dist_wtd_avg_neon.h (2609B)
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_AOM_DSP_ARM_DIST_WTD_AVG_NEON_H_ 13 #define AOM_AOM_DSP_ARM_DIST_WTD_AVG_NEON_H_ 14 15 #include <arm_neon.h> 16 17 #include "aom_dsp/aom_dsp_common.h" 18 #include "av1/common/enums.h" 19 20 static inline uint8x8_t dist_wtd_avg_u8x8(uint8x8_t a, uint8x8_t b, 21 uint8x8_t wta, uint8x8_t wtb) { 22 uint16x8_t wtd_sum = vmull_u8(a, wta); 23 24 wtd_sum = vmlal_u8(wtd_sum, b, wtb); 25 26 return vrshrn_n_u16(wtd_sum, DIST_PRECISION_BITS); 27 } 28 29 static inline uint16x4_t dist_wtd_avg_u16x4(uint16x4_t a, uint16x4_t b, 30 uint16x4_t wta, uint16x4_t wtb) { 31 uint32x4_t wtd_sum = vmull_u16(a, wta); 32 33 wtd_sum = vmlal_u16(wtd_sum, b, wtb); 34 35 return vrshrn_n_u32(wtd_sum, DIST_PRECISION_BITS); 36 } 37 38 static inline uint8x16_t dist_wtd_avg_u8x16(uint8x16_t a, uint8x16_t b, 39 uint8x16_t wta, uint8x16_t wtb) { 40 uint16x8_t wtd_sum_lo = vmull_u8(vget_low_u8(a), vget_low_u8(wta)); 41 uint16x8_t wtd_sum_hi = vmull_u8(vget_high_u8(a), vget_high_u8(wta)); 42 43 wtd_sum_lo = vmlal_u8(wtd_sum_lo, vget_low_u8(b), vget_low_u8(wtb)); 44 wtd_sum_hi = vmlal_u8(wtd_sum_hi, vget_high_u8(b), vget_high_u8(wtb)); 45 46 uint8x8_t wtd_avg_lo = vrshrn_n_u16(wtd_sum_lo, DIST_PRECISION_BITS); 47 uint8x8_t wtd_avg_hi = vrshrn_n_u16(wtd_sum_hi, DIST_PRECISION_BITS); 48 49 return vcombine_u8(wtd_avg_lo, wtd_avg_hi); 50 } 51 52 static inline uint16x8_t dist_wtd_avg_u16x8(uint16x8_t a, uint16x8_t b, 53 uint16x8_t wta, uint16x8_t wtb) { 54 uint32x4_t wtd_sum_lo = vmull_u16(vget_low_u16(a), vget_low_u16(wta)); 55 uint32x4_t wtd_sum_hi = vmull_u16(vget_high_u16(a), vget_high_u16(wta)); 56 57 wtd_sum_lo = vmlal_u16(wtd_sum_lo, vget_low_u16(b), vget_low_u16(wtb)); 58 wtd_sum_hi = vmlal_u16(wtd_sum_hi, vget_high_u16(b), vget_high_u16(wtb)); 59 60 uint16x4_t wtd_avg_lo = vrshrn_n_u32(wtd_sum_lo, DIST_PRECISION_BITS); 61 uint16x4_t wtd_avg_hi = vrshrn_n_u32(wtd_sum_hi, DIST_PRECISION_BITS); 62 63 return vcombine_u16(wtd_avg_lo, wtd_avg_hi); 64 } 65 66 #endif // AOM_AOM_DSP_ARM_DIST_WTD_AVG_NEON_H_