8 #define _mm512_maskz_cvtepi16_epi8(k, a) simde_mm512_maskz_cvtepi16_epi8(k, a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m512i simde_mm512_cvtepi8_epi16 (simde__m256i a) { #if defined(SIMDE_X86_AVX512BW_NATIVE) return _mm512_cvtepi8_epi16(a); #else simde__m512i_private r_; simde__m256i_private a_ = simde__m256i_to_private(a); #if defined(SIMDE_CONVERT_VECTOR_) SIMDE_CONVERT_VECTOR_(r_.i16, a_.i8); #else SIMDE_VECTORIZE for (size_t i = 0 ; i < (sizeof(r_.i16) / sizeof(r_.i16[0])) ; i++) { r_.i16[i] = a_.i8[i]; } #endif return simde__m512i_from_private(r_); #endif } #if defined(SIMDE_X86_AVX512BW_ENABLE_NATIVE_ALIASES) #undef _mm512_cvtepi8_epi16 #define _mm512_cvtepi8_epi16(a) simde_mm512_cvtepi8_epi16(a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m512 simde_mm512_cvtepi32_ps (simde__m512i a) { #if defined(SIMDE_X86_AVX512F_NATIVE) return _mm512_cvtepi32_ps(a); #else simde__m512_private r_; simde__m512i_private a_ = simde__m512i_to_private(a); #if defined(SIMDE_CONVERT_VECTOR_) SIMDE_CONVERT_VECTOR_(r_.f32, a_.i32); #else SIMDE_VECTORIZE for (size_t i = 0 ; i < (sizeof(r_.i32) / sizeof(r_.i32[0])) ; i++) { r_.f32[i] = HEDLEY_STATIC_CAST(simde_float32, a_.i32[i]); } #endif return simde__m512_from_private(r_); #endif } #if defined(SIMDE_X86_AVX512F_ENABLE_NATIVE_ALIASES) #undef _mm512_cvtepi32_ps #define _mm512_cvtepi32_ps(a) simde_mm512_cvtepi32_ps(a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m256i simde_mm512_cvtepi64_epi32 (simde__m512i a) { #if defined(SIMDE_X86_AVX512F_NATIVE) return _mm512_cvtepi64_epi32(a); #else simde__m256i_private r_; simde__m512i_private a_ = simde__m512i_to_private(a); #if defined(SIMDE_CONVERT_VECTOR_) SIMDE_CONVERT_VECTOR_(r_.i32, a_.i64); #else SIMDE_VECTORIZE for (size_t i = 0 ; i < (sizeof(r_.i32) / sizeof(r_.i32[0])) ; i++) { r_.i32[i] = HEDLEY_STATIC_CAST(int32_t, a_.i64[i]); } #endif return simde__m256i_from_private(r_); #endif } #if defined(SIMDE_X86_AVX512F_ENABLE_NATIVE_ALIASES) #undef _mm512_cvtepi64_epi32 #define _mm512_cvtepi64_epi32(a) simde_mm512_cvtepi64_epi32(a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m512 simde_mm512_cvtepu32_ps (simde__m512i a) { #if defined(SIMDE_X86_AVX512F_NATIVE) return _mm512_cvtepu32_ps(a); #else simde__m512_private r_; simde__m512i_private a_ = simde__m512i_to_private(a); #if defined(SIMDE_X86_SSE2_NATIVE) for (size_t i = 0 ; i < (sizeof(r_.m128) / sizeof(r_.m128[0])) ; i++) { /* https://stackoverflow.com/a/34067907/501126 */ const __m128 tmp = _mm_cvtepi32_ps(_mm_srli_epi32(a_.m128i[i], 1)); r_.m128[i] = _mm_add_ps( _mm_add_ps(tmp, tmp), _mm_cvtepi32_ps(_mm_and_si128(a_.m128i[i], _mm_set1_epi32(1))) ); } #elif defined(SIMDE_CONVERT_VECTOR_) SIMDE_CONVERT_VECTOR_(r_.f32, a_.u32); #else SIMDE_VECTORIZE for (size_t i = 0 ; i < (sizeof(r_.u32) / sizeof(r_.u32[0])) ; i++) { r_.f32[i] = HEDLEY_STATIC_CAST(float, a_.u32[i]); } #endif return simde__m512_from_private(r_); #endif } #if defined(SIMDE_X86_AVX512F_ENABLE_NATIVE_ALIASES) #undef _mm512_cvtepu32_ps #define _mm512_cvtepu32_ps(a) simde_mm512_cvtepu32_ps(a) #endif SIMDE_FUNCTION_ATTRIBUTES simde__m512 simde_mm512_cvtph_ps(simde__m256i a) { #if defined(SIMDE_X86_AVX512F_NATIVE) return _mm512_cvtph_ps(a); #endif simde__m256i_private a_ = simde__m256i_to_private(a); simde__m512_private r_; SIMDE_VECTORIZE for (size_t i = 0 ; i < (sizeof(r_.f32) / sizeof(r_.f32[0])) ; i++) { r_.f32[i] = simde_float16_to_float32(simde_uint16_as_float16(a_.u16[i])); } return simde__m512_from_private(r_); } #if defined(SIMDE_X86_AVX512F_ENABLE_NATIVE_ALIASES) #undef _mm512_cvtph_ps #define _mm512_cvtph_ps(a) simde_mm512_cvtph_ps(a) #endif SIMDE_END_DECLS_ HEDLEY_DIAGNOSTIC_POP #endif /* !defined(SIMDE_X86_AVX512_CVT_H) */