Skip to content
Snippets Groups Projects

Compare revisions

Changes are shown as if the source revision was being merged into the target revision. Learn more about comparing revisions.

Source

Select target project
No results found

Target

Select target project
  • videolan/x264
  • EwoutH/x264
  • gramner/x264
  • BugMaster/x264
  • MaskRay/x264
  • thresh/x264
  • tpm/x264
  • wolfired/x264
  • ifb/x264
  • robinstorm/x264
  • ltnokiago/x264
  • janne/x264
  • Kromjunya/x264
  • trisnaayu0596/x264
  • felipegarcia1402/x264
  • coder2004/x264
  • philou/x264
  • walagnatalia/x264
  • DonDiego/x264
  • JHammler/x264
  • qyot27/x264
  • dwbuiten/x264
  • Kagami/x264
  • andriy-andreyev/x264
  • gxw/x264
  • trofi/x264
  • kierank/x264
  • aureliendavid/x264
  • galad/x264
  • roommini/x264
  • ocrete/x264
  • mstorsjo/x264
  • yinsj0116/x264
  • mamonet/x264
  • 1div0/x264
  • ko1265/x264
  • sergiomb2/x264
  • xutongda/x264
  • wenzhiwu/x264
  • arrowd/x264
  • FranceBB/x264
  • ziemek99/x264
  • longervision/x264
  • xopok/x264
  • jbk/x264
  • szatmary/x264
  • pekdon/x264
  • Jiangguyu/x264
  • jrtc27/x264
  • kankanol1/x264
  • gxwLite/x264
  • brad/x264
  • Gc6026/x264
  • jdek/x264
  • appcrash/x264
  • tguillem/x264
  • As/x264
  • wevian/x264
  • wangluls/x264
  • RellikJaeger/x264
  • hum/x264
  • rogerhardiman/x264
  • jankowalski12611/x264
  • zhijie1996/x264
  • yinshiyou/x264
  • Freed-Wu/x264
  • yajcoca/x264
  • bUd/x264
  • chienvannguyen2020/x264
  • nurbinakhatun386/x264
  • Siberiawind/x-264-meson
  • HecaiYuan/x264
  • david.chen/x264
  • Ytsejam76/x264
  • robUx4/x264
  • zhaoshiz/x-264-arm64ec
  • yintong.ustc/x-264-bd-ventana
  • nekobasu/x264
  • Courmisch/x264
  • BD-qjy/x264
  • quink/x264
  • markos/x264
82 results
Show changes
Commits on Source (9)
  • Yin Shiyou's avatar
    loongarch: Init LSX/LASX support · 1ecc51ee
    Yin Shiyou authored and Yin Shiyou's avatar Yin Shiyou committed
    
    LSX/LASX is the LOONGARCH 128-bit/256-bit SIMD Architecture.
    
    Signed-off-by: default avatarShiyou Yin <yinshiyou-hf@loongson.cn>
    Signed-off-by: default avatarXiwei Gu <guxiwei-hf@loongson.cn>
    1ecc51ee
  • Yin Shiyou's avatar
    loongarch: Add loongson_asm.S and loongson_utils.S · 25ffd616
    Yin Shiyou authored and Yin Shiyou's avatar Yin Shiyou committed
    
    Common macros and functions for loongson optimization.
    
    Signed-off-by: default avatarShiyou Yin <yinshiyou-hf@loongson.cn>
    25ffd616
  • Yin Shiyou's avatar
    loongarch: Improve the performance of deblock series functions. · d7d283f6
    Yin Shiyou authored and Yin Shiyou's avatar Yin Shiyou committed
    
    Performance has improved from 4.76fps to 4.92fps.
    Tested with following command:
    ./configure && make -j5
    ./x264 --threads 4 -o out.mkv yuv_1920x1080.yuv
    
    functions           performance     performance
                            (c)            (asm)
    deblock_luma[0]         79               39
    deblock_luma[1]         91               18
    deblock_luma_intra[0]   63               44
    deblock_luma_intra[1]   71               18
    deblock_strength        104              33
    
    Signed-off-by: default avatarHao Chen <chenhao@loongson.cn>
    d7d283f6
  • Yin Shiyou's avatar
    loongarch: Improve the performance of sad/sad_x3/sad_x4 series functions · 00b8e3b9
    Yin Shiyou authored and Yin Shiyou's avatar Yin Shiyou committed
    
    Performance has improved from 4.92fps to 6.32fps.
    Tested with following command:
    ./configure && make -j5
    ./x264 --threads 4 -o out.mkv yuv_1920x1080.yuv
    
    functions           performance     performance
                            (c)            (asm)
    sad_4x4                 13               3
    sad_4x8                 26               7
    sad_4x16                57               13
    sad_8x4                 24               3
    sad_8x8                 54               8
    sad_8x16                108              13
    sad_16x8                95               8
    sad_16x16               189              13
    sad_x3_4x4              37               6
    sad_x3_4x8              71               13
    sad_x3_8x4              70               8
    sad_x3_8x8              162              14
    sad_x3_8x16             323              25
    sad_x3_16x8             279              15
    sad_x3_16x16            555              27
    sad_x4_4x4              49               8
    sad_x4_4x8              95               17
    sad_x4_8x4              94               8
    sad_x4_8x8              214              16
    sad_x4_8x16             429              33
    sad_x4_16x8             372              18
    sad_x4_16x16            740              34
    
    Signed-off-by: default avatarwanglu <wanglu@loongson.cn>
    00b8e3b9
  • Yin Shiyou's avatar
    loongarch: Improve the performance of predict series functions · d8ed272a
    Yin Shiyou authored and Yin Shiyou's avatar Yin Shiyou committed
    
    Performance has improved from 6.32fps to 6.34fps.
    Tested with following command:
    ./configure && make -j5
    ./x264 --threads 4 -o out.mkv yuv_1920x1080.yuv
    
    functions           performance     performance
                            (c)            (asm)
    intra_predict_4x4_dc     3               2
    intra_predict_4x4_dc8    1               1
    intra_predict_4x4_dcl    2               1
    intra_predict_4x4_dct    2               1
    intra_predict_4x4_ddl    7               2
    intra_predict_4x4_h      2               1
    intra_predict_4x4_v      1               1
    intra_predict_8x8_dc     8               2
    intra_predict_8x8_dc8    1               1
    intra_predict_8x8_dcl    5               2
    intra_predict_8x8_dct    5               2
    intra_predict_8x8_ddl    27              3
    intra_predict_8x8_ddr    26              3
    intra_predict_8x8_h      4               2
    intra_predict_8x8_v      3               1
    intra_predict_8x8_vl     29              3
    intra_predict_8x8_vr     31              4
    intra_predict_8x8c_dc    8               5
    intra_predict_8x8c_dc8   1               1
    intra_predict_8x8c_dcl   5               3
    intra_predict_8x8c_dct   5               3
    intra_predict_8x8c_h     4               2
    intra_predict_8x8c_p     58              30
    intra_predict_8x8c_v     4               1
    intra_predict_16x16_dc   32              8
    intra_predict_16x16_dc8  9               4
    intra_predict_16x16_dcl  26              6
    intra_predict_16x16_dct  26              6
    intra_predict_16x16_h    23              7
    intra_predict_16x16_p    182             44
    intra_predict_16x16_v    22              4
    
    Signed-off-by: default avatarXiwei Gu <guxiwei-hf@loongson.cn>
    d8ed272a
  • Yin Shiyou's avatar
    loongarch: Improve the performance of quant series functions · 65e7bac5
    Yin Shiyou authored and Yin Shiyou's avatar Yin Shiyou committed
    
    Performance has improved from 6.34fps to 6.78fps.
    Tested with following command:
    ./configure && make -j5
    ./x264 --threads 4 -o out.mkv yuv_1920x1080.yuv
    
    functions           performance     performance
                            (c)            (asm)
    coeff_last15             3               2
    coeff_last16             3               1
    coeff_last64             42              6
    decimate_score15         8               12
    decimate_score16         8               11
    decimate_score64         61              43
    dequant_4x4_cqm          16              5
    dequant_4x4_dc_cqm       13              5
    dequant_4x4_dc_flat      13              5
    dequant_4x4_flat         16              5
    dequant_8x8_cqm          71              9
    dequant_8x8_flat         71              9
    
