generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Merge "fdct8x8 highbd neon: use tran_low_t for output"

Browse Source
This commit is contained in:
Johann Koenig 2017-02-13 23:11:30 +00:00 committed by Gerrit Code Review
commit 4526ec7907
3 changed files with 20 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
test/fdct8x8_test.cc
View File

@ -673,7 +673,7 @@ INSTANTIATE_TEST_CASE_P(
#if HAVE_NEON && !CONFIG_EMULATE_HARDWARE
#if CONFIG_VP9_HIGHBITDEPTH
INSTANTIATE_TEST_CASE_P(NEON, FwdTrans8x8DCT,
::testing::Values(make_tuple(&vpx_fdct8x8_c,
::testing::Values(make_tuple(&vpx_fdct8x8_neon,
&vpx_idct8x8_64_add_neon,
0, VPX_BITS_8)));
#else // !CONFIG_VP9_HIGHBITDEPTH

27
vpx_dsp/arm/fwd_txfm_neon.c
View File

@ -12,8 +12,11 @@
#include "./vpx_config.h"
#include "vpx_dsp/txfm_common.h"
#include "vpx_dsp/vpx_dsp_common.h"
#include "vpx_dsp/arm/idct_neon.h"
void vpx_fdct8x8_neon(const int16_t *input, int16_t *final_output, int stride) {
void vpx_fdct8x8_neon(const int16_t *input, tran_low_t *final_output,
int stride) {
int i;
// stage 1
int16x8_t input_0 = vshlq_n_s16(vld1q_s16(&input[0 * stride]), 2);
@ -191,18 +194,18 @@ void vpx_fdct8x8_neon(const int16_t *input, int16_t *final_output, int stride) {
input_6 = vhsubq_s16(input_6, sign_in6);
input_7 = vhsubq_s16(input_7, sign_in7);
// store results
vst1q_s16(&final_output[0 * 8], input_0);
vst1q_s16(&final_output[1 * 8], input_1);
vst1q_s16(&final_output[2 * 8], input_2);
vst1q_s16(&final_output[3 * 8], input_3);
vst1q_s16(&final_output[4 * 8], input_4);
vst1q_s16(&final_output[5 * 8], input_5);
vst1q_s16(&final_output[6 * 8], input_6);
vst1q_s16(&final_output[7 * 8], input_7);
store_s16q_to_tran_low(final_output + 0 * 8, input_0);
store_s16q_to_tran_low(final_output + 1 * 8, input_1);
store_s16q_to_tran_low(final_output + 2 * 8, input_2);
store_s16q_to_tran_low(final_output + 3 * 8, input_3);
store_s16q_to_tran_low(final_output + 4 * 8, input_4);
store_s16q_to_tran_low(final_output + 5 * 8, input_5);
store_s16q_to_tran_low(final_output + 6 * 8, input_6);
store_s16q_to_tran_low(final_output + 7 * 8, input_7);
}
}
void vpx_fdct8x8_1_neon(const int16_t *input, int16_t *output, int stride) {
void vpx_fdct8x8_1_neon(const int16_t *input, tran_low_t *output, int stride) {
int r;
int16x8_t sum = vld1q_s16(&input[0]);
for (r = 1; r < 8; ++r) {
@ -214,7 +217,11 @@ void vpx_fdct8x8_1_neon(const int16_t *input, int16_t *output, int stride) {
const int64x2_t b = vpaddlq_s32(a);
const int32x2_t c = vadd_s32(vreinterpret_s32_s64(vget_low_s64(b)),
vreinterpret_s32_s64(vget_high_s64(b)));
#if CONFIG_VP9_HIGHBITDEPTH
output[0] = vget_lane_s32(c, 0);
#else
output[0] = vget_lane_s16(vreinterpret_s16_s32(c), 0);
#endif
output[1] = 0;
}
}

4
vpx_dsp/vpx_dsp_rtcd_defs.pl
View File

@ -536,10 +536,10 @@ if (vpx_config("CONFIG_VP9_HIGHBITDEPTH") eq "yes") {
specialize qw/vpx_fdct4x4_1 sse2/;
add_proto qw/void vpx_fdct8x8/, "const int16_t *input, tran_low_t *output, int stride";
specialize qw/vpx_fdct8x8 sse2/;
specialize qw/vpx_fdct8x8 neon sse2/;
add_proto qw/void vpx_fdct8x8_1/, "const int16_t *input, tran_low_t *output, int stride";
specialize qw/vpx_fdct8x8_1 sse2/;
specialize qw/vpx_fdct8x8_1 neon sse2/;
add_proto qw/void vpx_fdct16x16/, "const int16_t *input, tran_low_t *output, int stride";
specialize qw/vpx_fdct16x16 sse2/;