blob: 3cb2567b2eafe6c724fc93dd6098dbbeb2602b6b [file] [log] [blame]
/*
* Copyright (c) 2018, Alliance for Open Media. All rights reserved
*
* This source code is subject to the terms of the BSD 2 Clause License and
* the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
* was not distributed with this source code in the LICENSE file, you can
* obtain it at www.aomedia.org/license/software. If the Alliance for Open
* Media Patent License 1.0 was not distributed with this source code in the
* PATENTS file, you can obtain it at www.aomedia.org/license/patent.
*/
#include <arm_neon.h>
#include <assert.h>
#include "config/aom_config.h"
#include "config/av1_rtcd.h"
#include "aom_dsp/txfm_common.h"
#include "aom_dsp/arm/mem_neon.h"
#include "aom_dsp/arm/transpose_neon.h"
#include "aom_ports/mem.h"
#include "av1/common/common.h"
#include "av1/common/arm/convolve_neon.h"
#if !defined(__aarch64__)
static INLINE void compute_avg_4x1(uint16x4_t dd0, uint16x4_t d0,
const uint16_t fwd_offset,
const uint16_t bck_offset,
const int16x4_t round_offset,
const int use_dist_wtd_comp_avg,
uint8x8_t *d0_u8) {
uint16x4_t avg0;
if (use_dist_wtd_comp_avg) {
uint32x4_t blend0;
blend0 = vmull_n_u16(dd0, fwd_offset);
blend0 = vmlal_n_u16(blend0, d0, bck_offset);
avg0 = vshrn_n_u32(blend0, DIST_PRECISION_BITS);
} else {
avg0 = vhadd_u16(dd0, d0);
}
int16x4_t dst0 = vsub_s16(vreinterpret_s16_u16(avg0), round_offset);
int16x8_t dst0q = vcombine_s16(dst0, vdup_n_s16(0));
*d0_u8 = vqrshrun_n_s16(dst0q, FILTER_BITS - ROUND0_BITS);
}
static INLINE void compute_avg_8x1(uint16x8_t dd0, uint16x8_t d0,
const uint16_t fwd_offset,
const uint16_t bck_offset,
const int16x8_t round_offset,
const int use_dist_wtd_comp_avg,
uint8x8_t *d0_u8) {
uint16x8_t avg0;
if (use_dist_wtd_comp_avg) {
uint32x4_t blend0_lo, blend0_hi;
blend0_lo = vmull_n_u16(vget_low_u16(dd0), fwd_offset);
blend0_lo = vmlal_n_u16(blend0_lo, vget_low_u16(d0), bck_offset);
blend0_hi = vmull_n_u16(vget_high_u16(dd0), fwd_offset);
blend0_hi = vmlal_n_u16(blend0_hi, vget_high_u16(d0), bck_offset);
avg0 = vcombine_u16(vshrn_n_u32(blend0_lo, DIST_PRECISION_BITS),
vshrn_n_u32(blend0_hi, DIST_PRECISION_BITS));
} else {
avg0 = vhaddq_u16(dd0, d0);
}
int16x8_t dst0 = vsubq_s16(vreinterpretq_s16_u16(avg0), round_offset);
*d0_u8 = vqrshrun_n_s16(dst0, FILTER_BITS - ROUND0_BITS);
}
#endif // !defined(__arch64__)
static INLINE void compute_avg_4x4(uint16x4_t dd0, uint16x4_t dd1,
uint16x4_t dd2, uint16x4_t dd3,
uint16x4_t d0, uint16x4_t d1, uint16x4_t d2,
uint16x4_t d3, const uint16_t fwd_offset,
const uint16_t bck_offset,
const int16x8_t round_offset,
const int use_dist_wtd_comp_avg,
uint8x8_t *d01_u8, uint8x8_t *d23_u8) {
uint16x4_t avg0, avg1, avg2, avg3;
if (use_dist_wtd_comp_avg) {
uint32x4_t blend0, blend1, blend2, blend3;
blend0 = vmull_n_u16(dd0, fwd_offset);
blend0 = vmlal_n_u16(blend0, d0, bck_offset);
blend1 = vmull_n_u16(dd1, fwd_offset);
blend1 = vmlal_n_u16(blend1, d1, bck_offset);
blend2 = vmull_n_u16(dd2, fwd_offset);
blend2 = vmlal_n_u16(blend2, d2, bck_offset);
blend3 = vmull_n_u16(dd3, fwd_offset);
blend3 = vmlal_n_u16(blend3, d3, bck_offset);
avg0 = vshrn_n_u32(blend0, DIST_PRECISION_BITS);
avg1 = vshrn_n_u32(blend1, DIST_PRECISION_BITS);
avg2 = vshrn_n_u32(blend2, DIST_PRECISION_BITS);
avg3 = vshrn_n_u32(blend3, DIST_PRECISION_BITS);
} else {
avg0 = vhadd_u16(dd0, d0);
avg1 = vhadd_u16(dd1, d1);
avg2 = vhadd_u16(dd2, d2);
avg3 = vhadd_u16(dd3, d3);
}
int16x8_t dst_01 = vreinterpretq_s16_u16(vcombine_u16(avg0, avg1));
int16x8_t dst_23 = vreinterpretq_s16_u16(vcombine_u16(avg2, avg3));
dst_01 = vsubq_s16(dst_01, round_offset);
dst_23 = vsubq_s16(dst_23, round_offset);
*d01_u8 = vqrshrun_n_s16(dst_01, FILTER_BITS - ROUND0_BITS);
*d23_u8 = vqrshrun_n_s16(dst_23, FILTER_BITS - ROUND0_BITS);
}
static INLINE void compute_avg_8x4(
uint16x8_t dd0, uint16x8_t dd1, uint16x8_t dd2, uint16x8_t dd3,
uint16x8_t d0, uint16x8_t d1, uint16x8_t d2, uint16x8_t d3,
const uint16_t fwd_offset, const uint16_t bck_offset,
const int16x8_t round_offset, const int use_dist_wtd_comp_avg,
uint8x8_t *d0_u8, uint8x8_t *d1_u8, uint8x8_t *d2_u8, uint8x8_t *d3_u8) {
uint16x8_t avg0, avg1, avg2, avg3;
if (use_dist_wtd_comp_avg) {
uint32x4_t blend0_lo, blend1_lo, blend2_lo, blend3_lo;
uint32x4_t blend0_hi, blend1_hi, blend2_hi, blend3_hi;
blend0_lo = vmull_n_u16(vget_low_u16(dd0), fwd_offset);
blend0_lo = vmlal_n_u16(blend0_lo, vget_low_u16(d0), bck_offset);
blend0_hi = vmull_n_u16(vget_high_u16(dd0), fwd_offset);
blend0_hi = vmlal_n_u16(blend0_hi, vget_high_u16(d0), bck_offset);
blend1_lo = vmull_n_u16(vget_low_u16(dd1), fwd_offset);
blend1_lo = vmlal_n_u16(blend1_lo, vget_low_u16(d1), bck_offset);
blend1_hi = vmull_n_u16(vget_high_u16(dd1), fwd_offset);
blend1_hi = vmlal_n_u16(blend1_hi, vget_high_u16(d1), bck_offset);
blend2_lo = vmull_n_u16(vget_low_u16(dd2), fwd_offset);
blend2_lo = vmlal_n_u16(blend2_lo, vget_low_u16(d2), bck_offset);
blend2_hi = vmull_n_u16(vget_high_u16(dd2), fwd_offset);
blend2_hi = vmlal_n_u16(blend2_hi, vget_high_u16(d2), bck_offset);
blend3_lo = vmull_n_u16(vget_low_u16(dd3), fwd_offset);
blend3_lo = vmlal_n_u16(blend3_lo, vget_low_u16(d3), bck_offset);
blend3_hi = vmull_n_u16(vget_high_u16(dd3), fwd_offset);
blend3_hi = vmlal_n_u16(blend3_hi, vget_high_u16(d3), bck_offset);
avg0 = vcombine_u16(vshrn_n_u32(blend0_lo, DIST_PRECISION_BITS),
vshrn_n_u32(blend0_hi, DIST_PRECISION_BITS));
avg1 = vcombine_u16(vshrn_n_u32(blend1_lo, DIST_PRECISION_BITS),
vshrn_n_u32(blend1_hi, DIST_PRECISION_BITS));
avg2 = vcombine_u16(vshrn_n_u32(blend2_lo, DIST_PRECISION_BITS),
vshrn_n_u32(blend2_hi, DIST_PRECISION_BITS));
avg3 = vcombine_u16(vshrn_n_u32(blend3_lo, DIST_PRECISION_BITS),
vshrn_n_u32(blend3_hi, DIST_PRECISION_BITS));
} else {
avg0 = vhaddq_u16(dd0, d0);
avg1 = vhaddq_u16(dd1, d1);
avg2 = vhaddq_u16(dd2, d2);
avg3 = vhaddq_u16(dd3, d3);
}
int16x8_t dst0 = vsubq_s16(vreinterpretq_s16_u16(avg0), round_offset);
int16x8_t dst1 = vsubq_s16(vreinterpretq_s16_u16(avg1), round_offset);
int16x8_t dst2 = vsubq_s16(vreinterpretq_s16_u16(avg2), round_offset);
int16x8_t dst3 = vsubq_s16(vreinterpretq_s16_u16(avg3), round_offset);
*d0_u8 = vqrshrun_n_s16(dst0, FILTER_BITS - ROUND0_BITS);
*d1_u8 = vqrshrun_n_s16(dst1, FILTER_BITS - ROUND0_BITS);
*d2_u8 = vqrshrun_n_s16(dst2, FILTER_BITS - ROUND0_BITS);
*d3_u8 = vqrshrun_n_s16(dst3, FILTER_BITS - ROUND0_BITS);
}
#if defined(__aarch64__) && defined(__ARM_FEATURE_MATMUL_INT8)
static INLINE int16x4_t convolve8_4_2d_h(uint8x16_t samples,
const int8x8_t x_filter,
const uint8x16x2_t permute_tbl,
const int32x4_t horiz_const) {
uint8x16_t permuted_samples[2];
int32x4_t sum;
// Permute samples ready for dot product.
// { 0, 1, 2, 3, 1, 2, 3, 4, 2, 3, 4, 5, 3, 4, 5, 6 }
permuted_samples[0] = vqtbl1q_u8(samples, permute_tbl.val[0]);
// { 4, 5, 6, 7, 5, 6, 7, 8, 6, 7, 8, 9, 7, 8, 9, 10 }
permuted_samples[1] = vqtbl1q_u8(samples, permute_tbl.val[1]);
// First 4 output values.
sum = vusdotq_lane_s32(horiz_const, permuted_samples[0], x_filter, 0);
sum = vusdotq_lane_s32(sum, permuted_samples[1], x_filter, 1);
// We halved the convolution filter values so -1 from the right shift.
return vshrn_n_s32(sum, ROUND0_BITS - 1);
}
static INLINE int16x8_t convolve8_8_2d_h(uint8x16_t samples,
const int8x8_t x_filter,
const uint8x16x3_t permute_tbl,
const int32x4_t horiz_const) {
uint8x16_t permuted_samples[3];
int32x4_t sum[2];
// Permute samples ready for dot product.
// { 0, 1, 2, 3, 1, 2, 3, 4, 2, 3, 4, 5, 3, 4, 5, 6 }
permuted_samples[0] = vqtbl1q_u8(samples, permute_tbl.val[0]);
// { 4, 5, 6, 7, 5, 6, 7, 8, 6, 7, 8, 9, 7, 8, 9, 10 }
permuted_samples[1] = vqtbl1q_u8(samples, permute_tbl.val[1]);
// { 8, 9, 10, 11, 9, 10, 11, 12, 10, 11, 12, 13, 11, 12, 13, 14 }
permuted_samples[2] = vqtbl1q_u8(samples, permute_tbl.val[2]);
// First 4 output values.
sum[0] = vusdotq_lane_s32(horiz_const, permuted_samples[0], x_filter, 0);
sum[0] = vusdotq_lane_s32(sum[0], permuted_samples[1], x_filter, 1);
// Second 4 output values.
sum[1] = vusdotq_lane_s32(horiz_const, permuted_samples[1], x_filter, 0);
sum[1] = vusdotq_lane_s32(sum[1], permuted_samples[2], x_filter, 1);
// Narrow and re-pack.
// We halved the convolution filter values so -1 from the right shift.
return vcombine_s16(vshrn_n_s32(sum[0], ROUND0_BITS - 1),
vshrn_n_s32(sum[1], ROUND0_BITS - 1));
}
static INLINE void dist_wtd_convolve_2d_horiz_8tap_neon(
const uint8_t *src, int src_stride, int16_t *im_block, const int im_stride,
const int16x8_t x_filter_s16, const int im_h, int w) {
const int bd = 8;
// A shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-rounding
// shifts - which are generally faster than rounding shifts on modern CPUs.