    Signed-off-by: default avatarShiyou Yin <yinshiyou-hf@loongson.cn>
    65e7bac5
  • Yin Shiyou's avatar
    loongarch: Improve the performance of mc series functions · 981c8f25
    Yin Shiyou authored
    
    Performance has improved from 6.78fps to 10.53fps.
    Tested with following command:
    ./configure && make -j5
    ./x264 --threads 4 -o out.mkv yuv_1920x1080.yuv
    
    functions           performance     performance
                            (c)            (asm)
    avg_4x2                  16              5
    avg_4x4                  30              6
    avg_4x8                  63              10
    avg_4x16                 124             19
    avg_8x4                  60              6
    avg_8x8                  119             10
    avg_8x16                 233             19
    avg_16x8                 229             21
    avg_16x16                451             41
    get_ref_4x4              30              9
    get_ref_4x8              52              11
    get_ref_8x4              45              9
    get_ref_8x8              80              11
    get_ref_8x16             156             16
    get_ref_12x10            137             13
    get_ref_16x8             147             11
    get_ref_16x16            282             16
    get_ref_20x18            278             22
    hpel_filter              5163            686
    lowres_init              5440            286
    mc_chroma_2x2            24              7
    mc_chroma_2x4            42              10
    mc_chroma_4x2            41              7
    mc_chroma_4x4            75              10
    mc_chroma_4x8            144             19
    mc_chroma_8x4            137             15
    mc_chroma_8x8            269             28
    mc_luma_4x4              30              10
    mc_luma_4x8              52              12
    mc_luma_8x4              44              10
    mc_luma_8x8              80              13
    mc_luma_8x16             156             19
    mc_luma_16x8             147             13
    mc_luma_16x16            281             19
    memcpy_aligned           14              9
    memzero_aligned          24              4
    offsetadd_w4             79              18
    offsetadd_w8             142             18
    offsetadd_w16            277             25
    offsetadd_w20            1118            38
    offsetsub_w4             75              18
    offsetsub_w8             140             18
    offsetsub_w16            265             25
    offsetsub_w20            989             39
    weight_w4                111             19
    weight_w8                205             19
    weight_w16               396             29
    weight_w20               1143            45
    deinterleave_chroma_fdec 76              9
    deinterleave_chroma_fenc 86              9
    plane_copy_deinterleave  733             90
    plane_copy_interleave    791             245
    store_interleave_chroma  82              12
    
    Signed-off-by: default avatarXiwei Gu <guxiwei-hf@loongson.cn>
    981c8f25
  • Yin Shiyou's avatar
    loongarch: Improve the performance of dct series functions · fa7f1fce
    Yin Shiyou authored
    
    Performance has improved from 10.53fps to 11.27fps.
    Tested with following command:
    ./configure && make -j5
    ./x264 --threads 4 -o out.mkv yuv_1920x1080.yuv
    
    functions           performance     performance
                            (c)            (asm)
    add4x4_idct              34              9
    add8x8_idct              139             31
    add8x8_idct8             269             39
    add8x8_idct_dc           67              7
    add16x16_idct            564             123
    add16x16_idct_dc         260             22
    dct4x4dc                 18              10
    idct4x4dc                16              9
    sub4x4_dct               25              7
    sub8x8_dct               101             12
    sub8x8_dct8              160             25
    sub16x16_dct             403             52
    sub16x16_dct8            646             68
    zigzag_scan_4x4_frame    4               1
    
    Signed-off-by: default avatarzhoupeng <zhoupeng@loongson.cn>
    fa7f1fce
  • Yin Shiyou's avatar
    loongarch: Improve the performance of pixel series functions · 5f84d403
    Yin Shiyou authored
    
    Performance has improved from 11.27fps to 20.50fps by using the
    following command:
    ./configure && make -j5
    ./x264 --threads 4 -o out.mkv yuv_1920x1080.yuv
    
    functions           performance     performance
                            (c)            (asm)
    hadamard_ac_8x8          117             21
    hadamard_ac_8x16         236             42
    hadamard_ac_16x8         235             31
    hadamard_ac_16x16        473             60
    intra_sad_x3_4x4         50              21
    intra_sad_x3_8x8         183             34
    intra_sad_x3_8x8c        181             36
    intra_sad_x3_16x16       643             68
    intra_satd_x3_4x4        83              61
    intra_satd_x3_8x8c       344             81
    intra_satd_x3_16x16      1389            136
    sa8d_8x8                 97              19
    sa8d_16x16               394             68
    satd_4x4                 24              8
    satd_4x8                 51              11
    satd_4x16                103             24
    satd_8x4                 52              9
    satd_8x8                 108             12
    satd_8x16                218             24
    satd_16x8                218             19
    satd_16x16               437             38
    ssd_4x4                  10              5
    ssd_4x8                  24              8
    ssd_4x16                 42              15
    ssd_8x4                  23              5
    ssd_8x8                  37              9
    ssd_8x16                 74              17
    ssd_16x8                 72              11
    ssd_16x16                140             23
    var2_8x8                 91              37
    var2_8x16                176             66
    var_8x8                  50              15
    var_8x16                 65              29
    var_16x16                132             56
    
