virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-08-15 14:13:16 +02:00
Code Issues Releases Activity

lavc/aarch64: add HEVC idct_dc NEON

Browse Source 
Signed-off-by: Josh Dekker <josh@itanimul.li>
This commit is contained in:
Josh Dekker
2021-02-01 10:30:52 +00:00
parent 00c916ef61
commit 75c2ddfa61
2 changed files with 67 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
libavcodec/aarch64/hevcdsp_idct_neon.S
View File

@@ -5,6 +5,7 @@
*
* Ported from arm/hevcdsp_idct_neon.S by
* Copyright (c) 2020 Reimar Döffinger
* Copyright (c) 2020 Josh Dekker
*
* This file is part of FFmpeg.
*
@@ -568,3 +569,53 @@ tr_16x4 secondpass_10, 20 - 10, 512, 1
idct_16x16 8
idct_16x16 10
// void ff_hevc_idct_NxN_dc_DEPTH_neon(int16_t *coeffs)
.macro idct_dc size, bitdepth
function ff_hevc_idct_\size\()x\size\()_dc_\bitdepth\()_neon, export=1
movi v1.8h, #((1 << (14 - \bitdepth))+1)
ld1r {v4.8h}, [x0]
add v4.8h, v4.8h, v1.8h
sshr v0.8h, v4.8h, #(15 - \bitdepth)
sshr v1.8h, v4.8h, #(15 - \bitdepth)
.if \size > 4
sshr v2.8h, v4.8h, #(15 - \bitdepth)
sshr v3.8h, v4.8h, #(15 - \bitdepth)
.if \size > 16 /* dc 32x32 */
mov x2, #4
1:
subs x2, x2, #1
.endif
add x12, x0, #64
mov x13, #128
.if \size > 8 /* dc 16x16 */
st1 {v0.8h-v3.8h}, [x0], x13
st1 {v0.8h-v3.8h}, [x12], x13
st1 {v0.8h-v3.8h}, [x0], x13
st1 {v0.8h-v3.8h}, [x12], x13
st1 {v0.8h-v3.8h}, [x0], x13
st1 {v0.8h-v3.8h}, [x12], x13
.endif /* dc 8x8 */
st1 {v0.8h-v3.8h}, [x0], x13
st1 {v0.8h-v3.8h}, [x12], x13
.if \size > 16 /* dc 32x32 */
bne 1b
.endif
.else /* dc 4x4 */
st1 {v0.8h-v1.8h}, [x0]
.endif
ret
endfunc
.endm
idct_dc 4, 8
idct_dc 4, 10
idct_dc 8, 8
idct_dc 8, 10
idct_dc 16, 8
idct_dc 16, 10
idct_dc 32, 8
idct_dc 32, 10

16
libavcodec/aarch64/hevcdsp_init_aarch64.c
View File

@@ -45,6 +45,14 @@ void ff_hevc_idct_8x8_8_neon(int16_t *coeffs, int col_limit);
void ff_hevc_idct_8x8_10_neon(int16_t *coeffs, int col_limit);
void ff_hevc_idct_16x16_8_neon(int16_t *coeffs, int col_limit);
void ff_hevc_idct_16x16_10_neon(int16_t *coeffs, int col_limit);
void ff_hevc_idct_4x4_dc_8_neon(int16_t *coeffs);
void ff_hevc_idct_8x8_dc_8_neon(int16_t *coeffs);
void ff_hevc_idct_16x16_dc_8_neon(int16_t *coeffs);
void ff_hevc_idct_32x32_dc_8_neon(int16_t *coeffs);
void ff_hevc_idct_4x4_dc_10_neon(int16_t *coeffs);
void ff_hevc_idct_8x8_dc_10_neon(int16_t *coeffs);
void ff_hevc_idct_16x16_dc_10_neon(int16_t *coeffs);
void ff_hevc_idct_32x32_dc_10_neon(int16_t *coeffs);
av_cold void ff_hevc_dsp_init_aarch64(HEVCDSPContext *c, const int bit_depth)
{
@@ -57,6 +65,10 @@ av_cold void ff_hevc_dsp_init_aarch64(HEVCDSPContext *c, const int bit_depth)
c->add_residual[3] = ff_hevc_add_residual_32x32_8_neon;
c->idct[1] = ff_hevc_idct_8x8_8_neon;
c->idct[2] = ff_hevc_idct_16x16_8_neon;
c->idct_dc[0] = ff_hevc_idct_4x4_dc_8_neon;
c->idct_dc[1] = ff_hevc_idct_8x8_dc_8_neon;
c->idct_dc[2] = ff_hevc_idct_16x16_dc_8_neon;
c->idct_dc[3] = ff_hevc_idct_32x32_dc_8_neon;
}
if (bit_depth == 10) {
c->add_residual[0] = ff_hevc_add_residual_4x4_10_neon;
@@ -65,5 +77,9 @@ av_cold void ff_hevc_dsp_init_aarch64(HEVCDSPContext *c, const int bit_depth)
c->add_residual[3] = ff_hevc_add_residual_32x32_10_neon;
c->idct[1] = ff_hevc_idct_8x8_10_neon;
c->idct[2] = ff_hevc_idct_16x16_10_neon;
c->idct_dc[0] = ff_hevc_idct_4x4_dc_10_neon;
c->idct_dc[1] = ff_hevc_idct_8x8_dc_10_neon;
c->idct_dc[2] = ff_hevc_idct_16x16_dc_10_neon;
c->idct_dc[3] = ff_hevc_idct_32x32_dc_10_neon;
}
}