highbd_convolve_sve2.h (1705B)
1 /* 2 * Copyright (c) 2023, Alliance for Open Media. All rights reserved. 3 * 4 * This source code is subject to the terms of the BSD 2 Clause License and 5 * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License 6 * was not distributed with this source code in the LICENSE file, you can 7 * obtain it at www.aomedia.org/license/software. If the Alliance for Open 8 * Media Patent License 1.0 was not distributed with this source code in the 9 * PATENTS file, you can obtain it at www.aomedia.org/license/patent. 10 */ 11 12 #ifndef AOM_AV1_COMMON_ARM_HIGHBD_CONVOLVE_SVE2_H_ 13 #define AOM_AV1_COMMON_ARM_HIGHBD_CONVOLVE_SVE2_H_ 14 15 #include <arm_neon.h> 16 17 #include "aom_dsp/arm/aom_neon_sve2_bridge.h" 18 19 // clang-format off 20 DECLARE_ALIGNED(16, static const uint16_t, kDotProdMergeBlockTbl[24]) = { 21 // Shift left and insert new last column in transposed 4x4 block. 22 1, 2, 3, 0, 5, 6, 7, 4, 23 // Shift left and insert two new columns in transposed 4x4 block. 24 2, 3, 0, 1, 6, 7, 4, 5, 25 // Shift left and insert three new columns in transposed 4x4 block. 26 3, 0, 1, 2, 7, 4, 5, 6, 27 }; 28 // clang-format on 29 30 static inline void aom_tbl2x4_s16(int16x8_t t0[4], int16x8_t t1[4], 31 uint16x8_t tbl, int16x8_t res[4]) { 32 res[0] = aom_tbl2_s16(t0[0], t1[0], tbl); 33 res[1] = aom_tbl2_s16(t0[1], t1[1], tbl); 34 res[2] = aom_tbl2_s16(t0[2], t1[2], tbl); 35 res[3] = aom_tbl2_s16(t0[3], t1[3], tbl); 36 } 37 38 static inline void aom_tbl2x2_s16(int16x8_t t0[2], int16x8_t t1[2], 39 uint16x8_t tbl, int16x8_t res[2]) { 40 res[0] = aom_tbl2_s16(t0[0], t1[0], tbl); 41 res[1] = aom_tbl2_s16(t0[1], t1[1], tbl); 42 } 43 44 #endif // AOM_AV1_COMMON_ARM_HIGHBD_CONVOLVE_SVE2_H_