    Signed-off-by: default avatarHecai Yuan <yuanhecai@loongson.cn>
    5f84d403
Showing
with 14718 additions and 1 deletion
......@@ -197,6 +197,32 @@ SRCS_X += common/mips/dct-c.c \
endif
endif
# LOONGARCH optimization
ifeq ($(SYS_ARCH),LOONGARCH)
ifneq ($(findstring HAVE_LSX 1, $(CONFIG)),)
SRCASM_X += common/loongarch/deblock-a.S \
common/loongarch/sad-a.S \
common/loongarch/predict-a.S \
common/loongarch/quant-a.S \
common/loongarch/mc-a.S \
common/loongarch/dct-a.S \
common/loongarch/pixel-a.S
SRCS_X += common/loongarch/predict-c.c \
common/loongarch/mc-c.c \
common/loongarch/pixel-c.c
OBJASM +=
ifneq ($(findstring HAVE_BITDEPTH8 1, $(CONFIG)),)
OBJASM += $(SRCASM_X:%.S=%-8.o)
endif
ifneq ($(findstring HAVE_BITDEPTH10 1, $(CONFIG)),)
OBJASM += $(SRCASM_X:%.S=%-10.o)
endif
endif
endif
endif
ifneq ($(HAVE_GETOPT_LONG),1)
......
......@@ -98,6 +98,9 @@ const x264_cpu_name_t x264_cpu_names[] =
{"NEON", X264_CPU_NEON},
#elif ARCH_MIPS
{"MSA", X264_CPU_MSA},
#elif ARCH_LOONGARCH
{"LSX", X264_CPU_LSX},
{"LASX", X264_CPU_LASX},
#endif
{"", 0},
};
......@@ -431,6 +434,25 @@ uint32_t x264_cpu_detect( void )
return X264_CPU_MSA;
}
#elif HAVE_LSX
#include <sys/auxv.h>
#define LA_HWCAP_LSX ( 1U << 4 )
#define LA_HWCAP_LASX ( 1U << 5 )
uint32_t x264_cpu_detect( void )
{
uint32_t flags = 0;
uint32_t hwcap = (uint32_t)getauxval( AT_HWCAP );
if( hwcap & LA_HWCAP_LSX )
flags |= X264_CPU_LSX;
if( hwcap & LA_HWCAP_LASX )
flags |= X264_CPU_LASX;
return flags;
}
#else
uint32_t x264_cpu_detect( void )
......
......@@ -41,7 +41,9 @@
#if HAVE_MSA
# include "mips/dct.h"
#endif
#if HAVE_LSX
# include "loongarch/dct.h"
#endif
static void dct4x4dc( dctcoef d[16] )
{
dctcoef tmp[16];
......@@ -727,6 +729,38 @@ void x264_dct_init( uint32_t cpu, x264_dct_function_t *dctf )
}
#endif
#if HAVE_LSX
if( cpu&X264_CPU_LSX )
{
dctf->sub4x4_dct = x264_sub4x4_dct_lsx;
dctf->add4x4_idct = x264_add4x4_idct_lsx;
dctf->dct4x4dc = x264_dct4x4dc_lsx;
dctf->idct4x4dc = x264_idct4x4dc_lsx;
dctf->sub8x8_dct8 = x264_sub8x8_dct8_lsx;
dctf->sub8x8_dct = x264_sub8x8_dct_lsx;
dctf->add8x8_idct = x264_add8x8_idct_lsx;
dctf->add8x8_idct8 = x264_add8x8_idct8_lsx;
dctf->add8x8_idct_dc = x264_add8x8_idct_dc_lsx;
dctf->add16x16_idct = x264_add16x16_idct_lsx;
dctf->sub16x16_dct = x264_sub16x16_dct_lsx;
dctf->add16x16_idct_dc = x264_add16x16_idct_dc_lsx;
dctf->sub16x16_dct8 = x264_sub16x16_dct8_lsx;
}
if( cpu&X264_CPU_LASX )
{
dctf->sub8x8_dct = x264_sub8x8_dct_lasx;
dctf->sub16x16_dct = x264_sub16x16_dct_lasx;
dctf->add8x8_idct = x264_add8x8_idct_lasx;
dctf->add8x8_idct8 = x264_add8x8_idct8_lasx;
dctf->add16x16_idct = x264_add16x16_idct_lasx;
dctf->sub16x16_dct8 = x264_sub16x16_dct8_lasx;
dctf->add8x8_idct_dc = x264_add8x8_idct_dc_lasx;
dctf->add16x16_idct_dc = x264_add16x16_idct_dc_lasx;
dctf->dct4x4dc = x264_dct4x4dc_lasx;
dctf->idct4x4dc = x264_idct4x4dc_lasx;
}
#endif
#endif // HIGH_BIT_DEPTH
}
......@@ -1087,5 +1121,12 @@ void x264_zigzag_init( uint32_t cpu, x264_zigzag_function_t *pf_progressive, x26
pf_progressive->scan_4x4 = x264_zigzag_scan_4x4_frame_msa;
}
#endif
#if HAVE_LSX
if( cpu&X264_CPU_LASX )
{
pf_progressive->scan_4x4 = x264_zigzag_scan_4x4_frame_lasx;
}
#endif
#endif // !HIGH_BIT_DEPTH
}
......@@ -680,6 +680,9 @@ void x264_macroblock_deblock( x264_t *h )
#if HAVE_MSA
#include "mips/deblock.h"
#endif
#if HAVE_LSX
#include "loongarch/deblock.h"
#endif
void x264_deblock_init( uint32_t cpu, x264_deblock_function_t *pf, int b_mbaff )
{
......@@ -816,6 +819,24 @@ void x264_deblock_init( uint32_t cpu, x264_deblock_function_t *pf, int b_mbaff )
pf->deblock_strength = x264_deblock_strength_msa;
}
#endif
#if HAVE_LSX
if( cpu&X264_CPU_LSX )
{
pf->deblock_luma_intra[1] = x264_deblock_v_luma_intra_lsx;
pf->deblock_luma_intra[0] = x264_deblock_h_luma_intra_lsx;
pf->deblock_strength = x264_deblock_strength_lsx;
}
if( cpu&X264_CPU_LASX )
{
pf->deblock_luma[1] = x264_deblock_v_luma_lasx;
pf->deblock_luma[0] = x264_deblock_h_luma_lasx;
pf->deblock_luma_intra[1] = x264_deblock_v_luma_intra_lasx;
pf->deblock_luma_intra[0] = x264_deblock_h_luma_intra_lasx;
pf->deblock_strength = x264_deblock_strength_lasx;
}
#endif
#endif // !HIGH_BIT_DEPTH
/* These functions are equivalent, so don't duplicate them. */
......
This diff is collapsed.
/*****************************************************************************
* dct.h: loongarch transform and zigzag
*****************************************************************************
* Copyright (C) 2023 x264 project
*
* Authors: Peng Zhou <zhoupeng@loongson.cn>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
*
* This program is also available under a commercial proprietary license.
* For more information, contact us at licensing@x264.com.
*****************************************************************************/
#ifndef X264_LOONGARCH_DCT_H
#define X264_LOONGARCH_DCT_H
#define x264_sub8x8_dct_lasx x264_template(sub8x8_dct_lasx)
void x264_sub8x8_dct_lasx( int16_t p_dst[4][16], uint8_t *p_src, uint8_t *p_ref );
#define x264_sub16x16_dct_lasx x264_template(sub16x16_dct_lasx)
void x264_sub16x16_dct_lasx( int16_t p_dst[16][16], uint8_t *p_src, uint8_t *p_ref );
#define x264_sub8x8_dct8_lsx x264_template(sub8x8_dct8_lsx)
void x264_sub8x8_dct8_lsx( int16_t pi_dct[64], uint8_t *p_pix1, uint8_t *p_pix2 );
#define x264_sub16x16_dct8_lasx x264_template(sub16x16_dct8_lasx)
void x264_sub16x16_dct8_lasx( int16_t pi_dct[4][64], uint8_t *p_pix1,
uint8_t *p_pix2 );
#define x264_add4x4_idct_lsx x264_template(add4x4_idct_lsx)
void x264_add4x4_idct_lsx( uint8_t *p_dst, int16_t pi_dct[16] );
#define x264_add8x8_idct_lasx x264_template(add8x8_idct_lasx)
void x264_add8x8_idct_lasx( uint8_t *p_dst, int16_t pi_dct[4][16] );
#define x264_add16x16_idct_lasx x264_template(add16x16_idct_lasx)
void x264_add16x16_idct_lasx( uint8_t *p_dst, int16_t pi_dct[16][16] );
#define x264_add8x8_idct8_lasx x264_template(add8x8_idct8_lasx)
void x264_add8x8_idct8_lasx( uint8_t *p_dst, int16_t pi_dct[64] );
#define x264_add8x8_idct_dc_lasx x264_template(add8x8_idct_dc_lasx)
void x264_add8x8_idct_dc_lasx( uint8_t *p_dst, int16_t dct[4] );
#define x264_add16x16_idct_dc_lasx x264_template(add16x16_idct_dc_lasx)
void x264_add16x16_idct_dc_lasx( uint8_t *p_dst, int16_t dct[16] );
#define x264_idct4x4dc_lasx x264_template(idct4x4dc_lasx)
void x264_idct4x4dc_lasx( int16_t d[16] );
#define x264_dct4x4dc_lasx x264_template(dct4x4dc_lasx)
