/* 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 folhighing 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 * 2020 Sean Maher (Copyright owned by Google, LLC) * 2023 Yi-Yen Chung (Copyright owned by Andes Technology) */ #if !defined(SIMDE_ARM_NEON_COMBINE_H) #define SIMDE_ARM_NEON_COMBINE_H #include "types.h" HEDLEY_DIAGNOSTIC_PUSH SIMDE_DISABLE_UNWANTED_DIAGNOSTICS SIMDE_BEGIN_DECLS_ SIMDE_FUNCTION_ATTRIBUTES simde_float16x8_t simde_vcombine_f16(simde_float16x4_t low, simde_float16x4_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) && defined(SIMDE_ARM_NEON_FP16) return vcombine_f16(low, high); #else simde_float16x8_private r_; simde_float16x4_private low_ = simde_float16x4_to_private(low), high_ = simde_float16x4_to_private(high); size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } return simde_float16x8_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_f16 #define vcombine_f16(low, high) simde_vcombine_f16((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_float32x4_t simde_vcombine_f32(simde_float32x2_t low, simde_float32x2_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_f32(low, high); #else simde_float32x4_private r_; simde_float32x2_private low_ = simde_float32x2_to_private(low), high_ = simde_float32x2_to_private(high); /* Note: __builtin_shufflevector can have a the output contain * twice the number of elements, __builtin_shuffle cannot. * Using SIMDE_SHUFFLE_VECTOR_ here would not work. */ #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1, 2, 3); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_float32x4_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_f32 #define vcombine_f32(low, high) simde_vcombine_f32((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_float64x2_t simde_vcombine_f64(simde_float64x1_t low, simde_float64x1_t high) { #if defined(SIMDE_ARM_NEON_A64V8_NATIVE) return vcombine_f64(low, high); #else simde_float64x2_private r_; simde_float64x1_private low_ = simde_float64x1_to_private(low), high_ = simde_float64x1_to_private(high); #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_float64x2_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A64V8_ENABLE_NATIVE_ALIASES) #undef vcombine_f64 #define vcombine_f64(low, high) simde_vcombine_f64((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_int8x16_t simde_vcombine_s8(simde_int8x8_t low, simde_int8x8_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_s8(low, high); #else simde_int8x16_private r_; simde_int8x8_private low_ = simde_int8x8_to_private(low), high_ = simde_int8x8_to_private(high); #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_int8x16_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_s8 #define vcombine_s8(low, high) simde_vcombine_s8((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_int16x8_t simde_vcombine_s16(simde_int16x4_t low, simde_int16x4_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_s16(low, high); #else simde_int16x8_private r_; simde_int16x4_private low_ = simde_int16x4_to_private(low), high_ = simde_int16x4_to_private(high); #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1, 2, 3, 4, 5, 6, 7); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_int16x8_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_s16 #define vcombine_s16(low, high) simde_vcombine_s16((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_int32x4_t simde_vcombine_s32(simde_int32x2_t low, simde_int32x2_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_s32(low, high); #else simde_int32x4_private r_; simde_int32x2_private low_ = simde_int32x2_to_private(low), high_ = simde_int32x2_to_private(high); #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1, 2, 3); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_int32x4_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_s32 #define vcombine_s32(low, high) simde_vcombine_s32((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_int64x2_t simde_vcombine_s64(simde_int64x1_t low, simde_int64x1_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_s64(low, high); #else simde_int64x2_private r_; simde_int64x1_private low_ = simde_int64x1_to_private(low), high_ = simde_int64x1_to_private(high); #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_int64x2_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_s64 #define vcombine_s64(low, high) simde_vcombine_s64((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_uint8x16_t simde_vcombine_u8(simde_uint8x8_t low, simde_uint8x8_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_u8(low, high); #else simde_uint8x16_private