// (The extra -1 is needed because we halved the filter values.)
const int32x4_t horiz_const = vdupq_n_s32((1 << (bd + FILTER_BITS - 2)) +
(1 << ((ROUND0_BITS - 1) - 1)));
// Horizontal filter.
const int8x8_t x_filter = vmovn_s16(x_filter_s16);
const uint8_t *src_ptr = src;
int16_t *dst_ptr = im_block;
int dst_stride = im_stride;
int height = im_h;
if (w == 4) {
const uint8x16x2_t permute_tbl = vld1q_u8_x2(dot_prod_permute_tbl);
uint8x16_t s0, s1, s2, s3;
int16x4_t d0, d1, d2, d3;
do {
load_u8_16x4(src_ptr, src_stride, &s0, &s1, &s2, &s3);
d0 = convolve8_4_2d_h(s0, x_filter, permute_tbl, horiz_const);
d1 = convolve8_4_2d_h(s1, x_filter, permute_tbl, horiz_const);
d2 = convolve8_4_2d_h(s2, x_filter, permute_tbl, horiz_const);
d3 = convolve8_4_2d_h(s3, x_filter, permute_tbl, horiz_const);
store_s16_4x4(dst_ptr, dst_stride, d0, d1, d2, d3);
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
height -= 4;
} while (height > 0);
} else {
const uint8x16x3_t permute_tbl = vld1q_u8_x3(dot_prod_permute_tbl);
uint8x16_t s0, s1, s2, s3;
int16x8_t d0, d1, d2, d3;
do {
const uint8_t *s = src_ptr;
int16_t *d = dst_ptr;
int width = w;
do {
load_u8_16x4(s, src_stride, &s0, &s1, &s2, &s3);
d0 = convolve8_8_2d_h(s0, x_filter, permute_tbl, horiz_const);
d1 = convolve8_8_2d_h(s1, x_filter, permute_tbl, horiz_const);
d2 = convolve8_8_2d_h(s2, x_filter, permute_tbl, horiz_const);
d3 = convolve8_8_2d_h(s3, x_filter, permute_tbl, horiz_const);
store_s16_8x4(d, dst_stride, d0, d1, d2, d3);
s += 8;
d += 8;
width -= 8;
} while (width > 0);
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
height -= 4;
} while (height > 0);
}
}
#elif defined(__aarch64__) && defined(__ARM_FEATURE_DOTPROD)
static INLINE int16x4_t convolve8_4_2d_h(uint8x16_t samples,
const int8x8_t x_filter,
const int32x4_t correction,
const uint8x16_t range_limit,
const uint8x16x2_t permute_tbl) {
int8x16_t clamped_samples, permuted_samples[2];
int32x4_t sum;
// Clamp sample range to [-128, 127] for 8-bit signed dot product.
clamped_samples = vreinterpretq_s8_u8(vsubq_u8(samples, range_limit));
// Permute samples ready for dot product.
// { 0, 1, 2, 3, 1, 2, 3, 4, 2, 3, 4, 5, 3, 4, 5, 6 }
permuted_samples[0] = vqtbl1q_s8(clamped_samples, permute_tbl.val[0]);
// { 4, 5, 6, 7, 5, 6, 7, 8, 6, 7, 8, 9, 7, 8, 9, 10 }
permuted_samples[1] = vqtbl1q_s8(clamped_samples, permute_tbl.val[1]);
// Accumulate dot product into 'correction' to account for range clamp.
sum = vdotq_lane_s32(correction, permuted_samples[0], x_filter, 0);
sum = vdotq_lane_s32(sum, permuted_samples[1], x_filter, 1);
// We halved the convolution filter values so -1 from the right shift.
return vshrn_n_s32(sum, ROUND0_BITS - 1);
}
static INLINE int16x8_t convolve8_8_2d_h(uint8x16_t samples,
const int8x8_t x_filter,
const int32x4_t correction,
const uint8x16_t range_limit,
const uint8x16x3_t permute_tbl) {
int8x16_t clamped_samples, permuted_samples[3];
int32x4_t sum[2];
// Clamp sample range to [-128, 127] for 8-bit signed dot product.
clamped_samples = vreinterpretq_s8_u8(vsubq_u8(samples, range_limit));
// Permute samples ready for dot product. */
// { 0, 1, 2, 3, 1, 2, 3, 4, 2, 3, 4, 5, 3, 4, 5, 6 }
permuted_samples[0] = vqtbl1q_s8(clamped_samples, permute_tbl.val[0]);
// { 4, 5, 6, 7, 5, 6, 7, 8, 6, 7, 8, 9, 7, 8, 9, 10 }
permuted_samples[1] = vqtbl1q_s8(clamped_samples, permute_tbl.val[1]);
// { 8, 9, 10, 11, 9, 10, 11, 12, 10, 11, 12, 13, 11, 12, 13, 14 }
permuted_samples[2] = vqtbl1q_s8(clamped_samples, permute_tbl.val[2]);
// Accumulate dot product into 'correction' to account for range clamp.
// First 4 output values.
sum[0] = vdotq_lane_s32(correction, permuted_samples[0], x_filter, 0);
sum[0] = vdotq_lane_s32(sum[0], permuted_samples[1], x_filter, 1);
// Second 4 output values.
sum[1] = vdotq_lane_s32(correction, permuted_samples[1], x_filter, 0);
sum[1] = vdotq_lane_s32(sum[1], permuted_samples[2], x_filter, 1);
// Narrow and re-pack.
// We halved the convolution filter values so -1 from the right shift.
return vcombine_s16(vshrn_n_s32(sum[0], ROUND0_BITS - 1),
vshrn_n_s32(sum[1], ROUND0_BITS - 1));
}
static INLINE void dist_wtd_convolve_2d_horiz_8tap_neon(
const uint8_t *src, int src_stride, int16_t *im_block, const int im_stride,
const int16x8_t x_filter_s16, const int im_h, int w) {
const int bd = 8;
const int32_t horiz_const = (1 << (bd + FILTER_BITS - 2));
// Dot product constants and other shims.
const int32_t correction_s32 = vaddlvq_s16(vshlq_n_s16(x_filter_s16, 7));
// Fold horiz_const into the dot-product filter correction constant. The
// additional shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-
// rounding shifts - which are generally faster than rounding shifts on
// modern CPUs. (The extra -1 is needed because we halved the filter values.)
const int32x4_t correction = vdupq_n_s32(correction_s32 + horiz_const +
(1 << ((ROUND0_BITS - 1) - 1)));
const uint8x16_t range_limit = vdupq_n_u8(128);
// Horizontal filter.
const int8x8_t x_filter = vmovn_s16(x_filter_s16);
const uint8_t *src_ptr = src;
int16_t *dst_ptr = im_block;
int dst_stride = im_stride;
int height = im_h;
if (w == 4) {
const uint8x16x2_t permute_tbl = vld1q_u8_x2(dot_prod_permute_tbl);
uint8x16_t s0, s1, s2, s3;
int16x4_t d0, d1, d2, d3;
do {
load_u8_16x4(src_ptr, src_stride, &s0, &s1, &s2, &s3);
d0 = convolve8_4_2d_h(s0, x_filter, correction, range_limit, permute_tbl);
d1 = convolve8_4_2d_h(s1, x_filter, correction, range_limit, permute_tbl);
d2 = convolve8_4_2d_h(s2, x_filter, correction, range_limit, permute_tbl);
d3 = convolve8_4_2d_h(s3, x_filter, correction, range_limit, permute_tbl);
store_s16_4x4(dst_ptr, dst_stride, d0, d1, d2, d3);
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
height -= 4;
} while (height > 0);
} else {
const uint8x16x3_t permute_tbl = vld1q_u8_x3(dot_prod_permute_tbl);
uint8x16_t s0, s1, s2, s3;
int16x8_t d0, d1, d2, d3;
do {
const uint8_t *s = src_ptr;
int16_t *d = dst_ptr;
int width = w;
do {
load_u8_16x4(s, src_stride, &s0, &s1, &s2, &s3);
d0 = convolve8_8_2d_h(s0, x_filter, correction, range_limit,
permute_tbl);
d1 = convolve8_8_2d_h(s1, x_filter, correction, range_limit,
permute_tbl);
d2 = convolve8_8_2d_h(s2, x_filter, correction, range_limit,
permute_tbl);
d3 = convolve8_8_2d_h(s3, x_filter, correction, range_limit,
permute_tbl);
store_s16_8x4(d, dst_stride, d0, d1, d2, d3);
s += 8;
d += 8;
width -= 8;
} while (width > 0);
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
height -= 4;
} while (height > 0);
}
}
#else // !(defined(__aarch64__) && defined(__ARM_FEATURE_DOTPROD))
static INLINE int16x4_t convolve8_4_2d_h(const int16x4_t s0, const int16x4_t s1,
const int16x4_t s2, const int16x4_t s3,
const int16x4_t s4, const int16x4_t s5,
const int16x4_t s6, const int16x4_t s7,
const int16x8_t x_filter,
const int16x4_t horiz_const) {
const int16x4_t x_filter_0_3 = vget_low_s16(x_filter);
const int16x4_t x_filter_4_7 = vget_high_s16(x_filter);
int16x4_t sum = horiz_const;
sum = vmla_lane_s16(sum, s0, x_filter_0_3, 0);
sum = vmla_lane_s16(sum, s1, x_filter_0_3, 1);
sum = vmla_lane_s16(sum, s2, x_filter_0_3, 2);
sum = vmla_lane_s16(sum, s3, x_filter_0_3, 3);
sum = vmla_lane_s16(sum, s4, x_filter_4_7, 0);
sum = vmla_lane_s16(sum, s5, x_filter_4_7, 1);
sum = vmla_lane_s16(sum, s6, x_filter_4_7, 2);
sum = vmla_lane_s16(sum, s7, x_filter_4_7, 3);
// We halved the convolution filter values so -1 from the right shift.
return vshr_n_s16(sum, ROUND0_BITS - 1);
}
static INLINE int16x8_t convolve8_8_2d_h(const int16x8_t s0, const int16x8_t s1,
const int16x8_t s2, const int16x8_t s3,
const int16x8_t s4, const int16x8_t s5,
const int16x8_t s6, const int16x8_t s7,
const int16x8_t x_filter,
const int16x8_t horiz_const) {
const int16x4_t x_filter_0_3 = vget_low_s16(x_filter);
const int16x4_t x_filter_4_7 = vget_high_s16(x_filter);
int16x8_t sum = horiz_const;
sum = vmlaq_lane_s16(sum, s0, x_filter_0_3, 0);
sum = vmlaq_lane_s16(sum, s1, x_filter_0_3, 1);
sum = vmlaq_lane_s16(sum, s2, x_filter_0_3, 2);
sum = vmlaq_lane_s16(sum, s3, x_filter_0_3, 3);
sum = vmlaq_lane_s16(sum, s4, x_filter_4_7, 0);
sum = vmlaq_lane_s16(sum, s5, x_filter_4_7, 1);
sum = vmlaq_lane_s16(sum, s6, x_filter_4_7, 2);
sum = vmlaq_lane_s16(sum, s7, x_filter_4_7, 3);
// We halved the convolution filter values so -1 from the right shift.
return vshrq_n_s16(sum, ROUND0_BITS - 1);
}
static INLINE void dist_wtd_convolve_2d_horiz_8tap_neon(
const uint8_t *src, int src_stride, int16_t *im_block, const int im_stride,
const int16x8_t x_filter, const int im_h, int w) {
const int bd = 8;
const uint8_t *src_ptr = src;
int16_t *dst_ptr = im_block;
int dst_stride = im_stride;
int height = im_h;
if (w == 4) {
int16x4_t s0, s1, s2, s3, s4, s5, s6, s7, d0;
uint8x8_t t0;
#if defined(__aarch64__)
int16x4_t s8, s9, s10, d1, d2, d3;
uint8x8_t t1, t2, t3;
#endif // defined(__aarch64__)
// A shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-rounding
// shifts - which are generally faster than rounding shifts on modern CPUs.
// (The extra -1 is needed because we halved the filter values.)
const int16x4_t horiz_const = vdup_n_s16((1 << (bd + FILTER_BITS - 2)) +
(1 << ((ROUND0_BITS - 1) - 1)));
do {
__builtin_prefetch(src_ptr + 0 * src_stride);
#if defined(__aarch64__)
__builtin_prefetch(src_ptr + 1 * src_stride);
__builtin_prefetch(src_ptr + 2 * src_stride);
__builtin_prefetch(src_ptr + 3 * src_stride);
load_u8_8x4(src_ptr, src_stride, &t0, &t1, &t2, &t3);
transpose_u8_8x4(&t0, &t1, &t2, &t3);
s0 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s1 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
s2 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t2)));
s3 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t3)));
s4 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s5 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
s6 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t2)));
__builtin_prefetch(dst_ptr + 0 * dst_stride);
__builtin_prefetch(dst_ptr + 1 * dst_stride);
__builtin_prefetch(dst_ptr + 2 * dst_stride);
__builtin_prefetch(dst_ptr + 3 * dst_stride);
load_u8_8x4(src_ptr + 7, src_stride, &t0, &t1, &t2, &t3);
transpose_u8_8x4(&t0, &t1, &t2, &t3);
s7 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s8 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
s9 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t2)));
s10 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t3)));
d0 = convolve8_4_2d_h(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
horiz_const);
d1 = convolve8_4_2d_h(s1, s2, s3, s4, s5, s6, s7, s8, x_filter,
horiz_const);
d2 = convolve8_4_2d_h(s2, s3, s4, s5, s6, s7, s8, s9, x_filter,
horiz_const);
d3 = convolve8_4_2d_h(s3, s4, s5, s6, s7, s8, s9, s10, x_filter,
horiz_const);
transpose_s16_4x4d(&d0, &d1, &d2, &d3);
store_s16_4x4(dst_ptr, dst_stride, d0, d1, d2, d3);
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
height -= 4;
#else // !defined(__aarch64__)
t0 = vld1_u8(src_ptr); // a0 a1 a2 a3 a4 a5 a6 a7
s0 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0))); // a0 a1 a2 a3
s4 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t0))); // a4 a5 a6 a7
__builtin_prefetch(dst_ptr);
t0 = vld1_u8(src_ptr + 8); // a8 a9 a10 a11
s7 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s1 = vext_s16(s0, s4, 1); // a1 a2 a3 a4
s2 = vext_s16(s0, s4, 2); // a2 a3 a4 a5
s3 = vext_s16(s0, s4, 3); // a3 a4 a5 a6
s5 = vext_s16(s4, s7, 1); // a5 a6 a7 a8
s6 = vext_s16(s4, s7, 2); // a6 a7 a8 a9
s7 = vext_s16(s4, s7, 3); // a7 a8 a9 a10
d0 = convolve8_4_2d_h(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
horiz_const);
vst1_s16(dst_ptr, d0);
src_ptr += src_stride;
dst_ptr += dst_stride;
height--;
#endif // defined(__aarch64__)
} while (height > 0);
} else {
int16x8_t s0, s1, s2, s3, s4, s5, s6, s7, s8, d0;
uint8x8_t t0;
#if defined(__aarch64__)
int16x8_t s9, s10, s11, s12, s13, s14;
int16x8_t d1, d2, d3, d4, d5, d6, d7;
uint8x8_t t1, t2, t3, t4, t5, t6, t7;
#endif // defined(__aarch64__)
// A shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-rounding
// shifts - which are generally faster than rounding shifts on modern CPUs.
