# HG changeset patch
# User Vignesh Vijayakumar
# Date 1502347959 -19800
# Thu Aug 10 12:22:39 2017 +0530
# Node ID f489bc0b864c48f557cc40b739e84fe1040e8728
# Parent 7bdf20f62d02f5714c1332695ffa8c7c6a9d8a5a
x86: AVX512 interp_4tap_horiz_pp_32xN
Color Space i444
Size | AVX2 performance | AVX512 performance
------------------------------------------------
32x8 | 23.96x | 31.57x
32x16 | 24.38x | 33.22x
32x24 | 22.41x | 36.92x
32x32 | 21.54x | 34.09x
32x64 | 23.27x | 29.14x
Color Space i422
Size | AVX2 performance | AVX512 performance
------------------------------------------------
32x16 | 25.55x | 33.16x
32x32 | 22.08x | 35.13x
32x48 | 24.01x | 34.53x
32x64 | 23.76x | 35.21x
diff -r 7bdf20f62d02 -r f489bc0b864c source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp Wed Apr 04 16:47:58 2018 -0700
+++ b/source/common/x86/asm-primitives.cpp Thu Aug 10 12:22:39 2017 +0530
@@ -4001,6 +4001,18 @@
p.chroma[X265_CSP_I444].pu[LUMA_64x48].filter_hpp =
PFX(interp_4tap_horiz_pp_64x48_avx512);
p.chroma[X265_CSP_I444].pu[LUMA_64x16].filter_hpp =
PFX(interp_4tap_horiz_pp_64x16_avx512);
+ p.chroma[X265_CSP_I444].pu[LUMA_32x16].filter_hpp =
PFX(interp_4tap_horiz_pp_32x16_avx512);
+ p.chroma[X265_CSP_I444].pu[LUMA_32x64].filter_hpp =
PFX(interp_4tap_horiz_pp_32x64_avx512);
+ p.chroma[X265_CSP_I444].pu[LUMA_32x24].filter_hpp =
PFX(interp_4tap_horiz_pp_32x24_avx512);
+ p.chroma[X265_CSP_I444].pu[LUMA_32x8].filter_hpp =
PFX(interp_4tap_horiz_pp_32x8_avx512);
+
+ p.chroma[X265_CSP_I444].pu[LUMA_32x32].filter_hpp =
PFX(interp_4tap_horiz_pp_32x32_avx512);
+
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_32x16].filter_hpp =
PFX(interp_4tap_horiz_pp_32x16_avx512);
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_32x32].filter_hpp =
PFX(interp_4tap_horiz_pp_32x32_avx512);
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_32x64].filter_hpp =
PFX(interp_4tap_horiz_pp_32x64_avx512);
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_32x48].filter_hpp =
PFX(interp_4tap_horiz_pp_32x48_avx512);
+
}
#endif
}
diff -r 7bdf20f62d02 -r f489bc0b864c source/common/x86/ipfilter8.asm
--- a/source/common/x86/ipfilter8.asm Wed Apr 04 16:47:58 2018 -0700
+++ b/source/common/x86/ipfilter8.asm Thu Aug 10 12:22:39 2017 +0530
@@ -150,6 +150,8 @@
const interp4_horiz_shuf_load2_avx512, times 2 db 8, 9, 10, 11, 9, 10, 11,
12, 10, 11, 12, 13, 11, 12, 13, 14
+const interp4_horiz_shuf_store1_avx512, dd 0 ,8, 1, 9, 4, 12, 5, 13, 2, 10, 3,
11, 6, 14, 7, 15
+
SECTION .text
cextern pb_128
@@ -9867,6 +9869,44 @@
movu [r2], m5
%endmacro
+%macro PROCESS_IPFILTER_CHROMA_PP_32x2_AVX512 0
+ ; register map
+ ; m0 - interpolate coeff
+ ; m1, m2 - shuffle order table
+ ; m3 - constant word 1
+ ; m4 - constant word 512
+ ; m9 - store shuffle order table
+
+ movu ym5, [r0]
+ vinserti32x8 m5, [r0 + 4], 1
+
+ pshufb m6, m5, m2
+ pshufb m5, m5, m1
+ pmaddubsw m5, m0
+ pmaddubsw m6, m0
+ pmaddwd m5, m3
+ pmaddwd m6, m3
+
+ movu ym7, [r0 + r1]
+ vinserti32x8 m7, [r0 + r1 + 4], 1
+
+ pshufb m8, m7, m2
+ pshufb m7, m7, m1
+ pmaddubsw m7, m0
+ pmaddubsw m8, m0
+ pmaddwd m7, m3
+ pmaddwd m8, m3
+
+ packssdw m5, m6
+ packssdw m7, m8
+ pmulhrsw m5, m4
+ pmulhrsw m7, m4
+ packuswb m5, m7
+ vpermd m5, m9, m5
+ movu [r2], ym5
+ vextracti32x8 [r2 + r3], m5, 1
+%endmacro
+
;-------------------------------------------------------------------------------------------------------------
; void interp_4tap_horiz_pp_64xN(pixel *src, intptr_t srcStride, int16_t *dst,
intptr_t dstStride, int coeffIdx
;-------------------------------------------------------------------------------------------------------------
@@ -9902,6 +9942,40 @@
IPFILTER_CHROMA_PP_64xN_AVX512 48
IPFILTER_CHROMA_PP_64xN_AVX512 16
+%macro IPFILTER_CHROMA_PP_32xN_AVX512 1
+INIT_ZMM avx512
+cglobal interp_4tap_horiz_pp_32x%1, 4,6,10
+ mov r4d, r4m
+
+%ifdef PIC
+ lea r5, [tab_ChromaCoeff]
+ vpbroadcastd m0, [r5 + r4 * 4]
+%else
+ vpbroadcastd m0, [tab_ChromaCoeff + r4 * 4]
+%endif
+
+ vbroadcasti32x8 m1, [interp4_horiz_shuf_load1_avx512]
+ vbroadcasti32x8 m2, [interp4_horiz_shuf_load2_avx512]
+ movu m9, [interp4_horiz_shuf_store1_avx512]
+ vbroadcasti32x8 m3, [pw_1]
+ vbroadcasti32x8 m4, [pw_512]
+ dec r0
+
+%rep %1/2 - 1
+ PROCESS_IPFILTER_CHROMA_PP_32x2_AVX512
+ lea r2, [r2 + 2 * r3]
+ lea r0, [r0 + 2 * r1]
+%endrep
+ PROCESS_IPFILTER_CHROMA_PP_32x2_AVX512
+ RET
+%endmacro
+
+ IPFILTER_CHROMA_PP_32xN_AVX512 16
+ IPFILTER_CHROMA_PP_32xN_AVX512 24
+ IPFILTER_CHROMA_PP_32xN_AVX512 8
+ IPFILTER_CHROMA_PP_32xN_AVX512 32
+ IPFILTER_CHROMA_PP_32xN_AVX512 64
+ IPFILTER_CHROMA_PP_32xN_AVX512 48
;-------------------------------------------------------------------------------------------------------------
;ipfilter_chroma_pp_avx512 code end
;-------------------------------------------------------------------------------------------------------------
_______________________________________________
x265-devel mailing list
x265-devel@videolan.org
https://mailman.videolan.org/listinfo/x265-devel
