Mercurial > x265
changeset 9513:bc19307b799c draft
asm-avx2: luma_vpp[4x16], luma_vps[4x16]: improve 1510c->756c, 1295c->728c
| author | Divya Manivannan <divya@multicorewareinc.com> |
|---|---|
| date | Fri, 13 Feb 2015 18:07:46 +0530 |
| parents | a2ff2a89f136 |
| children | c98d0dccbb36 |
| files | source/common/x86/asm-primitives.cpp source/common/x86/ipfilter16.asm |
| diffstat | 2 files changed, 241 insertions(+-), 0 deletions(-) [+] |
line wrap: on
line diff
--- a/source/common/x86/asm-primitives.cpp Fri Feb 13 17:24:51 2015 +0530 +++ b/source/common/x86/asm-primitives.cpp Fri Feb 13 18:07:46 2015 +0530 @@ -1144,11 +1144,13 @@ void setupAssemblyPrimitives(EncoderPrim p.pu[LUMA_4x4].luma_vpp = x265_interp_8tap_vert_pp_4x4_avx2; p.pu[LUMA_4x8].luma_vpp = x265_interp_8tap_vert_pp_4x8_avx2; + p.pu[LUMA_4x16].luma_vpp = x265_interp_8tap_vert_pp_4x16_avx2; p.pu[LUMA_8x4].luma_vpp = x265_interp_8tap_vert_pp_8x4_avx2; p.pu[LUMA_16x4].luma_vpp = x265_interp_8tap_vert_pp_16x4_avx2; p.pu[LUMA_4x4].luma_vps = x265_interp_8tap_vert_ps_4x4_avx2; p.pu[LUMA_4x8].luma_vps = x265_interp_8tap_vert_ps_4x8_avx2; + p.pu[LUMA_4x16].luma_vps = x265_interp_8tap_vert_ps_4x16_avx2; p.pu[LUMA_8x4].luma_vps = x265_interp_8tap_vert_ps_8x4_avx2; p.pu[LUMA_16x4].luma_vps = x265_interp_8tap_vert_ps_16x4_avx2; }
--- a/source/common/x86/ipfilter16.asm Fri Feb 13 17:24:51 2015 +0530 +++ b/source/common/x86/ipfilter16.asm Fri Feb 13 18:07:46 2015 +0530 @@ -4576,6 +4576,245 @@ cglobal interp_8tap_vert_%1_4x8, 4, 7, 8 FILTER_VER_LUMA_AVX2_4x8 pp FILTER_VER_LUMA_AVX2_4x8 ps +%macro PROCESS_LUMA_AVX2_W4_16R 1 + movq xm0, [r0] + movq xm1, [r0 + r1] + punpcklwd xm0, xm1 + movq xm2, [r0 + r1 * 2] + punpcklwd xm1, xm2 + vinserti128 m0, m0, xm1, 1 ; m0 = [2 1 1 0] + pmaddwd m0, [r5] + movq xm3, [r0 + r4] + punpcklwd xm2, xm3 + lea r0, [r0 + 4 * r1] + movq xm4, [r0] + punpcklwd xm3, xm4 + vinserti128 m2, m2, xm3, 1 ; m2 = [4 3 3 2] + pmaddwd m5, m2, [r5 + 1 * mmsize] + pmaddwd m2, [r5] + paddd m0, m5 + movq xm3, [r0 + r1] + punpcklwd xm4, xm3 + movq xm1, [r0 + r1 * 2] + punpcklwd xm3, xm1 + vinserti128 m4, m4, xm3, 1 ; m4 = [6 5 5 4] + pmaddwd m5, m4, [r5 + 2 * mmsize] + paddd m0, m5 + pmaddwd m5, m4, [r5 + 1 * mmsize] + paddd m2, m5 + pmaddwd m4, [r5] + movq xm3, [r0 + r4] + punpcklwd xm1, xm3 + lea r0, [r0 + 4 * r1] + movq xm6, [r0] + punpcklwd xm3, xm6 + vinserti128 m1, m1, xm3, 1 ; m1 = [8 7 7 6] + pmaddwd m5, m1, [r5 + 3 * mmsize] + paddd m0, m5 + pmaddwd m5, m1, [r5 + 2 * mmsize] + paddd m2, m5 + pmaddwd m5, m1, [r5 + 1 * mmsize] + paddd m4, m5 + pmaddwd m1, [r5] + movq xm3, [r0 + r1] + punpcklwd xm6, xm3 + movq xm5, [r0 + 2 * r1] + punpcklwd xm3, xm5 + vinserti128 m6, m6, xm3, 1 ; m6 = [10 9 9 8] + pmaddwd m3, m6, [r5 + 3 * mmsize] + paddd m2, m3 + pmaddwd m3, m6, [r5 + 2 * mmsize] + paddd m4, m3 + pmaddwd m3, m6, [r5 + 1 * mmsize] + paddd m1, m3 + pmaddwd m6, [r5] + + paddd m0, m7 + paddd m2, m7 +%ifidn %1,pp + psrad m0, 6 + psrad m2, 6 +%else + psrad m0, 2 + psrad m2, 2 +%endif + packssdw m0, m2 +%ifidn %1,pp + pxor m3, m3 + CLIPW m0, m3, [pw_pixel_max] +%endif + + vextracti128 xm2, m0, 1 + movq [r2], xm0 + movq [r2 + r3], xm2 + movhps [r2 + r3 * 2], xm0 + movhps [r2 + r6], xm2 + + movq xm2, [r0 + r4] + punpcklwd xm5, xm2 + lea r0, [r0 + 4 * r1] + movq xm0, [r0] + punpcklwd xm2, xm0 + vinserti128 m5, m5, xm2, 1 ; m5 = [12 11 11 10] + pmaddwd m2, m5, [r5 + 3 * mmsize] + paddd m4, m2 + pmaddwd m2, m5, [r5 + 2 * mmsize] + paddd m1, m2 + pmaddwd m2, m5, [r5 + 1 * mmsize] + paddd m6, m2 + pmaddwd m5, [r5] + movq xm2, [r0 + r1] + punpcklwd xm0, xm2 + movq xm3, [r0 + 2 * r1] + punpcklwd xm2, xm3 + vinserti128 m0, m0, xm2, 1 ; m0 = [14 13 13 12] + pmaddwd m2, m0, [r5 + 3 * mmsize] + paddd m1, m2 + pmaddwd m2, m0, [r5 + 2 * mmsize] + paddd m6, m2 + pmaddwd m2, m0, [r5 + 1 * mmsize] + paddd m5, m2 + pmaddwd m0, [r5] + + paddd m4, m7 + paddd m1, m7 +%ifidn %1,pp + psrad m4, 6 + psrad m1, 6 +%else + psrad m4, 2 + psrad m1, 2 +%endif + packssdw m4, m1 +%ifidn %1,pp + pxor m2, m2 + CLIPW m4, m2, [pw_pixel_max] +%endif + + vextracti128 xm1, m4, 1 + lea r2, [r2 + r3 * 4] + movq [r2], xm4 + movq [r2 + r3], xm1 + movhps [r2 + r3 * 2], xm4 + movhps [r2 + r6], xm1 + + movq xm4, [r0 + r4] + punpcklwd xm3, xm4 + lea r0, [r0 + 4 * r1] + movq xm1, [r0] + punpcklwd xm4, xm1 + vinserti128 m3, m3, xm4, 1 ; m3 = [16 15 15 14] + pmaddwd m4, m3, [r5 + 3 * mmsize] + paddd m6, m4 + pmaddwd m4, m3, [r5 + 2 * mmsize] + paddd m5, m4 + pmaddwd m4, m3, [r5 + 1 * mmsize] + paddd m0, m4 + pmaddwd m3, [r5] + movq xm4, [r0 + r1] + punpcklwd xm1, xm4 + movq xm2, [r0 + 2 * r1] + punpcklwd xm4, xm2 + vinserti128 m1, m1, xm4, 1 ; m1 = [18 17 17 16] + pmaddwd m4, m1, [r5 + 3 * mmsize] + paddd m5, m4 + pmaddwd m4, m1, [r5 + 2 * mmsize] + paddd m0, m4 + pmaddwd m1, [r5 + 1 * mmsize] + paddd m3, m1 + + paddd m6, m7 + paddd m5, m7 +%ifidn %1,pp + psrad m6, 6 + psrad m5, 6 +%else + psrad m6, 2 + psrad m5, 2 +%endif + packssdw m6, m5 +%ifidn %1,pp + pxor m1, m1 + CLIPW m6, m1, [pw_pixel_max] +%endif + + vextracti128 xm5, m6, 1 + lea r2, [r2 + r3 * 4] + movq [r2], xm6 + movq [r2 + r3], xm5 + movhps [r2 + r3 * 2], xm6 + movhps [r2 + r6], xm5 + + movq xm4, [r0 + r4] + punpcklwd xm2, xm4 + lea r0, [r0 + 4 * r1] + movq xm6, [r0] + punpcklwd xm4, xm6 + vinserti128 m2, m2, xm4, 1 ; m2 = [20 19 19 18] + pmaddwd m4, m2, [r5 + 3 * mmsize] + paddd m0, m4 + pmaddwd m2, [r5 + 2 * mmsize] + paddd m3, m2 + movq xm4, [r0 + r1] + punpcklwd xm6, xm4 + movq xm2, [r0 + 2 * r1] + punpcklwd xm4, xm2 + vinserti128 m6, m6, xm4, 1 ; m6 = [22 21 21 20] + pmaddwd m6, [r5 + 3 * mmsize] + paddd m3, m6 + + paddd m0, m7 + paddd m3, m7 +%ifidn %1,pp + psrad m0, 6 + psrad m3, 6 +%else + psrad m0, 2 + psrad m3, 2 +%endif + packssdw m0, m3 +%ifidn %1,pp + CLIPW m0, m1, [pw_pixel_max] +%endif + + vextracti128 xm3, m0, 1 + lea r2, [r2 + r3 * 4] + movq [r2], xm0 + movq [r2 + r3], xm3 + movhps [r2 + r3 * 2], xm0 + movhps [r2 + r6], xm3 +%endmacro + +%macro FILTER_VER_LUMA_AVX2_4x16 1 +INIT_YMM avx2 +cglobal interp_8tap_vert_%1_4x16, 4, 7, 8 + mov r4d, r4m + shl r4d, 7 + add r1d, r1d + add r3d, r3d + +%ifdef PIC + lea r5, [tab_LumaCoeffVer] + add r5, r4 +%else + lea r5, [tab_LumaCoeffVer + r4] +%endif + + lea r4, [r1 * 3] + sub r0, r4 +%ifidn %1,pp + vbroadcasti128 m7, [pd_32] +%else + vbroadcasti128 m7, [pd_n32768] +%endif + lea r6, [r3 * 3] + PROCESS_LUMA_AVX2_W4_16R %1 + RET +%endmacro + +FILTER_VER_LUMA_AVX2_4x16 pp +FILTER_VER_LUMA_AVX2_4x16 ps + ;--------------------------------------------------------------------------------------------------------------- ; void interp_8tap_vert_ps_%1x%2(pixel *src, intptr_t srcStride, int16_t *dst, intptr_t dstStride, int coeffIdx) ;---------------------------------------------------------------------------------------------------------------