// (The extra -1 is needed because we halved the filter values.)
const int16x8_t horiz_const = vdupq_n_s16((1 << (bd + FILTER_BITS - 2)) +
(1 << ((ROUND0_BITS - 1) - 1)));
do {
const uint8_t *s;
int16_t *d = dst_ptr;
int width = w;
#if defined(__aarch64__)
__builtin_prefetch(src_ptr + 0 * src_stride);
__builtin_prefetch(src_ptr + 1 * src_stride);
__builtin_prefetch(src_ptr + 2 * src_stride);
__builtin_prefetch(src_ptr + 3 * src_stride);
__builtin_prefetch(src_ptr + 4 * src_stride);
__builtin_prefetch(src_ptr + 5 * src_stride);
__builtin_prefetch(src_ptr + 6 * src_stride);
__builtin_prefetch(src_ptr + 7 * src_stride);
load_u8_8x8(src_ptr, src_stride, &t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
transpose_u8_8x8(&t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
s0 = vreinterpretq_s16_u16(vmovl_u8(t0));
s1 = vreinterpretq_s16_u16(vmovl_u8(t1));
s2 = vreinterpretq_s16_u16(vmovl_u8(t2));
s3 = vreinterpretq_s16_u16(vmovl_u8(t3));
s4 = vreinterpretq_s16_u16(vmovl_u8(t4));
s5 = vreinterpretq_s16_u16(vmovl_u8(t5));
s6 = vreinterpretq_s16_u16(vmovl_u8(t6));
s = src_ptr + 7;
__builtin_prefetch(dst_ptr + 0 * dst_stride);
__builtin_prefetch(dst_ptr + 1 * dst_stride);
__builtin_prefetch(dst_ptr + 2 * dst_stride);
__builtin_prefetch(dst_ptr + 3 * dst_stride);
__builtin_prefetch(dst_ptr + 4 * dst_stride);
__builtin_prefetch(dst_ptr + 5 * dst_stride);
__builtin_prefetch(dst_ptr + 6 * dst_stride);
__builtin_prefetch(dst_ptr + 7 * dst_stride);
do {
load_u8_8x8(s, src_stride, &t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
transpose_u8_8x8(&t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
s7 = vreinterpretq_s16_u16(vmovl_u8(t0));
s8 = vreinterpretq_s16_u16(vmovl_u8(t1));
s9 = vreinterpretq_s16_u16(vmovl_u8(t2));
s10 = vreinterpretq_s16_u16(vmovl_u8(t3));
s11 = vreinterpretq_s16_u16(vmovl_u8(t4));
s12 = vreinterpretq_s16_u16(vmovl_u8(t5));
s13 = vreinterpretq_s16_u16(vmovl_u8(t6));
s14 = vreinterpretq_s16_u16(vmovl_u8(t7));
d0 = convolve8_8_2d_h(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
horiz_const);
d1 = convolve8_8_2d_h(s1, s2, s3, s4, s5, s6, s7, s8, x_filter,
horiz_const);
d2 = convolve8_8_2d_h(s2, s3, s4, s5, s6, s7, s8, s9, x_filter,
horiz_const);
d3 = convolve8_8_2d_h(s3, s4, s5, s6, s7, s8, s9, s10, x_filter,
horiz_const);
d4 = convolve8_8_2d_h(s4, s5, s6, s7, s8, s9, s10, s11, x_filter,
horiz_const);
d5 = convolve8_8_2d_h(s5, s6, s7, s8, s9, s10, s11, s12, x_filter,
horiz_const);
d6 = convolve8_8_2d_h(s6, s7, s8, s9, s10, s11, s12, s13, x_filter,
horiz_const);
d7 = convolve8_8_2d_h(s7, s8, s9, s10, s11, s12, s13, s14, x_filter,
horiz_const);
transpose_s16_8x8(&d0, &d1, &d2, &d3, &d4, &d5, &d6, &d7);
store_s16_8x8(d, dst_stride, d0, d1, d2, d3, d4, d5, d6, d7);
s0 = s8;
s1 = s9;
s2 = s10;
s3 = s11;
s4 = s12;
s5 = s13;
s6 = s14;
s += 8;
d += 8;
width -= 8;
} while (width > 0);
src_ptr += 8 * src_stride;
dst_ptr += 8 * dst_stride;
height -= 8;
#else // !defined(__aarch64__)
t0 = vld1_u8(src_ptr);
s0 = vreinterpretq_s16_u16(vmovl_u8(t0)); // a0 a1 a2 a3 a4 a5 a6 a7
s = src_ptr + 8;
__builtin_prefetch(dst_ptr);
do {
t0 = vld1_u8(s); // a8 a9 a10 a11 a12 a13 a14 a15
s8 = vreinterpretq_s16_u16(vmovl_u8(t0));
s1 = vextq_s16(s0, s8, 1); // a1 a2 a3 a4 a5 a6 a7 a8
s2 = vextq_s16(s0, s8, 2); // a2 a3 a4 a5 a6 a7 a8 a9
s3 = vextq_s16(s0, s8, 3); // a3 a4 a5 a6 a7 a8 a9 a10
s4 = vextq_s16(s0, s8, 4); // a4 a5 a6 a7 a8 a9 a10 a11
s5 = vextq_s16(s0, s8, 5); // a5 a6 a7 a8 a9 a10 a11 a12
s6 = vextq_s16(s0, s8, 6); // a6 a7 a8 a9 a10 a11 a12 a13
s7 = vextq_s16(s0, s8, 7); // a7 a8 a9 a10 a11 a12 a13 a14
d0 = convolve8_8_2d_h(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
horiz_const);
vst1q_s16(d, d0);
s0 = s8;
s += 8;
d += 8;
width -= 8;
} while (width > 0);
src_ptr += src_stride;
dst_ptr += dst_stride;
height--;
#endif // defined(__aarch64__)
} while (height > 0);
}
}
#endif // defined(__aarch64__) && defined(__ARM_FEATURE_DOTPROD)
static INLINE uint16x4_t
convolve6_4_2d_v(const int16x4_t s0, const int16x4_t s1, const int16x4_t s2,
const int16x4_t s3, const int16x4_t s4, const int16x4_t s5,
const int16x8_t y_filter, const int32x4_t offset_const) {
const int16x4_t y_filter_0_3 = vget_low_s16(y_filter);
const int16x4_t y_filter_4_7 = vget_high_s16(y_filter);
int32x4_t sum = offset_const;
// Filter values at indices 0 and 7 are 0.
sum = vmlal_lane_s16(sum, s0, y_filter_0_3, 1);
sum = vmlal_lane_s16(sum, s1, y_filter_0_3, 2);
sum = vmlal_lane_s16(sum, s2, y_filter_0_3, 3);
sum = vmlal_lane_s16(sum, s3, y_filter_4_7, 0);
sum = vmlal_lane_s16(sum, s4, y_filter_4_7, 1);
sum = vmlal_lane_s16(sum, s5, y_filter_4_7, 2);
return vqrshrun_n_s32(sum, COMPOUND_ROUND1_BITS);
}
static INLINE uint16x8_t
convolve6_8_2d_v(const int16x8_t s0, const int16x8_t s1, const int16x8_t s2,
const int16x8_t s3, const int16x8_t s4, const int16x8_t s5,
const int16x8_t y_filter, const int32x4_t offset_const) {
const int16x4_t y_filter_0_3 = vget_low_s16(y_filter);
const int16x4_t y_filter_4_7 = vget_high_s16(y_filter);
int32x4_t sum0 = offset_const;
// Filter values at indices 0 and 7 are 0.
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s0), y_filter_0_3, 1);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s1), y_filter_0_3, 2);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s2), y_filter_0_3, 3);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s3), y_filter_4_7, 0);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s4), y_filter_4_7, 1);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s5), y_filter_4_7, 2);
int32x4_t sum1 = offset_const;
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s0), y_filter_0_3, 1);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s1), y_filter_0_3, 2);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s2), y_filter_0_3, 3);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s3), y_filter_4_7, 0);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s4), y_filter_4_7, 1);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s5), y_filter_4_7, 2);
return vcombine_u16(vqrshrun_n_s32(sum0, COMPOUND_ROUND1_BITS),
vqrshrun_n_s32(sum1, COMPOUND_ROUND1_BITS));
}
static INLINE void dist_wtd_convolve_2d_vert_6tap_neon(
int16_t *src_ptr, const int src_stride, uint8_t *dst8_ptr, int dst8_stride,
ConvolveParams *conv_params, const int16x8_t y_filter, int h, int w) {
const int bd = 8;
const int offset_bits = bd + 2 * FILTER_BITS - ROUND0_BITS;
const int32x4_t offset_const = vdupq_n_s32(1 << offset_bits);
const int16_t round_offset = (1 << (offset_bits - COMPOUND_ROUND1_BITS)) +
(1 << (offset_bits - COMPOUND_ROUND1_BITS - 1));
const int16x8_t round_offset_vec = vdupq_n_s16(round_offset);
const int do_average = conv_params->do_average;
const int use_dist_wtd_comp_avg = conv_params->use_dist_wtd_comp_avg;
const uint16_t fwd_offset = conv_params->fwd_offset;
const uint16_t bck_offset = conv_params->bck_offset;
CONV_BUF_TYPE *dst_ptr = conv_params->dst;
const int dst_stride = conv_params->dst_stride;
if (w == 4) {
int16x4_t s0, s1, s2, s3, s4, s5;
uint16x4_t dd0, d0;
uint8x8_t d01_u8;
#if defined(__aarch64__)
int16x4_t s6, s7, s8;
uint16x4_t dd1, dd2, dd3, d1, d2, d3;
uint8x8_t d23_u8;
#endif // defined(__aarch64__)
load_s16_4x5(src_ptr, src_stride, &s0, &s1, &s2, &s3, &s4);
src_ptr += 5 * src_stride;
do {
#if defined(__aarch64__)
load_s16_4x4(src_ptr, src_stride, &s5, &s6, &s7, &s8);
d0 = convolve6_4_2d_v(s0, s1, s2, s3, s4, s5, y_filter, offset_const);
d1 = convolve6_4_2d_v(s1, s2, s3, s4, s5, s6, y_filter, offset_const);
d2 = convolve6_4_2d_v(s2, s3, s4, s5, s6, s7, y_filter, offset_const);
d3 = convolve6_4_2d_v(s3, s4, s5, s6, s7, s8, y_filter, offset_const);
if (do_average) {
load_u16_4x4(dst_ptr, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_4x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d01_u8, &d23_u8);
store_u8_4x1(dst8_ptr + 0 * dst8_stride, d01_u8, 0);
store_u8_4x1(dst8_ptr + 1 * dst8_stride, d01_u8, 1);
store_u8_4x1(dst8_ptr + 2 * dst8_stride, d23_u8, 0);
store_u8_4x1(dst8_ptr + 3 * dst8_stride, d23_u8, 1);
dst8_ptr += 4 * dst8_stride;
} else {
store_u16_4x4(dst_ptr, dst_stride, d0, d1, d2, d3);
}
s0 = s4;
s1 = s5;
s2 = s6;
s3 = s7;
s4 = s8;
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
h -= 4;
#else // !defined(__aarch64__)
s5 = vld1_s16(src_ptr);
d0 = convolve6_4_2d_v(s0, s1, s2, s3, s4, s5, y_filter, offset_const);
if (do_average) {
dd0 = vld1_u16(dst_ptr);
compute_avg_4x1(dd0, d0, fwd_offset, bck_offset,
vget_low_s16(round_offset_vec), use_dist_wtd_comp_avg,
&d01_u8);
store_u8_4x1(dst8_ptr, d01_u8, 0);
dst8_ptr += dst8_stride;
} else {
vst1_u16(dst_ptr, d0);
}
s0 = s1;
s1 = s2;
s2 = s3;
s3 = s4;
s4 = s5;
src_ptr += src_stride;
dst_ptr += dst_stride;
h--;
#endif // defined(__aarch64__)
} while (h != 0);
} else {
int16x8_t s0, s1, s2, s3, s4, s5;
uint16x8_t dd0, d0;
uint8x8_t d0_u8;
#if defined(__aarch64__)
int16x8_t s6, s7, s8;
uint16x8_t dd1, dd2, dd3, d1, d2, d3;
uint8x8_t d1_u8, d2_u8, d3_u8;
#endif // defined(__aarch64__)
do {
int16_t *s = src_ptr;
CONV_BUF_TYPE *d = dst_ptr;
uint8_t *d_u8 = dst8_ptr;
int height = h;
load_s16_8x5(s, src_stride, &s0, &s1, &s2, &s3, &s4);
s += 5 * src_stride;
do {
#if defined(__aarch64__)
load_s16_8x4(s, src_stride, &s5, &s6, &s7, &s8);
d0 = convolve6_8_2d_v(s0, s1, s2, s3, s4, s5, y_filter, offset_const);
d1 = convolve6_8_2d_v(s1, s2, s3, s4, s5, s6, y_filter, offset_const);
