/* Copyright 2016-2017 INRIA and Microsoft Corporation * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #ifndef __Vec_H #define __Vec_H #ifdef __MSVC__ #define forceinline __forceinline inline #elif (defined(__GNUC__) || defined(__clang__)) #define forceinline __attribute__((always_inline)) inline #else #define forceinline inline #endif #if defined(__SSSE3__) || defined(__AVX2__) || defined(__AVX__) #include <emmintrin.h> #include <tmmintrin.h> #define VEC128 #define vec_size 4 typedef __m128i vec; static forceinline vec vec_rotate_left_8(vec v) { __m128i x = _mm_set_epi8(14, 13, 12, 15, 10, 9, 8, 11, 6, 5, 4, 7, 2, 1, 0, 3); return _mm_shuffle_epi8(v, x); } static forceinline vec vec_rotate_left_16(vec v) { __m128i x = _mm_set_epi8(13, 12, 15, 14, 9, 8, 11, 10, 5, 4, 7, 6, 1, 0, 3, 2); return _mm_shuffle_epi8(v, x); } static forceinline vec vec_rotate_left(vec v, unsigned int n) { if (n == 8) return vec_rotate_left_8(v); if (n == 16) return vec_rotate_left_16(v); return _mm_xor_si128(_mm_slli_epi32(v, n), _mm_srli_epi32(v, 32 - n)); } static forceinline vec vec_rotate_right(vec v, unsigned int n) { return (vec_rotate_left(v, 32 - n)); } #define vec_shuffle_right(x, n) \ _mm_shuffle_epi32(x, _MM_SHUFFLE((3 + (n)) % 4, (2 + (n)) % 4, (1 + (n)) % 4, (n) % 4)) #define vec_shuffle_left(x, n) vec_shuffle_right((x), 4 - (n)) static forceinline vec vec_load_32x4(uint32_t x1, uint32_t x2, uint32_t x3, uint32_t x4) { return _mm_set_epi32(x4, x3, x2, x1); } static forceinline vec vec_load_32x8(uint32_t x1, uint32_t x2, uint32_t x3, uint32_t x4, uint32_t x5, uint32_t x6, uint32_t x7, uint32_t x8) { return _mm_set_epi32(x4, x3, x2, x1); } static forceinline vec vec_load_le(const unsigned char* in) { return _mm_loadu_si128((__m128i*)(in)); } static forceinline vec vec_load128_le(const unsigned char* in) { return vec_load_le(in); } static forceinline void vec_store_le(unsigned char* out, vec v) { _mm_storeu_si128((__m128i*)(out), v); } static forceinline vec vec_add(vec v1, vec v2) { return _mm_add_epi32(v1, v2); } static forceinline vec vec_add_u32(vec v1, uint32_t x) { vec v2 = vec_load_32x4(x, 0, 0, 0); return _mm_add_epi32(v1, v2); } static forceinline vec vec_increment(vec v1) { vec one = vec_load_32x4(1, 0, 0, 0); return _mm_add_epi32(v1, one); } static forceinline vec vec_xor(vec v1, vec v2) { return _mm_xor_si128(v1, v2); } #define vec_zero() _mm_set_epi32(0, 0, 0, 0) #elif defined(__ARM_NEON__) || defined(__ARM_NEON) #include <arm_neon.h> typedef uint32x4_t vec; static forceinline vec vec_xor(vec v1, vec v2) { return veorq_u32(v1, v2); } #define vec_rotate_left(x, n) \ vsriq_n_u32(vshlq_n_u32((x), (n)), (x), 32 - (n)) #define vec_rotate_right(a, b) \ vec_rotate_left((b), 32 - (b)) #define vec_shuffle_right(x, n) \ vextq_u32((x), (x), (n)) #define vec_shuffle_left(a, b) \ vec_shuffle_right((a), 4 - (b)) static forceinline vec vec_load_32x4(uint32_t x1, uint32_t x2, uint32_t x3, uint32_t x4) { uint32_t a[4] = { x1, x2, x3, x4 }; return vld1q_u32(a); } static forceinline vec vec_load_32(uint32_t