void x264_dct4x4dc_lasx( int16_t d[16] );
#define x264_zigzag_scan_4x4_frame_lasx x264_template(zigzag_scan_4x4_frame_lasx)
void x264_zigzag_scan_4x4_frame_lasx( int16_t level[16], int16_t dct[16] );
#define x264_sub4x4_dct_lsx x264_template(sub4x4_dct_lsx)
void x264_sub4x4_dct_lsx( int16_t p_dst[16], uint8_t *p_src, uint8_t *p_ref );
#define x264_sub8x8_dct_lsx x264_template(sub8x8_dct_lsx)
void x264_sub8x8_dct_lsx( int16_t p_dst[4][16], uint8_t *p_src, uint8_t *p_ref );
#define x264_sub16x16_dct_lsx x264_template(sub16x16_dct_lsx)
void x264_sub16x16_dct_lsx( int16_t p_dst[16][16], uint8_t *p_src, uint8_t *p_ref );
#define x264_sub8x8_dct8_lsx x264_template(sub8x8_dct8_lsx)
void x264_sub8x8_dct8_lsx( int16_t pi_dct[64], uint8_t *p_pix1, uint8_t *p_pix2 );
#define x264_sub16x16_dct8_lsx x264_template(sub16x16_dct8_lsx)
void x264_sub16x16_dct8_lsx( int16_t pi_dct[4][64], uint8_t *p_pix1,
uint8_t *p_pix2 );
#define x264_add4x4_idct_lsx x264_template(add4x4_idct_lsx)
void x264_add4x4_idct_lsx( uint8_t *p_dst, int16_t pi_dct[16] );
#define x264_add8x8_idct_lsx x264_template(add8x8_idct_lsx)
void x264_add8x8_idct_lsx( uint8_t *p_dst, int16_t pi_dct[4][16] );
#define x264_add16x16_idct_lsx x264_template(add16x16_idct_lsx)
void x264_add16x16_idct_lsx( uint8_t *p_dst, int16_t pi_dct[16][16] );
#define x264_add8x8_idct8_lsx x264_template(add8x8_idct8_lsx)
void x264_add8x8_idct8_lsx( uint8_t *p_dst, int16_t pi_dct[64] );
#define x264_add8x8_idct_dc_lsx x264_template(add8x8_idct_dc_lsx)
void x264_add8x8_idct_dc_lsx( uint8_t *p_dst, int16_t dct[4] );
#define x264_add16x16_idct_dc_lsx x264_template(add16x16_idct_dc_lsx)
void x264_add16x16_idct_dc_lsx( uint8_t *p_dst, int16_t dct[16] );
#define x264_idct4x4dc_lsx x264_template(idct4x4dc_lsx)
void x264_idct4x4dc_lsx( int16_t d[16] );
#define x264_dct4x4dc_lsx x264_template(dct4x4dc_lsx)
void x264_dct4x4dc_lsx( int16_t d[16] );
#define x264_zigzag_scan_4x4_frame_lsx x264_template(zigzag_scan_4x4_frame_lsx)
void x264_zigzag_scan_4x4_frame_lsx( int16_t level[16], int16_t dct[16] );
#endif
This diff is collapsed.
/*****************************************************************************
* deblock.h: loongarch deblock
*****************************************************************************
* Copyright (C) 2023 x264 project
*
* Authors: Hao Chen <chenhao@loongson.cn>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
*
* This program is also available under a commercial proprietary license.
* For more information, contact us at licensing@x264.com.
*****************************************************************************/
#ifndef X264_LOONGARCH_DEBLOCK_H
#define X264_LOONGARCH_DEBLOCK_H
#if !HIGH_BIT_DEPTH
#define x264_deblock_v_luma_lasx x264_template(deblock_v_luma_lasx)
void x264_deblock_v_luma_lasx( uint8_t *pix, intptr_t stride, int alpha, int beta, int8_t *tc0 );
#define x264_deblock_h_luma_lasx x264_template(deblock_h_luma_lasx)
void x264_deblock_h_luma_lasx( uint8_t *pix, intptr_t stride, int alpha, int beta, int8_t *tc0 );
#define x264_deblock_v_luma_intra_lsx x264_template(deblock_v_luma_intra_lsx)
void x264_deblock_v_luma_intra_lsx( uint8_t *pix, intptr_t stride, int alpha, int beta );
#define x264_deblock_h_luma_intra_lsx x264_template(deblock_h_luma_intra_lsx)
void x264_deblock_h_luma_intra_lsx( uint8_t *pix, intptr_t stride, int alpha, int beta );
#define x264_deblock_v_luma_intra_lasx x264_template(deblock_v_luma_intra_lasx)
void x264_deblock_v_luma_intra_lasx( uint8_t *pix, intptr_t stride, int alpha, int beta );
#define x264_deblock_h_luma_intra_lasx x264_template(deblock_h_luma_intra_lasx)
void x264_deblock_h_luma_intra_lasx( uint8_t *pix, intptr_t stride, int alpha, int beta );
#define x264_deblock_strength_lsx x264_template(deblock_strength_lsx)
void x264_deblock_strength_lsx( uint8_t nnz[X264_SCAN8_SIZE], int8_t ref[2][X264_SCAN8_LUMA_SIZE],
int16_t mv[2][X264_SCAN8_LUMA_SIZE][2], uint8_t bs[2][8][4],
int mvy_limit, int bframe );
#define x264_deblock_strength_lasx x264_template(deblock_strength_lasx)
void x264_deblock_strength_lasx( uint8_t nnz[X264_SCAN8_SIZE], int8_t ref[2][X264_SCAN8_LUMA_SIZE],
int16_t mv[2][X264_SCAN8_LUMA_SIZE][2], uint8_t bs[2][8][4],
int mvy_limit, int bframe );
#endif
#endif
/*********************************************************************
* Copyright (c) 2022 Loongson Technology Corporation Limited
* Contributed by Xiwei Gu <guxiwei-hf@loongson.cn>
* Shiyou Yin <yinshiyou-hf@loongson.cn>
*
* Permission to use, copy, modify, and/or distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*********************************************************************/
/*
* This file is a LoongArch assembly helper file and available under ISC
* license. It provides a large number of macros and alias to simplify
* writing assembly code, especially for LSX and LASX optimizations.
*
* Any one can modify it or add new features for his/her own purposes.
* Contributing a patch will be appreciated as it might be useful for
* others as well. Send patches to loongson contributor mentioned above.
*
* MAJOR version: Usage changes, incompatible with previous version.
* MINOR version: Add new macros/functions, or bug fixes.
* MICRO version: Comment changes or implementation changes.
*/
#define LML_VERSION_MAJOR 0
#define LML_VERSION_MINOR 2
#define LML_VERSION_MICRO 2
#define ASM_PREF
#define DEFAULT_ALIGN 5
.macro function name, align=DEFAULT_ALIGN
.macro endfunc
jirl $r0, $r1, 0x0
.size ASM_PREF\name, . - ASM_PREF\name
.purgem endfunc
.endm
.text ;
.align \align ;
.globl ASM_PREF\name ;
.type ASM_PREF\name, @function ;
ASM_PREF\name: ;
.endm
.macro const name, align=DEFAULT_ALIGN
.macro endconst
.size \name, . - \name
.purgem endconst
.endm
.section .rodata
.align \align
\name:
.endm
/*
*============================================================================
* LoongArch register alias
*============================================================================
*/
#define a0 $a0
#define a1 $a1
#define a2 $a2
#define a3 $a3
#define a4 $a4
#define a5 $a5
#define a6 $a6
#define a7 $a7
#define t0 $t0
#define t1 $t1
#define t2 $t2
#define t3 $t3
#define t4 $t4
#define t5 $t5
#define t6 $t6
#define t7 $t7
#define t8 $t8
#define s0 $s0
#define s1 $s1
#define s2 $s2
#define s3 $s3
#define s4 $s4
#define s5 $s5
#define s6 $s6
#define s7 $s7
#define s8 $s8
#define zero $zero
#define sp $sp
#define ra $ra