d2 = convolve6_8_2d_v(s2, s3, s4, s5, s6, s7, y_filter, offset_const);
d3 = convolve6_8_2d_v(s3, s4, s5, s6, s7, s8, y_filter, offset_const);
if (do_average) {
load_u16_8x4(d, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_8x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d0_u8, &d1_u8, &d2_u8, &d3_u8);
store_u8_8x4(d_u8, dst8_stride, d0_u8, d1_u8, d2_u8, d3_u8);
d_u8 += 4 * dst8_stride;
} else {
store_u16_8x4(d, dst_stride, d0, d1, d2, d3);
}
s0 = s4;
s1 = s5;
s2 = s6;
s3 = s7;
s4 = s8;
s += 4 * src_stride;
d += 4 * dst_stride;
height -= 4;
#else // !defined(__aarch64__)
s5 = vld1q_s16(s);
d0 = convolve6_8_2d_v(s0, s1, s2, s3, s4, s5, y_filter, offset_const);
if (do_average) {
dd0 = vld1q_u16(d);
compute_avg_8x1(dd0, d0, fwd_offset, bck_offset, round_offset_vec,
use_dist_wtd_comp_avg, &d0_u8);
vst1_u8(d_u8, d0_u8);
d_u8 += dst8_stride;
} else {
vst1q_u16(d, d0);
}
s0 = s1;
s1 = s2;
s2 = s3;
s3 = s4;
s4 = s5;
s += src_stride;
d += dst_stride;
height--;
#endif // defined(__aarch64__)
} while (height != 0);
src_ptr += 8;
dst_ptr += 8;
dst8_ptr += 8;
w -= 8;
} while (w != 0);
}
}
static INLINE uint16x4_t
convolve8_4_2d_v(const int16x4_t s0, const int16x4_t s1, const int16x4_t s2,
const int16x4_t s3, const int16x4_t s4, const int16x4_t s5,
const int16x4_t s6, const int16x4_t s7,
const int16x8_t y_filter, const int32x4_t offset_const) {
const int16x4_t y_filter_0_3 = vget_low_s16(y_filter);
const int16x4_t y_filter_4_7 = vget_high_s16(y_filter);
int32x4_t sum = offset_const;
sum = vmlal_lane_s16(sum, s0, y_filter_0_3, 0);
sum = vmlal_lane_s16(sum, s1, y_filter_0_3, 1);
sum = vmlal_lane_s16(sum, s2, y_filter_0_3, 2);
sum = vmlal_lane_s16(sum, s3, y_filter_0_3, 3);
sum = vmlal_lane_s16(sum, s4, y_filter_4_7, 0);
sum = vmlal_lane_s16(sum, s5, y_filter_4_7, 1);
sum = vmlal_lane_s16(sum, s6, y_filter_4_7, 2);
sum = vmlal_lane_s16(sum, s7, y_filter_4_7, 3);
return vqrshrun_n_s32(sum, COMPOUND_ROUND1_BITS);
}
static INLINE uint16x8_t
convolve8_8_2d_v(const int16x8_t s0, const int16x8_t s1, const int16x8_t s2,
const int16x8_t s3, const int16x8_t s4, const int16x8_t s5,
const int16x8_t s6, const int16x8_t s7,
const int16x8_t y_filter, const int32x4_t offset_const) {
const int16x4_t y_filter_0_3 = vget_low_s16(y_filter);
const int16x4_t y_filter_4_7 = vget_high_s16(y_filter);
int32x4_t sum0 = offset_const;
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s0), y_filter_0_3, 0);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s1), y_filter_0_3, 1);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s2), y_filter_0_3, 2);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s3), y_filter_0_3, 3);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s4), y_filter_4_7, 0);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s5), y_filter_4_7, 1);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s6), y_filter_4_7, 2);
sum0 = vmlal_lane_s16(sum0, vget_low_s16(s7), y_filter_4_7, 3);
int32x4_t sum1 = offset_const;
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s0), y_filter_0_3, 0);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s1), y_filter_0_3, 1);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s2), y_filter_0_3, 2);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s3), y_filter_0_3, 3);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s4), y_filter_4_7, 0);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s5), y_filter_4_7, 1);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s6), y_filter_4_7, 2);
sum1 = vmlal_lane_s16(sum1, vget_high_s16(s7), y_filter_4_7, 3);
return vcombine_u16(vqrshrun_n_s32(sum0, COMPOUND_ROUND1_BITS),
vqrshrun_n_s32(sum1, COMPOUND_ROUND1_BITS));
}
static INLINE void dist_wtd_convolve_2d_vert_8tap_neon(
int16_t *src_ptr, const int src_stride, uint8_t *dst8_ptr, int dst8_stride,
ConvolveParams *conv_params, const int16x8_t y_filter, int h, int w) {
const int bd = 8;
const int offset_bits = bd + 2 * FILTER_BITS - ROUND0_BITS;
const int32x4_t offset_const = vdupq_n_s32(1 << offset_bits);
const int16_t round_offset = (1 << (offset_bits - COMPOUND_ROUND1_BITS)) +
(1 << (offset_bits - COMPOUND_ROUND1_BITS - 1));
const int16x8_t round_offset_vec = vdupq_n_s16(round_offset);
const int do_average = conv_params->do_average;
const int use_dist_wtd_comp_avg = conv_params->use_dist_wtd_comp_avg;
const uint16_t fwd_offset = conv_params->fwd_offset;
const uint16_t bck_offset = conv_params->bck_offset;
CONV_BUF_TYPE *dst_ptr = conv_params->dst;
const int dst_stride = conv_params->dst_stride;
if (w == 4) {
int16x4_t s0, s1, s2, s3, s4, s5, s6, s7;
uint16x4_t dd0, d0;
uint8x8_t d01_u8;
#if defined(__aarch64__)
int16x4_t s8, s9, s10;
uint16x4_t dd1, dd2, dd3, d1, d2, d3;
uint8x8_t d23_u8;
#endif // defined(__aarch64__)
load_s16_4x7(src_ptr, src_stride, &s0, &s1, &s2, &s3, &s4, &s5, &s6);
src_ptr += 7 * src_stride;
do {
#if defined(__aarch64__)
load_s16_4x4(src_ptr, src_stride, &s7, &s8, &s9, &s10);
d0 = convolve8_4_2d_v(s0, s1, s2, s3, s4, s5, s6, s7, y_filter,
offset_const);
d1 = convolve8_4_2d_v(s1, s2, s3, s4, s5, s6, s7, s8, y_filter,
offset_const);
d2 = convolve8_4_2d_v(s2, s3, s4, s5, s6, s7, s8, s9, y_filter,
offset_const);
d3 = convolve8_4_2d_v(s3, s4, s5, s6, s7, s8, s9, s10, y_filter,
offset_const);
if (do_average) {
load_u16_4x4(dst_ptr, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_4x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d01_u8, &d23_u8);
store_u8_4x1(dst8_ptr + 0 * dst8_stride, d01_u8, 0);
store_u8_4x1(dst8_ptr + 1 * dst8_stride, d01_u8, 1);
store_u8_4x1(dst8_ptr + 2 * dst8_stride, d23_u8, 0);
store_u8_4x1(dst8_ptr + 3 * dst8_stride, d23_u8, 1);
dst8_ptr += 4 * dst8_stride;
} else {
store_u16_4x4(dst_ptr, dst_stride, d0, d1, d2, d3);
}
s0 = s4;
s1 = s5;
s2 = s6;
s3 = s7;
s4 = s8;
s5 = s9;
s6 = s10;
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
h -= 4;
#else // !defined(__aarch64__)
s7 = vld1_s16(src_ptr);
d0 = convolve8_4_2d_v(s0, s1, s2, s3, s4, s5, s6, s7, y_filter,
offset_const);
if (do_average) {
dd0 = vld1_u16(dst_ptr);
compute_avg_4x1(dd0, d0, fwd_offset, bck_offset,
vget_low_s16(round_offset_vec), use_dist_wtd_comp_avg,
&d01_u8);
store_u8_4x1(dst8_ptr, d01_u8, 0);
dst8_ptr += dst8_stride;
} else {
vst1_u16(dst_ptr, d0);
}
s0 = s1;
s1 = s2;
s2 = s3;
s3 = s4;
s4 = s5;
s5 = s6;
s6 = s7;
src_ptr += src_stride;
dst_ptr += dst_stride;
h--;
#endif // defined(__aarch64__)
} while (h != 0);
} else {
int16x8_t s0, s1, s2, s3, s4, s5, s6, s7;
uint16x8_t dd0, d0;
uint8x8_t d0_u8;
#if defined(__aarch64__)
int16x8_t s8, s9, s10;
uint16x8_t dd1, dd2, dd3, d1, d2, d3;
uint8x8_t d1_u8, d2_u8, d3_u8;
#endif // defined(__aarch64__)
do {
int16_t *s = src_ptr;
CONV_BUF_TYPE *d = dst_ptr;
uint8_t *d_u8 = dst8_ptr;
int height = h;
load_s16_8x7(s, src_stride, &s0, &s1, &s2, &s3, &s4, &s5, &s6);
s += 7 * src_stride;
do {
#if defined(__aarch64__)
load_s16_8x4(s, src_stride, &s7, &s8, &s9, &s10);
d0 = convolve8_8_2d_v(s0, s1, s2, s3, s4, s5, s6, s7, y_filter,
offset_const);
d1 = convolve8_8_2d_v(s1, s2, s3, s4, s5, s6, s7, s8, y_filter,
offset_const);
d2 = convolve8_8_2d_v(s2, s3, s4, s5, s6, s7, s8, s9, y_filter,
offset_const);
d3 = convolve8_8_2d_v(s3, s4, s5, s6, s7, s8, s9, s10, y_filter,
offset_const);
if (do_average) {
load_u16_8x4(d, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_8x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d0_u8, &d1_u8, &d2_u8, &d3_u8);
store_u8_8x4(d_u8, dst8_stride, d0_u8, d1_u8, d2_u8, d3_u8);
d_u8 += 4 * dst8_stride;
} else {
store_u16_8x4(d, dst_stride, d0, d1, d2, d3);
}
s0 = s4;
s1 = s5;
s2 = s6;
s3 = s7;
s4 = s8;
s5 = s9;
s6 = s10;
s += 4 * src_stride;
d += 4 * dst_stride;
height -= 4;
#else // !defined(__aarch64__)
s7 = vld1q_s16(s);
d0 = convolve8_8_2d_v(s0, s1, s2, s3, s4, s5, s6, s7, y_filter,
offset_const);
if (do_average) {
dd0 = vld1q_u16(d);
compute_avg_8x1(dd0, d0, fwd_offset, bck_offset, round_offset_vec,
use_dist_wtd_comp_avg, &d0_u8);
vst1_u8(d_u8, d0_u8);
d_u8 += dst8_stride;
} else {
vst1q_u16(d, d0);
}
s0 = s1;
s1 = s2;
s2 = s3;
s3 = s4;
s4 = s5;
s5 = s6;
s6 = s7;
s += src_stride;
d += dst_stride;
height--;
#endif // defined(__aarch64__)
} while (height != 0);
src_ptr += 8;
dst_ptr += 8;
dst8_ptr += 8;
w -= 8;
} while (w != 0);
}
}
void av1_dist_wtd_convolve_2d_neon(const uint8_t *src, int src_stride,
uint8_t *dst8, int dst8_stride, int w, int h,
const InterpFilterParams *filter_params_x,
const InterpFilterParams *filter_params_y,
const int subpel_x_qn, const int subpel_y_qn,
ConvolveParams *conv_params) {
assert(w % 4 == 0);
assert(h % 4 == 0);
DECLARE_ALIGNED(16, int16_t,
im_block[(MAX_SB_SIZE + HORIZ_EXTRA_ROWS) * MAX_SB_SIZE]);
const int y_filter_taps = get_filter_tap(filter_params_y, subpel_y_qn);
const int clamped_y_taps = y_filter_taps < 6 ? 6 : y_filter_taps;
const int im_h = h + filter_params_y->taps - 1;
const int im_stride = MAX_SB_SIZE;
const int vert_offset = filter_params_y->taps / 2 - 1;
const int horiz_offset = filter_params_x->taps / 2 - 1;
const uint8_t *src_ptr = src - vert_offset * src_stride - horiz_offset;
const int16_t *x_filter_ptr = av1_get_interp_filter_subpel_kernel(
filter_params_x, subpel_x_qn & SUBPEL_MASK);
const int16_t *y_filter_ptr = av1_get_interp_filter_subpel_kernel(
filter_params_y, subpel_y_qn & SUBPEL_MASK);
// Filter values are even, so downshift by 1 to reduce intermediate precision
// requirements.