x1) { uint32_t a[4] = { x1, x1, x1, x1 }; return vld1q_u32(a); } static forceinline vec vec_load_32x8(uint32_t x1, uint32_t x2, uint32_t x3, uint32_t x4, uint32_t x5, uint32_t x6, uint32_t x7, uint32_t x8) { return vec_load_32x4(x1, x2, x3, x4); } static forceinline vec vec_load_le(const unsigned char* in) { return vld1q_u32((uint32_t*)in); } static forceinline vec vec_load128_le(const unsigned char* in) { return vec_load_le(in); } static forceinline void vec_store_le(unsigned char* out, vec v) { vst1q_u32((uint32_t*)out, v); } static forceinline vec vec_add(vec v1, vec v2) { return vaddq_u32(v1, v2); } static forceinline vec vec_add_u32(vec v1, uint32_t x) { vec v2 = vec_load_32x4(x, 0, 0, 0); return vec_add(v1, v2); } static forceinline vec vec_increment(vec v1) { vec one = vec_load_32x4(1, 0, 0, 0); return vec_add(v1, one); } #define vec_zero() vec_load_32x4(0, 0, 0, 0) #else #define VEC128 #define vec_size 4 typedef struct { uint32_t v[4]; } vec; static forceinline vec vec_xor(vec v1, vec v2) { vec r; r.v[0] = v1.v[0] ^ v2.v[0]; r.v[1] = v1.v[1] ^ v2.v[1]; r.v[2] = v1.v[2] ^ v2.v[2]; r.v[3] = v1.v[3] ^ v2.v[3]; return r; } static forceinline vec vec_rotate_left(vec v, unsigned int n) { vec r; r.v[0] = (v.v[0] << n) ^ (v.v[0] >> (32 - n)); r.v[1] = (v.v[1] << n) ^ (v.v[1] >> (32 - n)); r.v[2] = (v.v[2] << n) ^ (v.v[2] >> (32 - n)); r.v[3] = (v.v[3] << n) ^ (v.v[3] >> (32 - n)); return r; } static forceinline vec vec_rotate_right(vec v, unsigned int n) { return (vec_rotate_left(v, 32 - n)); } static forceinline vec vec_shuffle_right(vec v, unsigned int n) { vec r; r.v[0] = v.v[n % 4]; r.v[1] = v.v[(n + 1) % 4]; r.v[2] = v.v[(n + 2) % 4]; r.v[3] = v.v[(n + 3) % 4]; return r; } static forceinline vec vec_shuffle_left(vec x, unsigned int n) { return vec_shuffle_right(x, 4 - n); } static forceinline vec vec_load_32x4(uint32_t x0, uint32_t x1, uint32_t x2, uint32_t x3) { vec v; v.v[0] = x0; v.v[1] = x1; v.v[2] = x2; v.v[3] = x3; return v; } static forceinline vec vec_load_32(uint32_t x0) { vec v; v.v[0] = x0; v.v[1] = x0; v.v[2] = x0; v.v[3] = x0; return v; } static forceinline vec vec_load_le(const uint8_t* in) { vec r; r.v[0] = load32_le((uint8_t*)in); r.v[1] = load32_le((uint8_t*)in + 4); r.v[2] = load32_le((uint8_t*)in + 8); r.v[3] = load32_le((uint8_t*)in + 12); return r; } static forceinline void vec_store_le(unsigned char* out, vec r) { store32_le(out, r.v[0]); store32_le(out + 4, r.v[1]); store32_le(out + 8, r.v[2]); store32_le(out + 12, r.v[3]); } static forceinline vec vec_load128_le(const unsigned char* in) { return vec_load_le(in); } static forceinline vec vec_add(vec v1, vec v2) { vec r; r.v[0] = v1.v[0] + v2.v[0]; r.v[1] = v1.v[1] + v2.v[1]; r.v[2] = v1.v[2] + v2.v[2]; r.v[3] = v1.v[3] + v2.v[3]; return r; } static forceinline vec vec_add_u32(vec v1, uint32_t x) { vec v2 = vec_load_32x4(x, 0, 0, 0); return vec_add(v1, v2); } static forceinline vec vec_increment(vec v1) { vec one = vec_load_32x4(1, 0, 0, 0); return vec_add(v1, one); } #define vec_zero() vec_load_32x4(0, 0, 0, 0) #endif #endif