| Message ID | 20240607173844.140233-1-ramiro.polla@gmail.com |
|---|---|
| State | New |
| Headers | show |
| Series | [FFmpeg-devel] swscale/x86: add avx2 {lum, chr}ConvertRange | expand |
| Context | Check | Description |
|---|---|---|
| andriy/configure_x86 | warning | Failed to apply patch |
On Fri, Jun 7, 2024 at 7:38 PM Ramiro Polla <ramiro.polla@gmail.com> wrote: > > chrRangeFromJpeg_8_c: 49.4 > chrRangeFromJpeg_8_sse4: 15.9 > chrRangeFromJpeg_8_avx2: 22.6 > chrRangeFromJpeg_24_c: 129.4 > chrRangeFromJpeg_24_sse4: 32.1 > chrRangeFromJpeg_24_avx2: 35.1 > chrRangeFromJpeg_128_c: 534.6 > chrRangeFromJpeg_128_sse4: 165.6 > chrRangeFromJpeg_128_avx2: 100.4 > chrRangeFromJpeg_144_c: 735.6 > chrRangeFromJpeg_144_sse4: 185.1 > chrRangeFromJpeg_144_avx2: 109.4 > chrRangeFromJpeg_256_c: 634.6 > chrRangeFromJpeg_256_sse4: 323.6 > chrRangeFromJpeg_256_avx2: 192.6 > chrRangeFromJpeg_512_c: 1242.4 > chrRangeFromJpeg_512_sse4: 662.1 > chrRangeFromJpeg_512_avx2: 409.1 > chrRangeToJpeg_8_c: 39.6 > chrRangeToJpeg_8_sse4: 15.9 > chrRangeToJpeg_8_avx2: 25.4 > chrRangeToJpeg_24_c: 118.9 > chrRangeToJpeg_24_sse4: 32.9 > chrRangeToJpeg_24_avx2: 30.1 > chrRangeToJpeg_128_c: 636.9 > chrRangeToJpeg_128_sse4: 164.4 > chrRangeToJpeg_128_avx2: 96.6 > chrRangeToJpeg_144_c: 716.4 > chrRangeToJpeg_144_sse4: 187.1 > chrRangeToJpeg_144_avx2: 109.4 > chrRangeToJpeg_256_c: 1258.6 > chrRangeToJpeg_256_sse4: 326.1 > chrRangeToJpeg_256_avx2: 193.9 > chrRangeToJpeg_512_c: 2489.4 > chrRangeToJpeg_512_sse4: 662.1 > chrRangeToJpeg_512_avx2: 382.4 > lumRangeFromJpeg_8_c: 13.6 > lumRangeFromJpeg_8_sse4: 14.4 > lumRangeFromJpeg_8_avx2: 19.6 > lumRangeFromJpeg_24_c: 38.9 > lumRangeFromJpeg_24_sse4: 18.9 > lumRangeFromJpeg_24_avx2: 23.9 > lumRangeFromJpeg_128_c: 239.4 > lumRangeFromJpeg_128_sse4: 81.9 > lumRangeFromJpeg_128_avx2: 51.6 > lumRangeFromJpeg_144_c: 285.1 > lumRangeFromJpeg_144_sse4: 92.1 > lumRangeFromJpeg_144_avx2: 59.6 > lumRangeFromJpeg_256_c: 857.1 > lumRangeFromJpeg_256_sse4: 164.4 > lumRangeFromJpeg_256_avx2: 101.9 > lumRangeFromJpeg_512_c: 1028.6 > lumRangeFromJpeg_512_sse4: 335.6 > lumRangeFromJpeg_512_avx2: 201.4 > lumRangeToJpeg_8_c: 20.4 > lumRangeToJpeg_8_sse4: 14.4 > lumRangeToJpeg_8_avx2: 20.4 > lumRangeToJpeg_24_c: 58.1 > lumRangeToJpeg_24_sse4: 18.9 > lumRangeToJpeg_24_avx2: 22.6 > lumRangeToJpeg_128_c: 327.4 > lumRangeToJpeg_128_sse4: 83.4 > lumRangeToJpeg_128_avx2: 53.6 > lumRangeToJpeg_144_c: 375.6 > lumRangeToJpeg_144_sse4: 93.9 > lumRangeToJpeg_144_avx2: 58.9 > lumRangeToJpeg_256_c: 649.6 > lumRangeToJpeg_256_sse4: 162.1 > lumRangeToJpeg_256_avx2: 101.9 > lumRangeToJpeg_512_c: 1289.1 > lumRangeToJpeg_512_sse4: 335.6 > lumRangeToJpeg_512_avx2: 201.4 > --- > libswscale/x86/range_convert.asm | 46 ++++++++++++++++++++++++++------ > libswscale/x86/swscale.c | 5 +++- > 2 files changed, 42 insertions(+), 9 deletions(-) > > diff --git a/libswscale/x86/range_convert.asm b/libswscale/x86/range_convert.asm > index 13983a386b..54c2f64769 100644 > --- a/libswscale/x86/range_convert.asm > +++ b/libswscale/x86/range_convert.asm [...] > @@ -66,10 +66,19 @@ cglobal %1, 2, 3, 3, dst, width, x > paddd m1, m5 > psrad m0, %4 > psrad m1, %4 > +%if mmsize == 16 > packssdw m0, m0 > packssdw m1, m1 > movq [dstq+xq*2], m0 > movq [dstq+xq*2+mmsize/2], m1 > +%else > + vextracti128 xm7, ym0, 1 > + packssdw xm0, xm7 > + vextracti128 xm7, ym1, 1 > + packssdw xm1, xm7 > + movdqu [dstq+xq*2], xm0 > + movdqu [dstq+xq*2+mmsize/2], xm1 > +%endif > add xq, mmsize / 2 > cmp xd, widthd > jl .loop Is there a cleaner way to do this packing in avx2 (or a macro to have the same code as non-avx2)? Also is there some cleaner way to move half the register into memory (instead of having to ifdef between mmsize)?