const int16x8_t x_filter = vshrq_n_s16(vld1q_s16(x_filter_ptr), 1);
const int16x8_t y_filter = vld1q_s16(y_filter_ptr);
dist_wtd_convolve_2d_horiz_8tap_neon(src_ptr, src_stride, im_block, im_stride,
x_filter, im_h, w);
if (clamped_y_taps == 6) {
dist_wtd_convolve_2d_vert_6tap_neon(im_block + im_stride, im_stride, dst8,
dst8_stride, conv_params, y_filter, h,
w);
} else {
dist_wtd_convolve_2d_vert_8tap_neon(im_block, im_stride, dst8, dst8_stride,
conv_params, y_filter, h, w);
}
}
void av1_dist_wtd_convolve_2d_copy_neon(const uint8_t *src, int src_stride,
uint8_t *dst8, int dst8_stride, int w,
int h, ConvolveParams *conv_params) {
assert(w % 4 == 0);
assert(h % 4 == 0);
const int bd = 8;
const int offset_bits = bd + 2 * FILTER_BITS - ROUND0_BITS;
const uint16_t round_offset = (1 << (offset_bits - COMPOUND_ROUND1_BITS)) +
(1 << (offset_bits - COMPOUND_ROUND1_BITS - 1));
const uint16x8_t round_offset_vec = vdupq_n_u16(round_offset);
const uint8x8_t shift_by_bits = vdup_n_u8(1 << (FILTER_BITS - ROUND0_BITS));
const int do_average = conv_params->do_average;
const int use_dist_wtd_comp_avg = conv_params->use_dist_wtd_comp_avg;
const uint16_t fwd_offset = conv_params->fwd_offset;
const uint16_t bck_offset = conv_params->bck_offset;
CONV_BUF_TYPE *dst = conv_params->dst;
const int dst_stride = conv_params->dst_stride;
int height = h;
if (w == 4) {
uint8x8_t s0, s1, s2, s3, d01, d23;
uint16x4_t d0, d1, d2, d3, dd0, dd1, dd2, dd3;
do {
load_u8_8x4(src, src_stride, &s0, &s1, &s2, &s3);
d0 = vget_low_u16(vmlal_u8(round_offset_vec, s0, shift_by_bits));
d1 = vget_low_u16(vmlal_u8(round_offset_vec, s1, shift_by_bits));
d2 = vget_low_u16(vmlal_u8(round_offset_vec, s2, shift_by_bits));
d3 = vget_low_u16(vmlal_u8(round_offset_vec, s3, shift_by_bits));
if (do_average) {
load_u16_4x4(dst, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_4x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, vreinterpretq_s16_u16(round_offset_vec),
use_dist_wtd_comp_avg, &d01, &d23);
store_u8_4x1(dst8 + 0 * dst8_stride, d01, 0);
store_u8_4x1(dst8 + 1 * dst8_stride, d01, 1);
store_u8_4x1(dst8 + 2 * dst8_stride, d23, 0);
store_u8_4x1(dst8 + 3 * dst8_stride, d23, 1);
} else {
store_u16_4x4(dst, dst_stride, d0, d1, d2, d3);
}
src += 4 * src_stride;
dst += 4 * dst_stride;
dst8 += 4 * dst8_stride;
height -= 4;
} while (height != 0);
} else {
uint8x8_t s0, s1, s2, s3, d0_u8, d1_u8, d2_u8, d3_u8;
uint16x8_t d0, d1, d2, d3, dd0, dd1, dd2, dd3;
do {
const uint8_t *s = src;
CONV_BUF_TYPE *d = dst;
uint8_t *d_u8 = dst8;
int width = w;
do {
load_u8_8x4(s, src_stride, &s0, &s1, &s2, &s3);
d0 = vmlal_u8(round_offset_vec, s0, shift_by_bits);
d1 = vmlal_u8(round_offset_vec, s1, shift_by_bits);
d2 = vmlal_u8(round_offset_vec, s2, shift_by_bits);
d3 = vmlal_u8(round_offset_vec, s3, shift_by_bits);
if (do_average) {
load_u16_8x4(d, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_8x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, vreinterpretq_s16_u16(round_offset_vec),
use_dist_wtd_comp_avg, &d0_u8, &d1_u8, &d2_u8,
&d3_u8);
store_u8_8x4(d_u8, dst8_stride, d0_u8, d1_u8, d2_u8, d3_u8);
} else {
store_u16_8x4(d, dst_stride, d0, d1, d2, d3);
}
s += 8;
d += 8;
d_u8 += 8;
width -= 8;
} while (width != 0);
src += 4 * src_stride;
dst += 4 * dst_stride;
dst8 += 4 * dst8_stride;
height -= 4;
} while (height != 0);
}
}
#if defined(__aarch64__) && defined(__ARM_FEATURE_MATMUL_INT8)
static INLINE uint16x4_t convolve8_4_x(uint8x16_t samples,
const int8x8_t x_filter,
const uint8x16x2_t permute_tbl,
const int32x4_t round_offset) {
uint8x16_t permuted_samples[2];
int32x4_t sum;
// Permute samples ready for dot product.
// { 0, 1, 2, 3, 1, 2, 3, 4, 2, 3, 4, 5, 3, 4, 5, 6 }
permuted_samples[0] = vqtbl1q_u8(samples, permute_tbl.val[0]);
// { 4, 5, 6, 7, 5, 6, 7, 8, 6, 7, 8, 9, 7, 8, 9, 10 }
permuted_samples[1] = vqtbl1q_u8(samples, permute_tbl.val[1]);
// First 4 output values.
sum = vusdotq_lane_s32(round_offset, permuted_samples[0], x_filter, 0);
sum = vusdotq_lane_s32(sum, permuted_samples[1], x_filter, 1);
// We halved the convolution filter values so -1 from the right shift.
return vreinterpret_u16_s16(vshrn_n_s32(sum, ROUND0_BITS - 1));
}
static INLINE uint16x8_t convolve8_8_x(uint8x16_t samples,
const int8x8_t x_filter,
const uint8x16x3_t permute_tbl,
const int32x4_t round_offset) {
uint8x16_t permuted_samples[3];
int32x4_t sum[2];
// Permute samples ready for dot product.
// { 0, 1, 2, 3, 1, 2, 3, 4, 2, 3, 4, 5, 3, 4, 5, 6 }
permuted_samples[0] = vqtbl1q_u8(samples, permute_tbl.val[0]);
// { 4, 5, 6, 7, 5, 6, 7, 8, 6, 7, 8, 9, 7, 8, 9, 10 }
permuted_samples[1] = vqtbl1q_u8(samples, permute_tbl.val[1]);
// { 8, 9, 10, 11, 9, 10, 11, 12, 10, 11, 12, 13, 11, 12, 13, 14 }
permuted_samples[2] = vqtbl1q_u8(samples, permute_tbl.val[2]);
// First 4 output values.
sum[0] = vusdotq_lane_s32(round_offset, permuted_samples[0], x_filter, 0);
sum[0] = vusdotq_lane_s32(sum[0], permuted_samples[1], x_filter, 1);
// Second 4 output values.
sum[1] = vusdotq_lane_s32(round_offset, permuted_samples[1], x_filter, 0);
sum[1] = vusdotq_lane_s32(sum[1], permuted_samples[2], x_filter, 1);
// Narrow and re-pack.
// We halved the convolution filter values so -1 from the right shift.
int16x8_t res = vcombine_s16(vshrn_n_s32(sum[0], ROUND0_BITS - 1),
vshrn_n_s32(sum[1], ROUND0_BITS - 1));
return vreinterpretq_u16_s16(res);
}
void av1_dist_wtd_convolve_x_neon(const uint8_t *src, int src_stride,
uint8_t *dst8, int dst8_stride, int w, int h,
const InterpFilterParams *filter_params_x,
const int subpel_x_qn,
ConvolveParams *conv_params) {
assert(w % 4 == 0);
assert(h % 4 == 0);
const int bd = 8;
const int offset_bits = bd + 2 * FILTER_BITS - ROUND0_BITS;
const int16_t round_offset = (1 << (offset_bits - COMPOUND_ROUND1_BITS)) +
(1 << (offset_bits - COMPOUND_ROUND1_BITS - 1));
const int16x8_t round_offset_vec = vdupq_n_s16(round_offset);
// A shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-rounding
// shifts - which are generally faster than rounding shifts on modern CPUs.
// (The extra -1 is needed because we halved the filter values.)
const int32x4_t round_offset_shim = vdupq_n_s32(
(round_offset << (ROUND0_BITS - 1)) + (1 << ((ROUND0_BITS - 1) - 1)));
const int do_average = conv_params->do_average;
const int use_dist_wtd_comp_avg = conv_params->use_dist_wtd_comp_avg;
const uint16_t fwd_offset = conv_params->fwd_offset;
const uint16_t bck_offset = conv_params->bck_offset;
// Horizontal filter.
const int16_t *x_filter_ptr = av1_get_interp_filter_subpel_kernel(
filter_params_x, subpel_x_qn & SUBPEL_MASK);
// Filter values are even, so downshift by 1 to reduce intermediate precision
// requirements.
const int8x8_t x_filter = vshrn_n_s16(vld1q_s16(x_filter_ptr), 1);
const int horiz_offset = filter_params_x->taps / 2 - 1;
const uint8_t *src_ptr = src - horiz_offset;
CONV_BUF_TYPE *dst_ptr = conv_params->dst;
uint8_t *dst8_ptr = dst8;
int dst_stride = conv_params->dst_stride;
int height = h;
if (w == 4) {
const uint8x16x2_t permute_tbl = vld1q_u8_x2(dot_prod_permute_tbl);
do {
uint8x16_t s0, s1, s2, s3;
uint16x4_t d0, d1, d2, d3, dd0, dd1, dd2, dd3;
uint8x8_t d01_u8, d23_u8;
load_u8_16x4(src_ptr, src_stride, &s0, &s1, &s2, &s3);
d0 = convolve8_4_x(s0, x_filter, permute_tbl, round_offset_shim);
d1 = convolve8_4_x(s1, x_filter, permute_tbl, round_offset_shim);
d2 = convolve8_4_x(s2, x_filter, permute_tbl, round_offset_shim);
d3 = convolve8_4_x(s3, x_filter, permute_tbl, round_offset_shim);
if (do_average) {
load_u16_4x4(dst_ptr, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_4x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d01_u8, &d23_u8);
store_u8_4x1(dst8_ptr + 0 * dst8_stride, d01_u8, 0);
store_u8_4x1(dst8_ptr + 1 * dst8_stride, d01_u8, 1);
store_u8_4x1(dst8_ptr + 2 * dst8_stride, d23_u8, 0);
store_u8_4x1(dst8_ptr + 3 * dst8_stride, d23_u8, 1);
} else {
store_u16_4x4(dst_ptr, dst_stride, d0, d1, d2, d3);
}
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
dst8_ptr += 4 * dst8_stride;
height -= 4;
} while (height != 0);
} else {
const uint8x16x3_t permute_tbl = vld1q_u8_x3(dot_prod_permute_tbl);
do {
const uint8_t *s = src_ptr;
CONV_BUF_TYPE *d = dst_ptr;
uint8_t *d_u8 = dst8_ptr;
int width = w;
do {
uint8x16_t s0, s1, s2, s3;
uint16x8_t d0, d1, d2, d3, dd0, dd1, dd2, dd3;
uint8x8_t d0_u8, d1_u8, d2_u8, d3_u8;
load_u8_16x4(s, src_stride, &s0, &s1, &s2, &s3);
d0 = convolve8_8_x(s0, x_filter, permute_tbl, round_offset_shim);
d1 = convolve8_8_x(s1, x_filter, permute_tbl, round_offset_shim);
d2 = convolve8_8_x(s2, x_filter, permute_tbl, round_offset_shim);
d3 = convolve8_8_x(s3, x_filter, permute_tbl, round_offset_shim);
if (do_average) {
load_u16_8x4(d, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_8x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d0_u8, &d1_u8, &d2_u8, &d3_u8);
store_u8_8x4(d_u8, dst8_stride, d0_u8, d1_u8, d2_u8, d3_u8);
} else {
store_u16_8x4(d, dst_stride, d0, d1, d2, d3);
}
s += 8;
d += 8;
d_u8 += 8;
width -= 8;
} while (width != 0);
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
dst8_ptr += 4 * dst8_stride;
height -= 4;
} while (height != 0);
}
}
#elif defined(__aarch64__) && defined(__ARM_FEATURE_DOTPROD)
static INLINE uint16x4_t convolve8_4_x(uint8x16_t samples,
const int8x8_t x_filter,
const int32x4_t correction,
const uint8x16_t range_limit,
const uint8x16x2_t permute_tbl) {
int8x16_t clamped_samples, permuted_samples[2];
int32x4_t sum;
// Clamp sample range to [-128, 127] for 8-bit signed dot product.
clamped_samples = vreinterpretq_s8_u8(vsubq_u8(samples, range_limit));
// Permute samples ready for dot product.
// { 0, 1, 2, 3, 1, 2, 3, 4, 2, 3, 4, 5, 3, 4, 5, 6 }
permuted_samples[0] = vqtbl1q_s8(clamped_samples, permute_tbl.val[0]);
// { 4, 5, 6, 7, 5, 6, 7, 8, 6, 7, 8, 9, 7, 8, 9, 10 }
permuted_samples[1] = vqtbl1q_s8(clamped_samples, permute_tbl.val[1]);
// Accumulate dot product into 'correction' to account for range clamp.
sum = vdotq_lane_s32(correction, permuted_samples[0], x_filter, 0);
sum = vdotq_lane_s32(sum, permuted_samples[1], x_filter, 1);
// We halved the convolution filter values so -1 from the right shift.
return vreinterpret_u16_s16(vshrn_n_s32(sum, ROUND0_BITS - 1));
}
static INLINE uint16x8_t convolve8_8_x(uint8x16_t samples,
const int8x8_t x_filter,
const int32x4_t correction,
const uint8x16_t range_limit,
const uint8x16x3_t permute_tbl) {
int8x16_t clamped_samples, permuted_samples[3];
int32x4_t sum[2];
// Clamp sample range to [-128, 127] for 8-bit signed dot product.
clamped_samples = vreinterpretq_s8_u8(vsubq_u8(samples, range_limit));
// Permute samples ready for dot product. */
// { 0, 1, 2, 3, 1, 2, 3, 4, 2, 3, 4, 5, 3, 4, 5, 6 }
permuted_samples[0] = vqtbl1q_s8(clamped_samples, permute_tbl.val[0]);
// { 4, 5, 6, 7, 5, 6, 7, 8, 6, 7, 8, 9, 7, 8, 9, 10 }
permuted_samples[1] = vqtbl1q_s8(clamped_samples, permute_tbl.val[1]);
// { 8, 9, 10, 11, 9, 10, 11, 12, 10, 11, 12, 13, 11, 12, 13, 14 }
permuted_samples[2] = vqtbl1q_s8(clamped_samples, permute_tbl.val[2]);
// Accumulate dot product into 'correction' to account for range clamp.