#define f0 $f0
#define f1 $f1
#define f2 $f2
#define f3 $f3
#define f4 $f4
#define f5 $f5
#define f6 $f6
#define f7 $f7
#define f8 $f8
#define f9 $f9
#define f10 $f10
#define f11 $f11
#define f12 $f12
#define f13 $f13
#define f14 $f14
#define f15 $f15
#define f16 $f16
#define f17 $f17
#define f18 $f18
#define f19 $f19
#define f20 $f20
#define f21 $f21
#define f22 $f22
#define f23 $f23
#define f24 $f24
#define f25 $f25
#define f26 $f26
#define f27 $f27
#define f28 $f28
#define f29 $f29
#define f30 $f30
#define f31 $f31
#define vr0 $vr0
#define vr1 $vr1
#define vr2 $vr2
#define vr3 $vr3
#define vr4 $vr4
#define vr5 $vr5
#define vr6 $vr6
#define vr7 $vr7
#define vr8 $vr8
#define vr9 $vr9
#define vr10 $vr10
#define vr11 $vr11
#define vr12 $vr12
#define vr13 $vr13
#define vr14 $vr14
#define vr15 $vr15
#define vr16 $vr16
#define vr17 $vr17
#define vr18 $vr18
#define vr19 $vr19
#define vr20 $vr20
#define vr21 $vr21
#define vr22 $vr22
#define vr23 $vr23
#define vr24 $vr24
#define vr25 $vr25
#define vr26 $vr26
#define vr27 $vr27
#define vr28 $vr28
#define vr29 $vr29
#define vr30 $vr30
#define vr31 $vr31
#define xr0 $xr0
#define xr1 $xr1
#define xr2 $xr2
#define xr3 $xr3
#define xr4 $xr4
#define xr5 $xr5
#define xr6 $xr6
#define xr7 $xr7
#define xr8 $xr8
#define xr9 $xr9
#define xr10 $xr10
#define xr11 $xr11
#define xr12 $xr12
#define xr13 $xr13
#define xr14 $xr14
#define xr15 $xr15
#define xr16 $xr16
#define xr17 $xr17
#define xr18 $xr18
#define xr19 $xr19
#define xr20 $xr20
#define xr21 $xr21
#define xr22 $xr22
#define xr23 $xr23
#define xr24 $xr24
#define xr25 $xr25
#define xr26 $xr26
#define xr27 $xr27
#define xr28 $xr28
#define xr29 $xr29
#define xr30 $xr30
#define xr31 $xr31
/*
*============================================================================
* LSX/LASX synthesize instructions
*============================================================================
*/
/*
* Description : Dot product of byte vector elements
* Arguments : Inputs - vj, vk
* Outputs - vd
* Return Type - halfword
*/
.macro vdp2.h.bu vd, vj, vk
vmulwev.h.bu \vd, \vj, \vk
vmaddwod.h.bu \vd, \vj, \vk
.endm
.macro vdp2.h.bu.b vd, vj, vk
vmulwev.h.bu.b \vd, \vj, \vk
vmaddwod.h.bu.b \vd, \vj, \vk
.endm
.macro vdp2.w.h vd, vj, vk
vmulwev.w.h \vd, \vj, \vk
vmaddwod.w.h \vd, \vj, \vk
.endm
.macro xvdp2.h.bu xd, xj, xk
xvmulwev.h.bu \xd, \xj, \xk
xvmaddwod.h.bu \xd, \xj, \xk
.endm
.macro xvdp2.h.bu.b xd, xj, xk
xvmulwev.h.bu.b \xd, \xj, \xk
xvmaddwod.h.bu.b \xd, \xj, \xk
.endm
.macro xvdp2.w.h xd, xj, xk
xvmulwev.w.h \xd, \xj, \xk
xvmaddwod.w.h \xd, \xj, \xk
.endm
/*
* Description : Dot product & addition of halfword vector elements
* Arguments : Inputs - vj, vk
* Outputs - vd
* Return Type - twice size of input
*/
.macro vdp2add.h.bu vd, vj, vk
vmaddwev.h.bu \vd, \vj, \vk
vmaddwod.h.bu \vd, \vj, \vk
.endm
.macro vdp2add.h.bu.b vd, vj, vk
vmaddwev.h.bu.b \vd, \vj, \vk
vmaddwod.h.bu.b \vd, \vj, \vk
.endm
.macro vdp2add.w.h vd, vj, vk
vmaddwev.w.h \vd, \vj, \vk
vmaddwod.w.h \vd, \vj, \vk
.endm
.macro xvdp2add.h.bu.b xd, xj, xk
xvmaddwev.h.bu.b \xd, \xj, \xk
xvmaddwod.h.bu.b \xd, \xj, \xk
.endm
.macro xvdp2add.w.h xd, xj, xk
xvmaddwev.w.h \xd, \xj, \xk
xvmaddwod.w.h \xd, \xj, \xk
.endm
/*
* Description : Range each element of vector
* clip: vj > vk ? vj : vk && vj < va ? vj : va
* clip255: vj < 255 ? vj : 255 && vj > 0 ? vj : 0
*/
.macro vclip.h vd, vj, vk, va
vmax.h \vd, \vj, \vk
vmin.h \vd, \vd, \va
.endm
.macro vclip255.w vd, vj
vmaxi.w \vd, \vj, 0
vsat.wu \vd, \vd, 7
.endm
.macro xvclip.h xd, xj, xk, xa
xvmax.h \xd, \xj, \xk
xvmin.h \xd, \xd, \xa
.endm
.macro xvclip255.h xd, xj
xvmaxi.h \xd, \xj, 0
xvsat.hu \xd, \xd, 7
.endm
.macro xvclip255.w xd, xj
xvmaxi.w \xd, \xj, 0
xvsat.wu \xd, \xd, 7
.endm
/*
* Description : Store elements of vector
* vd : Data vector to be stroed
* rk : Address of data storage
* ra : Offset of address
* si : Index of data in vd
*/
.macro vstelmx.b vd, rk, ra, si
add.d \rk, \rk, \ra
vstelm.b \vd, \rk, 0, \si
.endm
.macro vstelmx.h vd, rk, ra, si
add.d \rk, \rk, \ra
vstelm.h \vd, \rk, 0, \si
.endm
.macro vstelmx.w vd, rk, ra, si
add.d \rk, \rk, \ra
vstelm.w \vd, \rk, 0, \si
.endm
.macro vstelmx.d vd, rk, ra, si
add.d \rk, \rk, \ra
vstelm.d \vd, \rk, 0, \si
.endm
.macro vmov xd, xj
vor.v \xd, \xj, \xj
.endm
.macro xmov xd, xj
xvor.v \xd, \xj, \xj
.endm
.macro xvstelmx.d xd, rk, ra, si
add.d \rk, \rk, \ra
xvstelm.d \xd, \rk, 0, \si
.endm
/*
*============================================================================
* LSX/LASX custom macros
*============================================================================
*/
/*
* Load 4 float, double, V128, v256 elements with stride.
*/
.macro FLDS_LOADX_4 src, stride, stride2, stride3, out0, out1, out2, out3
fld.s \out0, \src, 0
fldx.s \out1, \src, \stride
fldx.s \out2, \src, \stride2
fldx.s \out3, \src, \stride3
.endm
.macro FLDD_LOADX_4 src, stride, stride2, stride3, out0, out1, out2, out3
fld.d \out0, \src, 0
fldx.d \out1, \src, \stride
fldx.d \out2, \src, \stride2
fldx.d \out3, \src, \stride3
.endm
.macro LSX_LOADX_4 src, stride, stride2, stride3, out0, out1, out2, out3
vld \out0, \src, 0
vldx \out1, \src, \stride
vldx \out2, \src, \stride2
vldx \out3, \src, \stride3
.endm
.macro LASX_LOADX_4 src, stride, stride2, stride3, out0, out1, out2, out3
xvld \out0, \src, 0
xvldx \out1, \src, \stride
xvldx \out2, \src, \stride2
xvldx \out3, \src, \stride3
.endm
/*
* Description : Transpose 4x4 block with half-word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3
* Outputs - out0, out1, out2, out3
*/
.macro LSX_TRANSPOSE4x4_H in0, in1, in2, in3, out0, out1, out2, out3, \
tmp0, tmp1
vilvl.h \tmp0, \in1, \in0
vilvl.h \tmp1, \in3, \in2
vilvl.w \out0, \tmp1, \tmp0
vilvh.w \out2, \tmp1, \tmp0
vilvh.d \out1, \out0, \out0
vilvh.d \out3, \out0, \out2
.endm
/*
* Description : Transpose 4x4 block with word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3
* Outputs - out0, out1, out2, out3
* Details :
* Example :
* 1, 2, 3, 4 1, 5, 9,13
* 5, 6, 7, 8 to 2, 6,10,14
* 9,10,11,12 =====> 3, 7,11,15
* 13,14,15,16 4, 8,12,16
*/
.macro LSX_TRANSPOSE4x4_W in0, in1, in2, in3, out0, out1, out2, out3, \
tmp0, tmp1
vilvl.w \tmp0, \in1, \in0
vilvh.w \out1, \in1, \in0
vilvl.w \tmp1, \in3, \in2
vilvh.w \out3, \in3, \in2
vilvl.d \out0, \tmp1, \tmp0
vilvl.d \out2, \out3, \out1
vilvh.d \out3, \out3, \out1
vilvh.d \out1, \tmp1, \tmp0
.endm
/*