r_; simde_uint8x8_private low_ = simde_uint8x8_to_private(low), high_ = simde_uint8x8_to_private(high); #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_uint8x16_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_u8 #define vcombine_u8(low, high) simde_vcombine_u8((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_uint16x8_t simde_vcombine_u16(simde_uint16x4_t low, simde_uint16x4_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_u16(low, high); #else simde_uint16x8_private r_; simde_uint16x4_private low_ = simde_uint16x4_to_private(low), high_ = simde_uint16x4_to_private(high); #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1, 2, 3, 4, 5, 6, 7); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_uint16x8_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_u16 #define vcombine_u16(low, high) simde_vcombine_u16((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_uint32x4_t simde_vcombine_u32(simde_uint32x2_t low, simde_uint32x2_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_u32(low, high); #else simde_uint32x4_private r_; simde_uint32x2_private low_ = simde_uint32x2_to_private(low), high_ = simde_uint32x2_to_private(high); #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1, 2, 3); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_uint32x4_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_u32 #define vcombine_u32(low, high) simde_vcombine_u32((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_uint64x2_t simde_vcombine_u64(simde_uint64x1_t low, simde_uint64x1_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_u64(low, high); #else simde_uint64x2_private r_; simde_uint64x1_private low_ = simde_uint64x1_to_private(low), high_ = simde_uint64x1_to_private(high); #if defined(SIMDE_VECTOR_SUBSCRIPT) && HEDLEY_HAS_BUILTIN(__builtin_shufflevector) r_.values = __builtin_shufflevector(low_.values, high_.values, 0, 1); #else size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } #endif return simde_uint64x2_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_u64 #define vcombine_u64(low, high) simde_vcombine_u64((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_poly8x16_t simde_vcombine_p8(simde_poly8x8_t low, simde_poly8x8_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_p8(low, high); #else simde_poly8x16_private r_; simde_poly8x8_private low_ = simde_poly8x8_to_private(low), high_ = simde_poly8x8_to_private(high); size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } return simde_poly8x16_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_p8 #define vcombine_p8(low, high) simde_vcombine_p8((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_poly16x8_t simde_vcombine_p16(simde_poly16x4_t low, simde_poly16x4_t high) { #if defined(SIMDE_ARM_NEON_A32V7_NATIVE) return vcombine_p16(low, high); #else simde_poly16x8_private r_; simde_poly16x4_private low_ = simde_poly16x4_to_private(low), high_ = simde_poly16x4_to_private(high); size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } return simde_poly16x8_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V7_ENABLE_NATIVE_ALIASES) #undef vcombine_p16 #define vcombine_p16(low, high) simde_vcombine_p16((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_poly64x2_t simde_vcombine_p64(simde_poly64x1_t low, simde_poly64x1_t high) { #if defined(SIMDE_ARM_NEON_A32V8_NATIVE) return vcombine_p64(low, high); #else simde_poly64x2_private r_; simde_poly64x1_private low_ = simde_poly64x1_to_private(low), high_ = simde_poly64x1_to_private(high); size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } return simde_poly64x2_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES) #undef vcombine_p64 #define vcombine_p64(low, high) simde_vcombine_p64((low), (high)) #endif SIMDE_FUNCTION_ATTRIBUTES simde_bfloat16x8_t simde_vcombine_bf16(simde_bfloat16x4_t low, simde_bfloat16x4_t high) { #if defined(SIMDE_ARM_NEON_A32V8_NATIVE) && defined(SIMDE_ARM_NEON_BF16) return vcombine_bf16(low, high); #else simde_bfloat16x8_private r_; simde_bfloat16x4_private low_ = simde_bfloat16x4_to_private(low), high_ = simde_bfloat16x4_to_private(high); size_t halfway = (sizeof(r_.values) / sizeof(r_.values[0])) / 2; SIMDE_VECTORIZE for (size_t i = 0 ; i < halfway ; i++) { r_.values[i] = low_.values[i]; r_.values[i + halfway] = high_.values[i]; } return simde_bfloat16x8_from_private(r_); #endif } #if defined(SIMDE_ARM_NEON_A32V8_ENABLE_NATIVE_ALIASES) #undef vcombine_bf16 #define vcombine_bf16(low, high) simde_vcombine_bf16((low), (high)) #endif SIMDE_END_DECLS_ HEDLEY_DIAGNOSTIC_POP #endif /* !defined(SIMDE_ARM_NEON_COMBINE_H) */