// First 4 output values.
sum[0] = vdotq_lane_s32(correction, permuted_samples[0], x_filter, 0);
sum[0] = vdotq_lane_s32(sum[0], permuted_samples[1], x_filter, 1);
// Second 4 output values.
sum[1] = vdotq_lane_s32(correction, permuted_samples[1], x_filter, 0);
sum[1] = vdotq_lane_s32(sum[1], permuted_samples[2], x_filter, 1);
// Narrow and re-pack.
// We halved the convolution filter values so -1 from the right shift.
int16x8_t res = vcombine_s16(vshrn_n_s32(sum[0], ROUND0_BITS - 1),
vshrn_n_s32(sum[1], ROUND0_BITS - 1));
return vreinterpretq_u16_s16(res);
}
void av1_dist_wtd_convolve_x_neon(const uint8_t *src, int src_stride,
uint8_t *dst8, int dst8_stride, int w, int h,
const InterpFilterParams *filter_params_x,
const int subpel_x_qn,
ConvolveParams *conv_params) {
assert(w % 4 == 0);
assert(h % 4 == 0);
const int bd = 8;
const int offset_bits = bd + 2 * FILTER_BITS - ROUND0_BITS;
const int16_t round_offset = (1 << (offset_bits - COMPOUND_ROUND1_BITS)) +
(1 << (offset_bits - COMPOUND_ROUND1_BITS - 1));
const int16x8_t round_offset_vec = vdupq_n_s16(round_offset);
const int do_average = conv_params->do_average;
const int use_dist_wtd_comp_avg = conv_params->use_dist_wtd_comp_avg;
const uint16_t fwd_offset = conv_params->fwd_offset;
const uint16_t bck_offset = conv_params->bck_offset;
// Horizontal filter.
const int16_t *x_filter_ptr = av1_get_interp_filter_subpel_kernel(
filter_params_x, subpel_x_qn & SUBPEL_MASK);
// Filter values are even, so downshift by 1 to reduce intermediate precision
// requirements.
const int8x8_t x_filter = vshrn_n_s16(vld1q_s16(x_filter_ptr), 1);
// Dot-product constants and other shims.
const uint8x16_t range_limit = vdupq_n_u8(128);
const int32_t correction_s32 = vaddlvq_s16(vshll_n_s8(x_filter, 7));
// Fold round_offset into the dot-product filter correction constant. The
// additional shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-
// rounding shifts - which are generally faster than rounding shifts on
// modern CPUs. (The extra -1 is needed because we halved the filter values.)
int32x4_t correction =
vdupq_n_s32(correction_s32 + (round_offset << (ROUND0_BITS - 1)) +
(1 << ((ROUND0_BITS - 1) - 1)));
const int horiz_offset = filter_params_x->taps / 2 - 1;
const uint8_t *src_ptr = src - horiz_offset;
CONV_BUF_TYPE *dst_ptr = conv_params->dst;
uint8_t *dst8_ptr = dst8;
int dst_stride = conv_params->dst_stride;
int height = h;
if (w == 4) {
const uint8x16x2_t permute_tbl = vld1q_u8_x2(dot_prod_permute_tbl);
do {
uint8x16_t s0, s1, s2, s3;
uint16x4_t d0, d1, d2, d3, dd0, dd1, dd2, dd3;
uint8x8_t d01_u8, d23_u8;
load_u8_16x4(src_ptr, src_stride, &s0, &s1, &s2, &s3);
d0 = convolve8_4_x(s0, x_filter, correction, range_limit, permute_tbl);
d1 = convolve8_4_x(s1, x_filter, correction, range_limit, permute_tbl);
d2 = convolve8_4_x(s2, x_filter, correction, range_limit, permute_tbl);
d3 = convolve8_4_x(s3, x_filter, correction, range_limit, permute_tbl);
if (do_average) {
load_u16_4x4(dst_ptr, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_4x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d01_u8, &d23_u8);
store_u8_4x1(dst8_ptr + 0 * dst8_stride, d01_u8, 0);
store_u8_4x1(dst8_ptr + 1 * dst8_stride, d01_u8, 1);
store_u8_4x1(dst8_ptr + 2 * dst8_stride, d23_u8, 0);
store_u8_4x1(dst8_ptr + 3 * dst8_stride, d23_u8, 1);
} else {
store_u16_4x4(dst_ptr, dst_stride, d0, d1, d2, d3);
}
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
dst8_ptr += 4 * dst8_stride;
height -= 4;
} while (height != 0);
} else {
const uint8x16x3_t permute_tbl = vld1q_u8_x3(dot_prod_permute_tbl);
do {
const uint8_t *s = src_ptr;
CONV_BUF_TYPE *d = dst_ptr;
uint8_t *d_u8 = dst8_ptr;
int width = w;
do {
uint8x16_t s0, s1, s2, s3;
uint16x8_t d0, d1, d2, d3, dd0, dd1, dd2, dd3;
uint8x8_t d0_u8, d1_u8, d2_u8, d3_u8;
load_u8_16x4(s, src_stride, &s0, &s1, &s2, &s3);
d0 = convolve8_8_x(s0, x_filter, correction, range_limit, permute_tbl);
d1 = convolve8_8_x(s1, x_filter, correction, range_limit, permute_tbl);
d2 = convolve8_8_x(s2, x_filter, correction, range_limit, permute_tbl);
d3 = convolve8_8_x(s3, x_filter, correction, range_limit, permute_tbl);
if (do_average) {
load_u16_8x4(d, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_8x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d0_u8, &d1_u8, &d2_u8, &d3_u8);
store_u8_8x4(d_u8, dst8_stride, d0_u8, d1_u8, d2_u8, d3_u8);
} else {
store_u16_8x4(d, dst_stride, d0, d1, d2, d3);
}
s += 8;
d += 8;
d_u8 += 8;
width -= 8;
} while (width != 0);
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
dst8_ptr += 4 * dst8_stride;
height -= 4;
} while (height != 0);
}
}
#else // !(defined(__aarch64__) && defined(__ARM_FEATURE_DOTPROD))
static INLINE uint16x4_t convolve8_4_x(const int16x4_t s0, const int16x4_t s1,
const int16x4_t s2, const int16x4_t s3,
const int16x4_t s4, const int16x4_t s5,
const int16x4_t s6, const int16x4_t s7,
const int16x8_t x_filter,
const int16x4_t round_offset) {
const int16x4_t x_filter_0_3 = vget_low_s16(x_filter);
const int16x4_t x_filter_4_7 = vget_high_s16(x_filter);
int16x4_t sum = vmul_lane_s16(s0, x_filter_0_3, 0);
sum = vmla_lane_s16(sum, s1, x_filter_0_3, 1);
sum = vmla_lane_s16(sum, s2, x_filter_0_3, 2);
sum = vmla_lane_s16(sum, s3, x_filter_0_3, 3);
sum = vmla_lane_s16(sum, s4, x_filter_4_7, 0);
sum = vmla_lane_s16(sum, s5, x_filter_4_7, 1);
sum = vmla_lane_s16(sum, s6, x_filter_4_7, 2);
sum = vmla_lane_s16(sum, s7, x_filter_4_7, 3);
// We halved the convolution filter values so -1 from the right shift.
int16x4_t res = vrsra_n_s16(round_offset, sum, ROUND0_BITS - 1);
return vreinterpret_u16_s16(res);
}
static INLINE uint16x8_t convolve8_8_x(const int16x8_t s0, const int16x8_t s1,
const int16x8_t s2, const int16x8_t s3,
const int16x8_t s4, const int16x8_t s5,
const int16x8_t s6, const int16x8_t s7,
const int16x8_t x_filter,
const int16x8_t round_offset) {
const int16x4_t x_filter_0_3 = vget_low_s16(x_filter);
const int16x4_t x_filter_4_7 = vget_high_s16(x_filter);
int16x8_t sum = vmulq_lane_s16(s0, x_filter_0_3, 0);
sum = vmlaq_lane_s16(sum, s1, x_filter_0_3, 1);
sum = vmlaq_lane_s16(sum, s2, x_filter_0_3, 2);
sum = vmlaq_lane_s16(sum, s3, x_filter_0_3, 3);
sum = vmlaq_lane_s16(sum, s4, x_filter_4_7, 0);
sum = vmlaq_lane_s16(sum, s5, x_filter_4_7, 1);
sum = vmlaq_lane_s16(sum, s6, x_filter_4_7, 2);
sum = vmlaq_lane_s16(sum, s7, x_filter_4_7, 3);
// We halved the convolution filter values so -1 from the right shift.
int16x8_t res = vrsraq_n_s16(round_offset, sum, ROUND0_BITS - 1);
return vreinterpretq_u16_s16(res);
}
void av1_dist_wtd_convolve_x_neon(const uint8_t *src, int src_stride,
uint8_t *dst8, int dst8_stride, int w, int h,
const InterpFilterParams *filter_params_x,
const int subpel_x_qn,
ConvolveParams *conv_params) {
assert(w % 4 == 0);
assert(h % 4 == 0);
const int bd = 8;
const int offset_bits = bd + 2 * FILTER_BITS - ROUND0_BITS;
const int16_t round_offset = (1 << (offset_bits - COMPOUND_ROUND1_BITS)) +
(1 << (offset_bits - COMPOUND_ROUND1_BITS - 1));
const int16x8_t round_offset_vec = vdupq_n_s16(round_offset);
const int do_average = conv_params->do_average;
const int use_dist_wtd_comp_avg = conv_params->use_dist_wtd_comp_avg;
const uint16_t fwd_offset = conv_params->fwd_offset;
const uint16_t bck_offset = conv_params->bck_offset;
// Horizontal filter.
const int16_t *x_filter_ptr = av1_get_interp_filter_subpel_kernel(
filter_params_x, subpel_x_qn & SUBPEL_MASK);
// Filter values are even, so downshift by 1 to reduce intermediate precision
// requirements.