diff --git a/libswscale/x86/range_convert.asm b/libswscale/x86/range_convert.asm index 13983a386b..54c2f64769 100644 --- a/libswscale/x86/range_convert.asm +++ b/libswscale/x86/range_convert.asm @@ -22,20 +22,20 @@ SECTION_RODATA -chr_to_mult: times 4 dd 4663 -chr_to_offset: times 4 dd -9289992 +chr_to_mult: times 8 dd 4663 +chr_to_offset: times 8 dd -9289992 %define chr_to_shift 12 -chr_from_mult: times 4 dd 1799 -chr_from_offset: times 4 dd 4081085 +chr_from_mult: times 8 dd 1799 +chr_from_offset: times 8 dd 4081085 %define chr_from_shift 11 -lum_to_mult: times 4 dd 19077 -lum_to_offset: times 4 dd -39057361 +lum_to_mult: times 8 dd 19077 +lum_to_offset: times 8 dd -39057361 %define lum_to_shift 14 -lum_from_mult: times 4 dd 14071 -lum_from_offset: times 4 dd 33561947 +lum_from_mult: times 8 dd 14071 +lum_from_offset: times 8 dd 33561947 %define lum_from_shift 14 SECTION .text @@ -66,10 +66,19 @@ cglobal %1, 2, 3, 3, dst, width, x paddd m1, m5 psrad m0, %4 psrad m1, %4 +%if mmsize == 16 packssdw m0, m0 packssdw m1, m1 movq [dstq+xq*2], m0 movq [dstq+xq*2+mmsize/2], m1 +%else + vextracti128 xm7, ym0, 1 + packssdw xm0, xm7 + vextracti128 xm7, ym1, 1 + packssdw xm1, xm7 + movdqu [dstq+xq*2], xm0 + movdqu [dstq+xq*2+mmsize/2], xm1 +%endif add xq, mmsize / 2 cmp xd, widthd jl .loop @@ -107,6 +116,7 @@ cglobal %1, 3, 4, 4, dstU, dstV, width, x psrad m1, %4 psrad m2, %4 psrad m3, %4 +%if mmsize == 16 packssdw m0, m0 packssdw m1, m1 packssdw m2, m2 @@ -115,6 +125,20 @@ cglobal %1, 3, 4, 4, dstU, dstV, width, x movq [dstUq+xq*2+mmsize/2], m1 movq [dstVq+xq*2], m2 movq [dstVq+xq*2+mmsize/2], m3 +%else + vextracti128 xm7, ym0, 1 + packssdw xm0, xm7 + vextracti128 xm7, ym1, 1 + packssdw xm1, xm7 + vextracti128 xm7, ym2, 1 + packssdw xm2, xm7 + vextracti128 xm7, ym3, 1 + packssdw xm3, xm7 + movdqu [dstUq+xq*2], xm0 + movdqu [dstUq+xq*2+mmsize/2], xm1 + movdqu [dstVq+xq*2], xm2 + movdqu [dstVq+xq*2+mmsize/2], xm3 +%endif add xq, mmsize / 2 cmp xd, widthd jl .loop @@ -127,4 +151,10 @@ LUMCONVERTRANGE lumRangeToJpeg, lum_to_mult, lum_to_offset, lum_to_shift CHRCONVERTRANGE chrRangeToJpeg, chr_to_mult, chr_to_offset, chr_to_shift LUMCONVERTRANGE lumRangeFromJpeg, lum_from_mult, lum_from_offset, lum_from_shift CHRCONVERTRANGE chrRangeFromJpeg, chr_from_mult, chr_from_offset, chr_from_shift + +INIT_YMM avx2 +LUMCONVERTRANGE lumRangeToJpeg, lum_to_mult, lum_to_offset, lum_to_shift +CHRCONVERTRANGE chrRangeToJpeg, chr_to_mult, chr_to_offset, chr_to_shift +LUMCONVERTRANGE lumRangeFromJpeg, lum_from_mult, lum_from_offset, lum_from_shift +CHRCONVERTRANGE chrRangeFromJpeg, chr_from_mult, chr_from_offset, chr_from_shift %endif diff --git a/libswscale/x86/swscale.c b/libswscale/x86/swscale.c index c5ddfb5605..df6953550a 100644 --- a/libswscale/x86/swscale.c +++ b/libswscale/x86/swscale.c @@ -467,12 +467,15 @@ void ff_lumRangeToJpeg_ ##opt(int16_t *dst, int width); \ void ff_chrRangeToJpeg_ ##opt(int16_t *dstU, int16_t *dstV, int width); \ RANGE_CONVERT_FUNCS_DECL(sse4); +RANGE_CONVERT_FUNCS_DECL(avx2); av_cold void ff_sws_init_range_convert_x86(SwsContext *c) { if (c->srcRange != c->dstRange && !isAnyRGB(c->dstFormat)) { int cpu_flags = av_get_cpu_flags(); - if (EXTERNAL_SSE4(cpu_flags)) { + if (EXTERNAL_AVX2_FAST(cpu_flags)) { + RANGE_CONVERT_FUNCS(avx2); + } else if (EXTERNAL_SSE4(cpu_flags)) { RANGE_CONVERT_FUNCS(sse4); } }