* Description : Transpose 8x8 block with half-word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7
* Outputs - out0, out1, out2, out3, out4, out5, out6, out7
*/
.macro LSX_TRANSPOSE8x8_H in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, \
out2, out3, out4, out5, out6, out7, tmp0, tmp1, tmp2, \
tmp3, tmp4, tmp5, tmp6, tmp7
vilvl.h \tmp0, \in6, \in4
vilvl.h \tmp1, \in7, \in5
vilvl.h \tmp2, \in2, \in0
vilvl.h \tmp3, \in3, \in1
vilvl.h \tmp4, \tmp1, \tmp0
vilvh.h \tmp5, \tmp1, \tmp0
vilvl.h \tmp6, \tmp3, \tmp2
vilvh.h \tmp7, \tmp3, \tmp2
vilvh.h \tmp0, \in6, \in4
vilvh.h \tmp1, \in7, \in5
vilvh.h \tmp2, \in2, \in0
vilvh.h \tmp3, \in3, \in1
vpickev.d \out0, \tmp4, \tmp6
vpickod.d \out1, \tmp4, \tmp6
vpickev.d \out2, \tmp5, \tmp7
vpickod.d \out3, \tmp5, \tmp7
vilvl.h \tmp4, \tmp1, \tmp0
vilvh.h \tmp5, \tmp1, \tmp0
vilvl.h \tmp6, \tmp3, \tmp2
vilvh.h \tmp7, \tmp3, \tmp2
vpickev.d \out4, \tmp4, \tmp6
vpickod.d \out5, \tmp4, \tmp6
vpickev.d \out6, \tmp5, \tmp7
vpickod.d \out7, \tmp5, \tmp7
.endm
/*
* Description : Transpose 16x8 block with byte elements in vectors
* Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7
* Outputs - out0, out1, out2, out3, out4, out5, out6, out7
*/
.macro LASX_TRANSPOSE16X8_B in0, in1, in2, in3, in4, in5, in6, in7, \
in8, in9, in10, in11, in12, in13, in14, in15, \
out0, out1, out2, out3, out4, out5, out6, out7,\
tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7
xvilvl.b \tmp0, \in2, \in0
xvilvl.b \tmp1, \in3, \in1
xvilvl.b \tmp2, \in6, \in4
xvilvl.b \tmp3, \in7, \in5
xvilvl.b \tmp4, \in10, \in8
xvilvl.b \tmp5, \in11, \in9
xvilvl.b \tmp6, \in14, \in12
xvilvl.b \tmp7, \in15, \in13
xvilvl.b \out0, \tmp1, \tmp0
xvilvh.b \out1, \tmp1, \tmp0
xvilvl.b \out2, \tmp3, \tmp2
xvilvh.b \out3, \tmp3, \tmp2
xvilvl.b \out4, \tmp5, \tmp4
xvilvh.b \out5, \tmp5, \tmp4
xvilvl.b \out6, \tmp7, \tmp6
xvilvh.b \out7, \tmp7, \tmp6
xvilvl.w \tmp0, \out2, \out0
xvilvh.w \tmp2, \out2, \out0
xvilvl.w \tmp4, \out3, \out1
xvilvh.w \tmp6, \out3, \out1
xvilvl.w \tmp1, \out6, \out4
xvilvh.w \tmp3, \out6, \out4
xvilvl.w \tmp5, \out7, \out5
xvilvh.w \tmp7, \out7, \out5
xvilvl.d \out0, \tmp1, \tmp0
xvilvh.d \out1, \tmp1, \tmp0
xvilvl.d \out2, \tmp3, \tmp2
xvilvh.d \out3, \tmp3, \tmp2
xvilvl.d \out4, \tmp5, \tmp4
xvilvh.d \out5, \tmp5, \tmp4
xvilvl.d \out6, \tmp7, \tmp6
xvilvh.d \out7, \tmp7, \tmp6
.endm
/*
* Description : Transpose 4x4 block with half-word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3
* Outputs - out0, out1, out2, out3
*/
.macro LASX_TRANSPOSE4x4_H in0, in1, in2, in3, out0, out1, out2, out3, \
tmp0, tmp1
xvilvl.h \tmp0, \in1, \in0
xvilvl.h \tmp1, \in3, \in2
xvilvl.w \out0, \tmp1, \tmp0
xvilvh.w \out2, \tmp1, \tmp0
xvilvh.d \out1, \out0, \out0
xvilvh.d \out3, \out0, \out2
.endm
/*
* Description : Transpose 4x8 block with half-word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3
* Outputs - out0, out1, out2, out3
*/
.macro LASX_TRANSPOSE4x8_H in0, in1, in2, in3, out0, out1, out2, out3, \
tmp0, tmp1
xvilvl.h \tmp0, \in2, \in0
xvilvl.h \tmp1, \in3, \in1
xvilvl.h \out2, \tmp1, \tmp0
xvilvh.h \out3, \tmp1, \tmp0
xvilvl.d \out0, \out2, \out2
xvilvh.d \out1, \out2, \out2
xvilvl.d \out2, \out3, \out3
xvilvh.d \out3, \out3, \out3
.endm
/*
* Description : Transpose 8x8 block with half-word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7
* Outputs - out0, out1, out2, out3, out4, out5, out6, out7
*/
.macro LASX_TRANSPOSE8x8_H in0, in1, in2, in3, in4, in5, in6, in7, \
out0, out1, out2, out3, out4, out5, out6, out7, \
tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7
xvilvl.h \tmp0, \in6, \in4
xvilvl.h \tmp1, \in7, \in5
xvilvl.h \tmp2, \in2, \in0
xvilvl.h \tmp3, \in3, \in1
xvilvl.h \tmp4, \tmp1, \tmp0
xvilvh.h \tmp5, \tmp1, \tmp0
xvilvl.h \tmp6, \tmp3, \tmp2
xvilvh.h \tmp7, \tmp3, \tmp2
xvilvh.h \tmp0, \in6, \in4
xvilvh.h \tmp1, \in7, \in5
xvilvh.h \tmp2, \in2, \in0
xvilvh.h \tmp3, \in3, \in1
xvpickev.d \out0, \tmp4, \tmp6
xvpickod.d \out1, \tmp4, \tmp6
xvpickev.d \out2, \tmp5, \tmp7
xvpickod.d \out3, \tmp5, \tmp7
xvilvl.h \tmp4, \tmp1, \tmp0
xvilvh.h \tmp5, \tmp1, \tmp0
xvilvl.h \tmp6, \tmp3, \tmp2
xvilvh.h \tmp7, \tmp3, \tmp2
xvpickev.d \out4, \tmp4, \tmp6
xvpickod.d \out5, \tmp4, \tmp6
xvpickev.d \out6, \tmp5, \tmp7
xvpickod.d \out7, \tmp5, \tmp7
.endm
/*
* Description : Transpose 2x4x4 block with half-word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3
* Outputs - out0, out1, out2, out3
*/
.macro LASX_TRANSPOSE2x4x4_H in0, in1, in2, in3, out0, out1, out2, out3, \
tmp0, tmp1, tmp2
xvilvh.h \tmp1, \in0, \in1
xvilvl.h \out1, \in0, \in1
xvilvh.h \tmp0, \in2, \in3
xvilvl.h \out3, \in2, \in3
xvilvh.w \tmp2, \out3, \out1
xvilvl.w \out3, \out3, \out1
xvilvl.w \out2, \tmp0, \tmp1
xvilvh.w \tmp1, \tmp0, \tmp1
xvilvh.d \out0, \out2, \out3
xvilvl.d \out2, \out2, \out3
xvilvh.d \out1, \tmp1, \tmp2
xvilvl.d \out3, \tmp1, \tmp2
.endm
/*
* Description : Transpose 4x4 block with word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3
* Outputs - out0, out1, out2, out3
* Details :
* Example :
* 1, 2, 3, 4, 1, 2, 3, 4 1,5, 9,13, 1,5, 9,13
* 5, 6, 7, 8, 5, 6, 7, 8 to 2,6,10,14, 2,6,10,14
* 9,10,11,12, 9,10,11,12 =====> 3,7,11,15, 3,7,11,15
* 13,14,15,16, 13,14,15,16 4,8,12,16, 4,8,12,16
*/
.macro LASX_TRANSPOSE4x4_W in0, in1, in2, in3, out0, out1, out2, out3, \
tmp0, tmp1
xvilvl.w \tmp0, \in1, \in0
xvilvh.w \out1, \in1, \in0
xvilvl.w \tmp1, \in3, \in2
xvilvh.w \out3, \in3, \in2
xvilvl.d \out0, \tmp1, \tmp0
xvilvl.d \out2, \out3, \out1
xvilvh.d \out3, \out3, \out1
xvilvh.d \out1, \tmp1, \tmp0
.endm
/*
* Description : Transpose 8x8 block with word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7
* Outputs - out0, out1, out2, out3, out4, out5, out6,
* _out7
* Example : LASX_TRANSPOSE8x8_W
* in0 : 1,2,3,4,5,6,7,8
* in1 : 2,2,3,4,5,6,7,8
* in2 : 3,2,3,4,5,6,7,8
* in3 : 4,2,3,4,5,6,7,8
* in4 : 5,2,3,4,5,6,7,8
* in5 : 6,2,3,4,5,6,7,8
* in6 : 7,2,3,4,5,6,7,8
* in7 : 8,2,3,4,5,6,7,8
*
* out0 : 1,2,3,4,5,6,7,8
* out1 : 2,2,2,2,2,2,2,2
* out2 : 3,3,3,3,3,3,3,3
* out3 : 4,4,4,4,4,4,4,4
* out4 : 5,5,5,5,5,5,5,5
* out5 : 6,6,6,6,6,6,6,6
* out6 : 7,7,7,7,7,7,7,7
* out7 : 8,8,8,8,8,8,8,8
*/
.macro LASX_TRANSPOSE8x8_W in0, in1, in2, in3, in4, in5, in6, in7,\
out0, out1, out2, out3, out4, out5, out6, out7,\
tmp0, tmp1, tmp2, tmp3
xvilvl.w \tmp0, \in2, \in0
xvilvl.w \tmp1, \in3, \in1
xvilvh.w \tmp2, \in2, \in0
xvilvh.w \tmp3, \in3, \in1
xvilvl.w \out0, \tmp1, \tmp0
xvilvh.w \out1, \tmp1, \tmp0