const int16x8_t x_filter = vshrq_n_s16(vld1q_s16(x_filter_ptr), 1);
const int horiz_offset = filter_params_x->taps / 2 - 1;
const uint8_t *src_ptr = src - horiz_offset;
CONV_BUF_TYPE *dst_ptr = conv_params->dst;
uint8_t *dst8_ptr = dst8;
int dst_stride = conv_params->dst_stride;
const uint8_t *s;
uint8_t *d_u8;
CONV_BUF_TYPE *d;
int width;
int height = h;
uint8x8_t t0;
#if defined(__aarch64__)
uint8x8_t t1, t2, t3, t4, t5, t6, t7;
#endif // defined(__aarch64__)
if ((w == 4) || (h == 4)) {
int16x4_t s0, s1, s2, s3, s4, s5, s6, s7, s8;
uint16x4_t d0, dd0;
uint8x8_t d01;
#if defined(__aarch64__)
int16x4_t s9, s10;
uint16x4_t d1, d2, d3, dd1, dd2, dd3;
uint8x8_t d23;
#endif // defined(__aarch64__)
do {
d = dst_ptr;
d_u8 = dst8_ptr;
width = w;
__builtin_prefetch(src_ptr + 0 * src_stride);
#if defined(__aarch64__)
__builtin_prefetch(src_ptr + 1 * src_stride);
__builtin_prefetch(src_ptr + 2 * src_stride);
__builtin_prefetch(src_ptr + 3 * src_stride);
load_u8_8x4(src_ptr, src_stride, &t0, &t1, &t2, &t3);
transpose_u8_8x4(&t0, &t1, &t2, &t3);
s0 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s1 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
s2 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t2)));
s3 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t3)));
s4 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s5 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
s6 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t2)));
__builtin_prefetch(d + 0 * dst_stride);
__builtin_prefetch(d + 1 * dst_stride);
__builtin_prefetch(d + 2 * dst_stride);
__builtin_prefetch(d + 3 * dst_stride);
s = src_ptr + 7;
do {
load_unaligned_u8_4x4(s, src_stride, &t0, &t1);
transpose_u8_4x4(&t0, &t1);
s7 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s8 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
s9 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s10 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
d0 = convolve8_4_x(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
vget_low_s16(round_offset_vec));
d1 = convolve8_4_x(s1, s2, s3, s4, s5, s6, s7, s8, x_filter,
vget_low_s16(round_offset_vec));
d2 = convolve8_4_x(s2, s3, s4, s5, s6, s7, s8, s9, x_filter,
vget_low_s16(round_offset_vec));
d3 = convolve8_4_x(s3, s4, s5, s6, s7, s8, s9, s10, x_filter,
vget_low_s16(round_offset_vec));
transpose_u16_4x4d(&d0, &d1, &d2, &d3);
if (do_average) {
__builtin_prefetch(d + 0 * dst_stride);
__builtin_prefetch(d + 1 * dst_stride);
__builtin_prefetch(d + 2 * dst_stride);
__builtin_prefetch(d + 3 * dst_stride);
__builtin_prefetch(d_u8 + 0 * dst8_stride);
__builtin_prefetch(d_u8 + 1 * dst8_stride);
__builtin_prefetch(d_u8 + 2 * dst8_stride);
__builtin_prefetch(d_u8 + 3 * dst8_stride);
load_u16_4x4(d, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_4x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d01, &d23);
store_u8_4x1(d_u8 + 0 * dst8_stride, d01, 0);
store_u8_4x1(d_u8 + 1 * dst8_stride, d01, 1);
store_u8_4x1(d_u8 + 2 * dst8_stride, d23, 0);
store_u8_4x1(d_u8 + 3 * dst8_stride, d23, 1);
} else {
store_u16_4x4(d, dst_stride, d0, d1, d2, d3);
}
s0 = s4;
s1 = s5;
s2 = s6;
s3 = s7;
s4 = s8;
s5 = s9;
s6 = s10;
s += 4;
d += 4;
d_u8 += 4;
width -= 4;
} while (width != 0);
src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
dst8_ptr += 4 * dst8_stride;
height -= 4;
#else // !defined(__aarch64__)
t0 = vld1_u8(src_ptr); // a0 a1 a2 a3 a4 a5 a6 a7
s0 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s4 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
__builtin_prefetch(d);
s = src_ptr + 8;
do {
t0 = vld1_u8(s); // a8 a9 a10 a11
s8 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s1 = vext_s16(s0, s4, 1); // a1 a2 a3 a4
s2 = vext_s16(s0, s4, 2); // a2 a3 a4 a5
s3 = vext_s16(s0, s4, 3); // a3 a4 a5 a6
s5 = vext_s16(s4, s8, 1); // a5 a6 a7 a8
s6 = vext_s16(s4, s8, 2); // a6 a7 a8 a9
s7 = vext_s16(s4, s8, 3); // a7 a8 a9 a10
d0 = convolve8_4_x(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
vget_low_s16(round_offset_vec));
if (do_average) {
__builtin_prefetch(d);
__builtin_prefetch(d_u8);
dd0 = vld1_u16(d);
compute_avg_4x1(dd0, d0, fwd_offset, bck_offset,
vget_low_s16(round_offset_vec), use_dist_wtd_comp_avg,
&d01);
store_u8_4x1(d_u8, d01, 0);
} else {
vst1_u16(d, d0);
}
s0 = s4;
s4 = s8;
s += 4;
d += 4;
d_u8 += 4;
width -= 4;
} while (width != 0);
src_ptr += src_stride;
dst_ptr += dst_stride;
dst8_ptr += dst8_stride;
height--;
#endif // defined(__aarch64__)
} while (height != 0);
} else {
int16x8_t s0, s1, s2, s3, s4, s5, s6, s7, s8;
uint16x8_t d0, dd0;
uint8x8_t d0_u8;
do {
d = dst_ptr;
d_u8 = dst8_ptr;
width = w;
#if defined(__aarch64__)
int16x8_t s9, s10, s11, s12, s13, s14;
uint16x8_t d1, d2, d3, d4, d5, d6, d7, dd1, dd2, dd3, dd4, dd5, dd6, dd7;
uint8x8_t d1_u8, d2_u8, d3_u8, d4_u8, d5_u8, d6_u8, d7_u8;
__builtin_prefetch(src_ptr + 0 * src_stride);
__builtin_prefetch(src_ptr + 1 * src_stride);
__builtin_prefetch(src_ptr + 2 * src_stride);
__builtin_prefetch(src_ptr + 3 * src_stride);
__builtin_prefetch(src_ptr + 4 * src_stride);
__builtin_prefetch(src_ptr + 5 * src_stride);
__builtin_prefetch(src_ptr + 6 * src_stride);
__builtin_prefetch(src_ptr + 7 * src_stride);
load_u8_8x8(src_ptr, src_stride, &t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
transpose_u8_8x8(&t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
s0 = vreinterpretq_s16_u16(vmovl_u8(t0));
s1 = vreinterpretq_s16_u16(vmovl_u8(t1));
s2 = vreinterpretq_s16_u16(vmovl_u8(t2));
s3 = vreinterpretq_s16_u16(vmovl_u8(t3));
s4 = vreinterpretq_s16_u16(vmovl_u8(t4));
s5 = vreinterpretq_s16_u16(vmovl_u8(t5));
s6 = vreinterpretq_s16_u16(vmovl_u8(t6));
__builtin_prefetch(dst_ptr + 0 * dst_stride);
__builtin_prefetch(dst_ptr + 1 * dst_stride);
__builtin_prefetch(dst_ptr + 2 * dst_stride);
__builtin_prefetch(dst_ptr + 3 * dst_stride);
__builtin_prefetch(dst_ptr + 4 * dst_stride);
__builtin_prefetch(dst_ptr + 5 * dst_stride);
__builtin_prefetch(dst_ptr + 6 * dst_stride);
__builtin_prefetch(dst_ptr + 7 * dst_stride);
s = src_ptr + 7;
do {
load_u8_8x8(s, src_stride, &t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
transpose_u8_8x8(&t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
s7 = vreinterpretq_s16_u16(vmovl_u8(t0));
s8 = vreinterpretq_s16_u16(vmovl_u8(t1));
s9 = vreinterpretq_s16_u16(vmovl_u8(t2));
s10 = vreinterpretq_s16_u16(vmovl_u8(t3));
s11 = vreinterpretq_s16_u16(vmovl_u8(t4));
s12 = vreinterpretq_s16_u16(vmovl_u8(t5));
s13 = vreinterpretq_s16_u16(vmovl_u8(t6));
s14 = vreinterpretq_s16_u16(vmovl_u8(t7));
d0 = convolve8_8_x(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
round_offset_vec);
d1 = convolve8_8_x(s1, s2, s3, s4, s5, s6, s7, s8, x_filter,
round_offset_vec);
d2 = convolve8_8_x(s2, s3, s4, s5, s6, s7, s8, s9, x_filter,
round_offset_vec);
d3 = convolve8_8_x(s3, s4, s5, s6, s7, s8, s9, s10, x_filter,
round_offset_vec);
d4 = convolve8_8_x(s4, s5, s6, s7, s8, s9, s10, s11, x_filter,
round_offset_vec);
d5 = convolve8_8_x(s5, s6, s7, s8, s9, s10, s11, s12, x_filter,
round_offset_vec);
d6 = convolve8_8_x(s6, s7, s8, s9, s10, s11, s12, s13, x_filter,
round_offset_vec);
d7 = convolve8_8_x(s7, s8, s9, s10, s11, s12, s13, s14, x_filter,
round_offset_vec);
transpose_u16_8x8(&d0, &d1, &d2, &d3, &d4, &d5, &d6, &d7);
if (conv_params->do_average) {
load_u16_8x4(d, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_8x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d0_u8, &d1_u8, &d2_u8, &d3_u8);
store_u8_8x4(d_u8, dst8_stride, d0_u8, d1_u8, d2_u8, d3_u8);
load_u16_8x4(d + 4 * dst_stride, dst_stride, &dd4, &dd5, &dd6, &dd7);
compute_avg_8x4(dd4, dd5, dd6, dd7, d4, d5, d6, d7, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d4_u8, &d5_u8, &d6_u8, &d7_u8);
store_u8_8x4(d_u8 + 4 * dst8_stride, dst8_stride, d4_u8, d5_u8, d6_u8,
d7_u8);
} else {
store_u16_8x8(d, dst_stride, d0, d1, d2, d3, d4, d5, d6, d7);
}
s0 = s8;
s1 = s9;
s2 = s10;
s3 = s11;
s4 = s12;
s5 = s13;
s6 = s14;
s += 8;
d += 8;
d_u8 += 8;
width -= 8;
} while (width != 0);
src_ptr += 8 * src_stride;
dst_ptr += 8 * dst_stride;
dst8_ptr += 8 * dst8_stride;
height -= 8;
#else // !defined(__aarch64__)
__builtin_prefetch(src_ptr);
t0 = vld1_u8(src_ptr);
s0 = vreinterpretq_s16_u16(vmovl_u8(t0)); // a0 a1 a2 a3 a4 a5 a6 a7
__builtin_prefetch(dst_ptr);
s = src_ptr + 8;
do {
t0 = vld1_u8(s); // a8 a9 a10 a11 a12 a13 a14 a15
s8 = vreinterpretq_s16_u16(vmovl_u8(t0));
s1 = vextq_s16(s0, s8, 1); // a1 a2 a3 a4 a5 a6 a7 a8
s2 = vextq_s16(s0, s8, 2); // a2 a3 a4 a5 a6 a7 a8 a9
s3 = vextq_s16(s0, s8, 3); // a3 a4 a5 a6 a7 a8 a9 a10
s4 = vextq_s16(s0, s8, 4); // a4 a5 a6 a7 a8 a9 a10 a11
s5 = vextq_s16(s0, s8, 5); // a5 a6 a7 a8 a9 a10 a11 a12
s6 = vextq_s16(s0, s8, 6); // a6 a7 a8 a9 a10 a11 a12 a13
s7 = vextq_s16(s0, s8, 7); // a7 a8 a9 a10 a11 a12 a13 a14
d0 = convolve8_8_x(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
round_offset_vec);
if (do_average) {
dd0 = vld1q_u16(d);
compute_avg_8x1(dd0, d0, fwd_offset, bck_offset, round_offset_vec,
use_dist_wtd_comp_avg, &d0_u8);
vst1_u8(d_u8, d0_u8);
} else {
vst1q_u16(d, d0);
}
s0 = s8;
s += 8;
d += 8;
d_u8 += 8;
width -= 8;
} while (width != 0);
src_ptr += src_stride;
dst_ptr += dst_stride;
dst8_ptr += dst8_stride;
height--;
#endif // defined(__aarch64__)
} while (height != 0);
}
}
#endif // defined(__aarch64__) && defined(__ARM_FEATURE_DOTPROD)
static INLINE uint16x4_t convolve6_4_y(const int16x4_t s0, const int16x4_t s1,
const int16x4_t s2, const int16x4_t s3,
const int16x4_t s4, const int16x4_t s5,
const int16x8_t y_filter,
const int16x4_t round_offset) {
const int16x4_t y_filter_0_3 = vget_low_s16(y_filter);
const int16x4_t y_filter_4_7 = vget_high_s16(y_filter);
// Filter values at indices 0 and 7 are 0.
int16x4_t sum = vmul_lane_s16(s0, y_filter_0_3, 1);
sum = vmla_lane_s16(sum, s1, y_filter_0_3, 2);
sum = vmla_lane_s16(sum, s2, y_filter_0_3, 3);
sum = vmla_lane_s16(sum, s3, y_filter_4_7, 0);
sum = vmla_lane_s16(sum, s4, y_filter_4_7, 1);
sum = vmla_lane_s16(sum, s5, y_filter_4_7, 2);
// We halved the convolution filter values so -1 from the right shift.
int16x4_t res = vrsra_n_s16(round_offset, sum, ROUND0_BITS - 1);
return vreinterpret_u16_s16(res);
}
static INLINE uint16x8_t convolve6_8_y(const int16x8_t s0, const int16x8_t s1,
const int16x8_t s2, const int16x8_t s3,
const int16x8_t s4, const int16x8_t s5,
const int16x8_t y_filter,
const int16x8_t round_offset) {
const int16x4_t y_filter_0_3 = vget_low_s16(y_filter);
const int16x4_t y_filter_4_7 = vget_high_s16(y_filter);
// Filter values at indices 0 and 7 are 0.
int16x8_t sum = vmulq_lane_s16(s0, y_filter_0_3, 1);
sum = vmlaq_lane_s16(sum, s1, y_filter_0_3, 2);
sum = vmlaq_lane_s16(sum, s2, y_filter_0_3, 3);
sum = vmlaq_lane_s16(sum, s3, y_filter_4_7, 0);
sum = vmlaq_lane_s16(sum, s4, y_filter_4_7, 1);
sum = vmlaq_lane_s16(sum, s5, y_filter_4_7, 2);
// We halved the convolution filter values so -1 from the right shift.
