diff options
author | Lauri Kasanen <cand@gmx.com> | 2019-03-24 15:23:46 +0300 |
---|---|---|
committer | Lauri Kasanen <cand@gmx.com> | 2019-03-31 12:41:33 +0300 |
commit | 9a2db4dc61d0c158719641164abb4f1351a79d14 (patch) | |
tree | fa3ed94340badeae178b9ec9bb1191974b8ea199 /libswscale | |
parent | a6a31ca3d9af907f6d10211af60d0762ee85284e (diff) |
swscale/ppc: VSX-optimize yuv2422_2
./ffmpeg -f lavfi -i yuvtestsrc=duration=1:size=1200x1440 -sws_flags area \
-s 1200x720 -f null -vframes 100 -pix_fmt $i -nostats \
-cpuflags 0 -v error -
5.1x speedup:
yuyv422
19339 UNITS in yuv2packed2, 16384 runs, 0 skips
3718 UNITS in yuv2packed2, 16383 runs, 1 skips
yvyu422
19438 UNITS in yuv2packed2, 16384 runs, 0 skips
3800 UNITS in yuv2packed2, 16380 runs, 4 skips
uyvy422
19128 UNITS in yuv2packed2, 16384 runs, 0 skips
3721 UNITS in yuv2packed2, 16380 runs, 4 skips
Diffstat (limited to 'libswscale')
-rw-r--r-- | libswscale/ppc/swscale_vsx.c | 69 |
1 files changed, 69 insertions, 0 deletions
diff --git a/libswscale/ppc/swscale_vsx.c b/libswscale/ppc/swscale_vsx.c index 0bb82ac742..1c4051b812 100644 --- a/libswscale/ppc/swscale_vsx.c +++ b/libswscale/ppc/swscale_vsx.c @@ -726,6 +726,61 @@ write422(const vector int16_t vy1, const vector int16_t vy2, } } +#define SETUP(x, buf0, buf1, alpha) { \ + x = vec_ld(0, buf0); \ + tmp = vec_mule(x, alpha); \ + tmp2 = vec_mulo(x, alpha); \ + tmp3 = vec_mergeh(tmp, tmp2); \ + tmp4 = vec_mergel(tmp, tmp2); \ +\ + x = vec_ld(0, buf1); \ + tmp = vec_mule(x, alpha); \ + tmp2 = vec_mulo(x, alpha); \ + tmp5 = vec_mergeh(tmp, tmp2); \ + tmp6 = vec_mergel(tmp, tmp2); \ +\ + tmp3 = vec_add(tmp3, tmp5); \ + tmp4 = vec_add(tmp4, tmp6); \ +\ + tmp3 = vec_sra(tmp3, shift19); \ + tmp4 = vec_sra(tmp4, shift19); \ + x = vec_packs(tmp3, tmp4); \ +} + +static av_always_inline void +yuv2422_2_vsx_template(SwsContext *c, const int16_t *buf[2], + const int16_t *ubuf[2], const int16_t *vbuf[2], + const int16_t *abuf[2], uint8_t *dest, int dstW, + int yalpha, int uvalpha, int y, + enum AVPixelFormat target) +{ + const int16_t *buf0 = buf[0], *buf1 = buf[1], + *ubuf0 = ubuf[0], *ubuf1 = ubuf[1], + *vbuf0 = vbuf[0], *vbuf1 = vbuf[1]; + const int16_t yalpha1 = 4096 - yalpha; + const int16_t uvalpha1 = 4096 - uvalpha; + vector int16_t vy1, vy2, vu, vv; + vector int32_t tmp, tmp2, tmp3, tmp4, tmp5, tmp6; + const vector int16_t vyalpha1 = vec_splats(yalpha1); + const vector int16_t vuvalpha1 = vec_splats(uvalpha1); + const vector uint32_t shift19 = vec_splats(19U); + int i; + av_assert2(yalpha <= 4096U); + av_assert2(uvalpha <= 4096U); + + for (i = 0; i < ((dstW + 1) >> 1); i += 8) { + + SETUP(vy1, &buf0[i * 2], &buf1[i * 2], vyalpha1) + SETUP(vy2, &buf0[(i + 4) * 2], &buf1[(i + 4) * 2], vyalpha1) + SETUP(vu, &ubuf0[i], &ubuf1[i], vuvalpha1) + SETUP(vv, &vbuf0[i], &vbuf1[i], vuvalpha1) + + write422(vy1, vy2, vu, vv, &dest[i * 4], target); + } +} + +#undef SETUP + static av_always_inline void yuv2422_1_vsx_template(SwsContext *c, const int16_t *buf0, const int16_t *ubuf[2], const int16_t *vbuf[2], @@ -786,7 +841,18 @@ yuv2422_1_vsx_template(SwsContext *c, const int16_t *buf0, } } +#define YUV2PACKEDWRAPPER2(name, base, ext, fmt) \ +static void name ## ext ## _2_vsx(SwsContext *c, const int16_t *buf[2], \ + const int16_t *ubuf[2], const int16_t *vbuf[2], \ + const int16_t *abuf[2], uint8_t *dest, int dstW, \ + int yalpha, int uvalpha, int y) \ +{ \ + name ## base ## _2_vsx_template(c, buf, ubuf, vbuf, abuf, \ + dest, dstW, yalpha, uvalpha, y, fmt); \ +} + #define YUV2PACKEDWRAPPER(name, base, ext, fmt) \ +YUV2PACKEDWRAPPER2(name, base, ext, fmt) \ static void name ## ext ## _1_vsx(SwsContext *c, const int16_t *buf0, \ const int16_t *ubuf[2], const int16_t *vbuf[2], \ const int16_t *abuf0, uint8_t *dest, int dstW, \ @@ -909,12 +975,15 @@ av_cold void ff_sws_init_swscale_vsx(SwsContext *c) switch (dstFormat) { case AV_PIX_FMT_YUYV422: c->yuv2packed1 = yuv2yuyv422_1_vsx; + c->yuv2packed2 = yuv2yuyv422_2_vsx; break; case AV_PIX_FMT_YVYU422: c->yuv2packed1 = yuv2yvyu422_1_vsx; + c->yuv2packed2 = yuv2yvyu422_2_vsx; break; case AV_PIX_FMT_UYVY422: c->yuv2packed1 = yuv2uyvy422_1_vsx; + c->yuv2packed2 = yuv2uyvy422_2_vsx; break; } } |