xvilvl.w \out2, \tmp3, \tmp2
xvilvh.w \out3, \tmp3, \tmp2
xvilvl.w \tmp0, \in6, \in4
xvilvl.w \tmp1, \in7, \in5
xvilvh.w \tmp2, \in6, \in4
xvilvh.w \tmp3, \in7, \in5
xvilvl.w \out4, \tmp1, \tmp0
xvilvh.w \out5, \tmp1, \tmp0
xvilvl.w \out6, \tmp3, \tmp2
xvilvh.w \out7, \tmp3, \tmp2
xmov \tmp0, \out0
xmov \tmp1, \out1
xmov \tmp2, \out2
xmov \tmp3, \out3
xvpermi.q \out0, \out4, 0x02
xvpermi.q \out1, \out5, 0x02
xvpermi.q \out2, \out6, 0x02
xvpermi.q \out3, \out7, 0x02
xvpermi.q \out4, \tmp0, 0x31
xvpermi.q \out5, \tmp1, 0x31
xvpermi.q \out6, \tmp2, 0x31
xvpermi.q \out7, \tmp3, 0x31
.endm
/*
* Description : Transpose 4x4 block with double-word elements in vectors
* Arguments : Inputs - in0, in1, in2, in3
* Outputs - out0, out1, out2, out3
* Example : LASX_TRANSPOSE4x4_D
* in0 : 1,2,3,4
* in1 : 1,2,3,4
* in2 : 1,2,3,4
* in3 : 1,2,3,4
*
* out0 : 1,1,1,1
* out1 : 2,2,2,2
* out2 : 3,3,3,3
* out3 : 4,4,4,4
*/
.macro LASX_TRANSPOSE4x4_D in0, in1, in2, in3, out0, out1, out2, out3, \
tmp0, tmp1
xvilvl.d \tmp0, \in1, \in0
xvilvh.d \out1, \in1, \in0
xvilvh.d \tmp1, \in3, \in2
xvilvl.d \out2, \in3, \in2
xvor.v \out0, \tmp0, \tmp0
xvor.v \out3, \tmp1, \tmp1
xvpermi.q \out0, \out2, 0x02
xvpermi.q \out2, \tmp0, 0x31
xvpermi.q \out3, \out1, 0x31
xvpermi.q \out1, \tmp1, 0x02
.endm
/*****************************************************************************
* loongson_util.S: loongson utility macros
*****************************************************************************
* Copyright (C) 2023 x264 project
*
* Authors: Shiyou Yin <yinshiyou-hf@loongson.cn>
* Xiwei Gu <guxiwei-hf@loongson.cn>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
*
* This program is also available under a commercial proprietary license.
* For more information, contact us at licensing@x264.com.
*****************************************************************************/
#define GLUE(a, b) a ## b
#define JOIN(a, b) GLUE(a, b)
/* Set prefix as needed. */
#define ASM_REF JOIN(JOIN(x264_, BIT_DEPTH), _)
#define FENC_STRIDE 16
#define FDEC_STRIDE 32
.macro function_x264 name, align=DEFAULT_ALIGN
.macro endfunc_x264
jirl $r0, $r1, 0x0
.size ASM_REF\name, . - ASM_REF\name
.purgem endfunc_x264
.endm
.text ;
.align \align ;
.globl ASM_REF\name ;
.type ASM_REF\name, @function ;
ASM_REF\name: ;
.endm
This diff is collapsed.
This diff is collapsed.
/*****************************************************************************
* mc.h: loongarch motion compensation
*****************************************************************************
* Copyright (C) 2023 x264 project
*
* Authors: Xiwei Gu <guxiwei-hf@loongson.cn>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
*
* This program is also available under a commercial proprietary license.
* For more information, contact us at licensing@x264.com.
*****************************************************************************/
#ifndef X264_LOONGARCH_MC_H
#define X264_LOONGARCH_MC_H
#define x264_mc_init_loongarch x264_template(mc_init_loongarch)
void x264_mc_init_loongarch( int cpu, x264_mc_functions_t *pf );
#define x264_pixel_avg_16x16_lsx x264_template(pixel_avg_16x16_lsx)
void x264_pixel_avg_16x16_lsx( pixel *, intptr_t, pixel *, intptr_t, pixel *, intptr_t, int );
#define x264_pixel_avg_16x8_lsx x264_template(pixel_avg_16x8_lsx)
void x264_pixel_avg_16x8_lsx( pixel *, intptr_t, pixel *, intptr_t, pixel *, intptr_t, int );
#define x264_pixel_avg_8x16_lsx x264_template(pixel_avg_8x16_lsx)
void x264_pixel_avg_8x16_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_8x8_lsx x264_template(pixel_avg_8x8_lsx)
void x264_pixel_avg_8x8_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_8x4_lsx x264_template(pixel_avg_8x4_lsx)
void x264_pixel_avg_8x4_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_4x16_lsx x264_template(pixel_avg_4x16_lsx)
void x264_pixel_avg_4x16_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_4x8_lsx x264_template(pixel_avg_4x8_lsx)
void x264_pixel_avg_4x8_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_4x4_lsx x264_template(pixel_avg_4x4_lsx)
void x264_pixel_avg_4x4_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_4x2_lsx x264_template(pixel_avg_4x2_lsx)
void x264_pixel_avg_4x2_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg2_w4_lsx x264_template(pixel_avg2_w4_lsx)
void x264_pixel_avg2_w4_lsx ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
#define x264_pixel_avg2_w8_lsx x264_template(pixel_avg2_w8_lsx)
void x264_pixel_avg2_w8_lsx ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
#define x264_pixel_avg2_w16_lsx x264_template(pixel_avg2_w16_lsx)
void x264_pixel_avg2_w16_lsx ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
#define x264_pixel_avg2_w20_lsx x264_template(pixel_avg2_w20_lsx)
void x264_pixel_avg2_w20_lsx ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
#define x264_mc_weight_w20_lsx x264_template(mc_weight_w20_lsx)
void x264_mc_weight_w20_lsx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w20_noden_lsx x264_template(mc_weight_w20_noden_lsx)
void x264_mc_weight_w20_noden_lsx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w16_lsx x264_template(mc_weight_w16_lsx)
void x264_mc_weight_w16_lsx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w16_noden_lsx x264_template(mc_weight_w16_noden_lsx)
void x264_mc_weight_w16_noden_lsx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w8_lsx x264_template(mc_weight_w8_lsx)
void x264_mc_weight_w8_lsx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w8_noden_lsx x264_template(mc_weight_w8_noden_lsx)
void x264_mc_weight_w8_noden_lsx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w4_lsx x264_template(mc_weight_w4_lsx)
void x264_mc_weight_w4_lsx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w4_noden_lsx x264_template(mc_weight_w4_noden_lsx)
void x264_mc_weight_w4_noden_lsx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_copy_w16_lsx x264_template(mc_copy_w16_lsx)
void x264_mc_copy_w16_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_mc_copy_w8_lsx x264_template(mc_copy_w8_lsx)
void x264_mc_copy_w8_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_mc_copy_w4_lsx x264_template(mc_copy_w4_lsx)
void x264_mc_copy_w4_lsx( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_store_interleave_chroma_lsx x264_template(store_interleave_chroma_lsx)
void x264_store_interleave_chroma_lsx( pixel *dst, intptr_t i_dst, pixel *srcu, pixel *srcv, int height );