int16x8_t res = vrsraq_n_s16(round_offset, sum, ROUND0_BITS - 1);
return vreinterpretq_u16_s16(res);
}
static INLINE void dist_wtd_convolve_y_6tap_neon(
const uint8_t *src_ptr, int src_stride, uint8_t *dst8_ptr,
const int dst8_stride, int w, int h, const int16x8_t y_filter,
ConvolveParams *conv_params) {
const int bd = 8;
const int offset_bits = bd + 2 * FILTER_BITS - ROUND0_BITS;
const int16_t round_offset = (1 << (offset_bits - COMPOUND_ROUND1_BITS)) +
(1 << (offset_bits - COMPOUND_ROUND1_BITS - 1));
const int16x8_t round_offset_vec = vdupq_n_s16(round_offset);
const int do_average = conv_params->do_average;
const int use_dist_wtd_comp_avg = conv_params->use_dist_wtd_comp_avg;
const uint16_t fwd_offset = conv_params->fwd_offset;
const uint16_t bck_offset = conv_params->bck_offset;
CONV_BUF_TYPE *dst_ptr = conv_params->dst;
const int dst_stride = conv_params->dst_stride;
int width = w;
if (w == 4 || h == 4) {
int16x4_t s0, s1, s2, s3, s4, s5;
uint16x4_t d0, dd0;
uint8x8_t t0, t1, t2, t3, t4, d01;
#if defined(__aarch64__)
int16x4_t s6, s7, s8;
uint16x4_t d1, d2, d3, dd1, dd2, dd3;
uint8x8_t d23;
#endif // defined(__aarch64__)
do {
const uint8_t *s = src_ptr;
CONV_BUF_TYPE *d = dst_ptr;
uint8_t *d_u8 = dst8_ptr;
int height = h;
t0 = load_unaligned_u8_4x1(s + 0 * src_stride);
t1 = load_unaligned_u8_4x1(s + 1 * src_stride);
t2 = load_unaligned_u8_4x1(s + 2 * src_stride);
t3 = load_unaligned_u8_4x1(s + 3 * src_stride);
t4 = load_unaligned_u8_4x1(s + 4 * src_stride);
s0 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t0)));
s1 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t1)));
s2 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t2)));
s3 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t3)));
s4 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t4)));
s += 5 * src_stride;
do {
#if defined(__aarch64__)
t0 = load_unaligned_u8_4x1(s + 0 * src_stride);
t1 = load_unaligned_u8_4x1(s + 1 * src_stride);
t2 = load_unaligned_u8_4x1(s + 2 * src_stride);
t3 = load_unaligned_u8_4x1(s + 3 * src_stride);
s5 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t0)));
s6 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t1)));
s7 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t2)));
s8 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t3)));
d0 = convolve6_4_y(s0, s1, s2, s3, s4, s5, y_filter,
vget_low_s16(round_offset_vec));
d1 = convolve6_4_y(s1, s2, s3, s4, s5, s6, y_filter,
vget_low_s16(round_offset_vec));
d2 = convolve6_4_y(s2, s3, s4, s5, s6, s7, y_filter,
vget_low_s16(round_offset_vec));
d3 = convolve6_4_y(s3, s4, s5, s6, s7, s8, y_filter,
vget_low_s16(round_offset_vec));
if (do_average) {
load_u16_4x4(d, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_4x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d01, &d23);
store_u8_4x1(d_u8 + 0 * dst8_stride, d01, 0);
store_u8_4x1(d_u8 + 1 * dst8_stride, d01, 1);
store_u8_4x1(d_u8 + 2 * dst8_stride, d23, 0);
store_u8_4x1(d_u8 + 3 * dst8_stride, d23, 1);
} else {
store_u16_4x4(d, dst_stride, d0, d1, d2, d3);
}
s0 = s4;
s1 = s5;
s2 = s6;
s3 = s7;
s4 = s8;
s += 4 * src_stride;
d += 4 * dst_stride;
d_u8 += 4 * dst8_stride;
height -= 4;
#else // !defined(__aarch64__)
t0 = load_unaligned_u8_4x1(s);
s5 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t0)));
d0 = convolve6_4_y(s0, s1, s2, s3, s4, s5, y_filter,
vget_low_s16(round_offset_vec));
if (do_average) {
dd0 = vld1_u16(d);
compute_avg_4x1(dd0, d0, fwd_offset, bck_offset,
vget_low_s16(round_offset_vec), use_dist_wtd_comp_avg,
&d01);
store_u8_4x1(d_u8, d01, 0);
} else {
vst1_u16(d, d0);
}
s0 = s1;
s1 = s2;
s2 = s3;
s3 = s4;
s4 = s5;
s += src_stride;
d += dst_stride;
d_u8 += dst8_stride;
height--;
#endif // defined(__aarch64__)
} while (height != 0);
src_ptr += 4;
dst_ptr += 4;
dst8_ptr += 4;
width -= 4;
} while (width != 0);
} else {
int16x8_t s0, s1, s2, s3, s4, s5;
uint16x8_t d0, dd0;
uint8x8_t d0_u8, t0, t1, t2, t3, t4;
#if defined(__aarch64__)
int16x8_t s6, s7, s8, s9, s10, s11, s12;
uint16x8_t d1, d2, d3, d4, d5, d6, d7, dd1, dd2, dd3, dd4, dd5, dd6, dd7;
uint8x8_t d1_u8, d2_u8, d3_u8, d4_u8, d5_u8, d6_u8, d7_u8, t5, t6, t7;
#endif // defined(__aarch64__)
do {
const uint8_t *s = src_ptr + (5 * src_stride);
CONV_BUF_TYPE *d = dst_ptr;
uint8_t *d_u8 = dst8_ptr;
int height = h;
load_u8_8x5(src_ptr, src_stride, &t0, &t1, &t2, &t3, &t4);
s0 = vreinterpretq_s16_u16(vmovl_u8(t0));
s1 = vreinterpretq_s16_u16(vmovl_u8(t1));
s2 = vreinterpretq_s16_u16(vmovl_u8(t2));
s3 = vreinterpretq_s16_u16(vmovl_u8(t3));
s4 = vreinterpretq_s16_u16(vmovl_u8(t4));
do {
#if defined(__aarch64__)
load_u8_8x8(s, src_stride, &t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
s5 = vreinterpretq_s16_u16(vmovl_u8(t0));
s6 = vreinterpretq_s16_u16(vmovl_u8(t1));
s7 = vreinterpretq_s16_u16(vmovl_u8(t2));
s8 = vreinterpretq_s16_u16(vmovl_u8(t3));
s9 = vreinterpretq_s16_u16(vmovl_u8(t4));
s10 = vreinterpretq_s16_u16(vmovl_u8(t5));
s11 = vreinterpretq_s16_u16(vmovl_u8(t6));
s12 = vreinterpretq_s16_u16(vmovl_u8(t7));
d0 = convolve6_8_y(s0, s1, s2, s3, s4, s5, y_filter, round_offset_vec);
d1 = convolve6_8_y(s1, s2, s3, s4, s5, s6, y_filter, round_offset_vec);
d2 = convolve6_8_y(s2, s3, s4, s5, s6, s7, y_filter, round_offset_vec);
d3 = convolve6_8_y(s3, s4, s5, s6, s7, s8, y_filter, round_offset_vec);
d4 = convolve6_8_y(s4, s5, s6, s7, s8, s9, y_filter, round_offset_vec);
d5 = convolve6_8_y(s5, s6, s7, s8, s9, s10, y_filter, round_offset_vec);
d6 =
convolve6_8_y(s6, s7, s8, s9, s10, s11, y_filter, round_offset_vec);
d7 = convolve6_8_y(s7, s8, s9, s10, s11, s12, y_filter,
round_offset_vec);
if (do_average) {
load_u16_8x4(d, dst_stride, &dd0, &dd1, &dd2, &dd3);
compute_avg_8x4(dd0, dd1, dd2, dd3, d0, d1, d2, d3, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d0_u8, &d1_u8, &d2_u8, &d3_u8);
store_u8_8x4(d_u8, dst8_stride, d0_u8, d1_u8, d2_u8, d3_u8);
d_u8 += 4 * dst8_stride;
load_u16_8x4(d + 4 * dst_stride, dst_stride, &dd4, &dd5, &dd6, &dd7);
compute_avg_8x4(dd4, dd5, dd6, dd7, d4, d5, d6, d7, fwd_offset,
bck_offset, round_offset_vec, use_dist_wtd_comp_avg,
&d4_u8, &d5_u8, &d6_u8, &d7_u8);
store_u8_8x4(d_u8, dst8_stride, d4_u8, d5_u8, d6_u8, d7_u8);
d_u8 += 4 * dst8_stride;
} else {
store_u16_8x8(d, dst_stride, d0, d1, d2, d3, d4, d5, d6, d7);
}
s0 = s8;
s1 = s9;
s2 = s10;
s3 = s11;
s4 = s12;
s += 8 * src_stride;
d += 8 * dst_stride;
height -= 8;
#else // !defined(__aarch64__)
s5 = vreinterpretq_s16_u16(vmovl_u8(vld1_u8(s)));
d0 = convolve6_8_y(s0, s1, s2, s3, s4, s5, y_filter, round_offset_vec);
s0 = s1;
s1 = s2;
s2 = s3;
s3 = s4;
s4 = s5;
if (do_average) {
dd0 = vld1q_u16(d);
compute_avg_8x1(dd0, d0, fwd_offset, bck_offset, round_offset_vec,
use_dist_wtd_comp_avg, &d0_u8);
vst1_u8(d_u8, d0_u8);
d_u8 += dst8_stride;
} else {
vst1q_u16(d, d0);
}
s += src_stride;
d += dst_stride;
height--;
#endif // defined(__aarch64__)
} while (height != 0);
src_ptr += 8;
dst_ptr += 8;
dst8_ptr += 8;
width -= 8;
} while (width != 0);
}
}
static INLINE uint16x4_t convolve8_4_y(const int16x4_t s0, const int16x4_t s1,
const int16x4_t s2, const int16x4_t s3,
const int16x4_t s4, const int16x4_t s5,
const int16x4_t s6, const int16x4_t s7,
const int16x8_t y_filter,
const int16x4_t round_offset) {
const int16x4_t y_filter_0_3 = vget_low_s16(y_filter);
const int16x4_t y_filter_4_7 = vget_high_s16(y_filter);
int16x4_t sum = vmul_lane_s16(s0, y_filter_0_3, 0);
sum = vmla_lane_s16(sum, s1, y_filter_0_3, 1);
sum = vmla_lane_s16(sum, s2, y_filter_0_3, 2);
sum = vmla_lane_s16(sum, s3, y_filter_0_3, 3);
sum = vmla_lane_s16(sum, s4, y_filter_4_7, 0);
sum = vmla_lane_s16(sum, s5, y_filter_4_7, 1);
sum = vmla_lane_s16(sum, s6, y_filter_4_7, 2);
sum = vmla_lane_s16(sum, s7, y_filter_4_7, 3);
// We halved the convolution filter values so -1 from the right shift.
int16x4_t res = vrsra_n_s16(round_offset, sum, ROUND0_BITS - 1);
return vreinterpret_u16_s16(res);
}
static INLINE uint16x8_t convolve8_8_y(const int16x8_t s0, const int16x8_t s1,
const int16x8_t s2, const int16x8_t s3,
const int16x8_t s4, const int16x8_t s5,
const int16x8_t s6, const int16x8_t s7,
const int16x8_t y_filter,
const int16x8_t round_offset) {
const int16x4_t y_filter_0_3 = vget_low_s16(y_filter);
const int16x4_t y_filter_4_7 = vget_high_s16(y_filter);
int16x8_t sum = vmulq_lane_s16(s0, y_filter_0_3, 0);
sum = vmlaq_lane_s16(sum, s1, y_filter_0_3, 1);
sum = vmlaq_lane_s16(sum, s2, y_filter_0_3, 2);
sum = vmlaq_lane_s16(sum, s3, y_filter_0_3, 3);
sum = vmlaq_lane_s16(sum, s4, y_filter_4_7, 0);
sum = vmlaq_lane_s16(sum, s5, y_filter_4_7, 1);
sum = vmlaq_lane_s16(sum, s6, y_filter_4_7, 2);
sum = vmlaq_lane_s16(sum, s7, y_filter_4_7, 3);
// We halved the convolution filter values so -1 from the right shift.
int16x8_t res = vrsraq_n_s16(round_offset, sum, ROUND0_BITS - 1);
return vreinterpretq_u16_s16(res);
}
static INLINE void dist_wtd_convolve_y_8tap_neon(
const uint8_t *src_ptr, int src_stride, uint8_t *dst8_ptr,
const int dst8_stride, int w, int h, const int16x8_t y_filter,
ConvolveParams *conv_params) {
const int bd = 8;
const int offset_bits = bd + 2 * FILTER_BITS - ROUND0_BITS;
const int16_t round_offset = (1 << (offset_bits - COMPOUND_ROUND1_BITS)) +
(1 << (offset_bits - COMPOUND_ROUND1_BITS - 1));
const int16x8_t round_offset_vec = vdupq_n_s16(round_offset);
const int do_average = conv_params->do_average;
const int use_dist_wtd_comp_avg = conv_params->use_dist_wtd_comp_avg;
const uint16_t fwd_offset = conv_params->fwd_offset;
const uint16_t bck_offset = conv_params->bck_offset;
CONV_BUF_TYPE *dst_ptr = conv_params->dst;
const int dst_stride = conv_params->dst_stride;
int width = w;
if ((w == 4) || (h == 4)) {
int16x4_t s0, s1, s2, s3, s4, s5, s6, s7;
uint16x4_t d0, dd0;
uint8x8_t t0, t1, t2, t3, t4, t5, t6, d01;
#if defined(__aarch64__)
int16x4_t s8, s9, s10;
uint16x4_t d1, d2, d3, dd1, dd2, dd3;
uint8x8_t d23;
#endif // defined(__aarch64__)
do {
const uint8_t *s = src_ptr;
CONV_BUF_TYPE *d = dst_ptr;
uint8_t *d_u8 = dst8_ptr;
int height = h;
__builtin_prefetch(s + 0 * src_stride);
__builtin_prefetch(s + 1 * src_stride);
__builtin_prefetch(s + 2 * src_stride);
__builtin_prefetch(s + 3 * src_stride);
t0 = load_unaligned_u8_4x1(s + 0 * src_stride);
t1 = load_unaligned_u8_4x1(s + 1 * src_stride);
t2 = load_unaligned_u8_4x1(s + 2 * src_stride);
t3 = load_unaligned_u8_4x1(s + 3 * src_stride);
t4 = load_unaligned_u8_4x1(s + 4 * src_stride);
t5 = load_unaligned_u8_4x1(s + 5 * src_stride);
t6 = load_unaligned_u8_4x1(s + 6 * src_stride);
s0 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t0)));
s1 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t1)));
s2 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t2)));
s3 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t3)));
s4 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t4)));
s5 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t5)));
s6 = vreinterpret_s16_u16(vget_low_u16(vmovl_u8(t6)));
__builtin_prefetch(d + 0 * dst_stride);
__builtin_prefetch(d + 1 * dst_stride);
__builtin_prefetch(d + 2 * dst_stride);
__builtin_prefetch(d + 3 * dst_stride);