#define x264_load_deinterleave_chroma_fenc_lsx x264_template(load_deinterleave_chroma_fenc_lsx)
void x264_load_deinterleave_chroma_fenc_lsx( pixel *dst, pixel *src, intptr_t i_src, int height );
#define x264_load_deinterleave_chroma_fdec_lsx x264_template(load_deinterleave_chroma_fdec_lsx)
void x264_load_deinterleave_chroma_fdec_lsx( pixel *dst, pixel *src, intptr_t i_src, int height );
#define x264_plane_copy_interleave_core_lsx x264_template(plane_copy_interleave_core_lsx)
void x264_plane_copy_interleave_core_lsx( pixel *dst, intptr_t i_dst,
pixel *srcu, intptr_t i_srcu,
pixel *srcv, intptr_t i_srcv, int w, int h );
#define x264_plane_copy_deinterleave_lsx x264_template(plane_copy_deinterleave_lsx)
void x264_plane_copy_deinterleave_lsx( pixel *dstu, intptr_t i_dstu,
pixel *dstv, intptr_t i_dstv,
pixel *src, intptr_t i_src, int w, int h );
#define x264_plane_copy_deinterleave_lasx x264_template(plane_copy_deinterleave_lasx)
void x264_plane_copy_deinterleave_lasx( pixel *dstu, intptr_t i_dstu,
pixel *dstv, intptr_t i_dstv,
pixel *src, intptr_t i_src, int w, int h );
#define x264_prefetch_fenc_420_lsx x264_template(prefetch_fenc_420_lsx)
void x264_prefetch_fenc_420_lsx( uint8_t *pix_y, intptr_t stride_y,
uint8_t *pix_uv, intptr_t stride_uv,
int32_t mb_x );
#define x264_prefetch_fenc_422_lsx x264_template(prefetch_fenc_422_lsx)
void x264_prefetch_fenc_422_lsx( uint8_t *pix_y, intptr_t stride_y,
uint8_t *pix_uv, intptr_t stride_uv,
int32_t mb_x );
#define x264_prefetch_ref_lsx x264_template(prefetch_ref_lsx)
void x264_prefetch_ref_lsx( uint8_t *pix, intptr_t stride, int32_t parity );
#define x264_memcpy_aligned_lsx x264_template(memcpy_aligned_lsx)
void *x264_memcpy_aligned_lsx( void *dst, const void *src, size_t n );
#define x264_memzero_aligned_lsx x264_template(memzero_aligned_lsx)
void x264_memzero_aligned_lsx( void *p_dst, size_t n );
#define x264_hpel_filter_lsx x264_template(hpel_filter_lsx)
void x264_hpel_filter_lsx( pixel *, pixel *, pixel *, pixel *, intptr_t, int, int, int16_t * );
#define x264_frame_init_lowres_core_lsx x264_template(frame_init_lowres_core_lsx)
void x264_frame_init_lowres_core_lsx( uint8_t *, uint8_t *, uint8_t *, uint8_t *,
uint8_t *, intptr_t, intptr_t, int, int );
#define x264_pixel_avg_16x8_lasx x264_template(pixel_avg_16x8_lasx)
void x264_pixel_avg_16x8_lasx( pixel *, intptr_t, pixel *, intptr_t, pixel *, intptr_t, int );
#define x264_pixel_avg_8x16_lasx x264_template(pixel_avg_8x16_lasx)
void x264_pixel_avg_8x16_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_8x8_lasx x264_template(pixel_avg_8x8_lasx)
void x264_pixel_avg_8x8_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_8x4_lasx x264_template(pixel_avg_8x4_lasx)
void x264_pixel_avg_8x4_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_4x16_lasx x264_template(pixel_avg_4x16_lasx)
void x264_pixel_avg_4x16_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_4x8_lasx x264_template(pixel_avg_4x8_lasx)
void x264_pixel_avg_4x8_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_4x4_lasx x264_template(pixel_avg_4x4_lasx)
void x264_pixel_avg_4x4_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg_4x2_lasx x264_template(pixel_avg_4x2_lasx)
void x264_pixel_avg_4x2_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_pixel_avg2_w4_lasx x264_template(pixel_avg2_w4_lasx)
void x264_pixel_avg2_w4_lasx ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
#define x264_pixel_avg2_w8_lasx x264_template(pixel_avg2_w8_lasx)
void x264_pixel_avg2_w8_lasx ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
#define x264_pixel_avg2_w16_lasx x264_template(pixel_avg2_w16_lasx)
void x264_pixel_avg2_w16_lasx ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
#define x264_pixel_avg2_w20_lasx x264_template(pixel_avg2_w20_lasx)
void x264_pixel_avg2_w20_lasx ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
#define x264_mc_weight_w20_lasx x264_template(mc_weight_w20_lasx)
void x264_mc_weight_w20_lasx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w20_noden_lasx x264_template(mc_weight_w20_noden_lasx)
void x264_mc_weight_w20_noden_lasx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w16_lasx x264_template(mc_weight_w16_lasx)
void x264_mc_weight_w16_lasx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w16_noden_lasx x264_template(mc_weight_w16_noden_lasx)
void x264_mc_weight_w16_noden_lasx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w8_lasx x264_template(mc_weight_w8_lasx)
void x264_mc_weight_w8_lasx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w8_noden_lasx x264_template(mc_weight_w8_noden_lasx)
void x264_mc_weight_w8_noden_lasx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w4_lasx x264_template(mc_weight_w4_lasx)
void x264_mc_weight_w4_lasx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_weight_w4_noden_lasx x264_template(mc_weight_w4_noden_lasx)
void x264_mc_weight_w4_noden_lasx( pixel *, intptr_t, pixel *, intptr_t, const x264_weight_t *, int );
#define x264_mc_copy_w16_lasx x264_template(mc_copy_w16_lasx)
void x264_mc_copy_w16_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_mc_copy_w8_lasx x264_template(mc_copy_w8_lasx)
void x264_mc_copy_w8_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_mc_copy_w4_lasx x264_template(mc_copy_w4_lasx)
void x264_mc_copy_w4_lasx( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
#define x264_plane_copy_interleave_core_lasx x264_template(plane_copy_interleave_core_lasx)
void x264_plane_copy_interleave_core_lasx( pixel *dst, intptr_t i_dst,
pixel *srcu, intptr_t i_srcu,
pixel *srcv, intptr_t i_srcv, int w, int h );
#define x264_plane_copy_deinterleave_lasx x264_template(plane_copy_deinterleave_lasx)
void x264_plane_copy_deinterleave_lasx( pixel *dstu, intptr_t i_dstu,
pixel *dstv, intptr_t i_dstv,
pixel *src, intptr_t i_src, int w, int h );
#define x264_memzero_aligned_lasx x264_template(memzero_aligned_lasx)
void x264_memzero_aligned_lasx( void *p_dst, size_t n );
#define x264_hpel_filter_lasx x264_template(hpel_filter_lasx)
void x264_hpel_filter_lasx( pixel *, pixel *, pixel *, pixel *, intptr_t, int, int, int16_t * );
#define x264_frame_init_lowres_core_lasx x264_template(frame_init_lowres_core_lasx)
void x264_frame_init_lowres_core_lasx( uint8_t *, uint8_t *, uint8_t *, uint8_t *, uint8_t *,
intptr_t, intptr_t, int, int );
#endif
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.