<div dir="ltr">Okay let me check and send it again.<div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Mar 16, 2015 at 4:32 PM, Deepthi Nandakumar <span dir="ltr"><<a href="mailto:deepthi@multicorewareinc.com" target="_blank">deepthi@multicorewareinc.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">This crashes on my system. Pls check.<br></div><div class="gmail_extra"><br><div class="gmail_quote"><div><div class="h5">On Mon, Mar 16, 2015 at 11:03 AM, <span dir="ltr"><<a href="mailto:aasaipriya@multicorewareinc.com" target="_blank">aasaipriya@multicorewareinc.com</a>></span> wrote:<br></div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5"># HG changeset patch<br>
# User Aasaipriya Chandran <<a href="mailto:aasaipriya@multicorewareinc.com" target="_blank">aasaipriya@multicorewareinc.com</a>><br>
# Date 1426483989 -19800<br>
#Â Â Â Mon Mar 16 11:03:09 2015 +0530<br>
# Node ID b71432c674193a204de05e8957574431eaba3610<br>
# Parent 19f6ba4dd66147698aa96deb2f364f30a7d63a7e<br>
asm: luma_hps[32x8, 32x16, 32x24, 32x32, 32x64] avx2 - improved 6143c->3210c, 9343c->4767c, 12806c->6737c, 16065c->8648c, 29073c->15496c<br>
<br>
diff -r 19f6ba4dd661 -r b71432c67419 source/common/x86/asm-primitives.cpp<br>
--- a/source/common/x86/asm-primitives.cpp   Mon Mar 16 10:47:09 2015 +0530<br>
+++ b/source/common/x86/asm-primitives.cpp   Mon Mar 16 11:03:09 2015 +0530<br>
@@ -1576,6 +1576,12 @@<br>
     p.pu[LUMA_16x32].luma_hps = x265_interp_8tap_horiz_ps_16x32_avx2;<br>
     p.pu[LUMA_16x64].luma_hps = x265_interp_8tap_horiz_ps_16x64_avx2;<br>
<br>
+Â Â Â Â p.pu[LUMA_32x32].luma_hps = x265_interp_8tap_horiz_ps_32x32_avx2;<br>
+Â Â Â Â p.pu[LUMA_32x16].luma_hps = x265_interp_8tap_horiz_ps_32x16_avx2;<br>
+Â Â Â Â p.pu[LUMA_32x24].luma_hps = x265_interp_8tap_horiz_ps_32x24_avx2;<br>
+Â Â Â Â p.pu[LUMA_32x8].luma_hps = x265_interp_8tap_horiz_ps_32x8_avx2;<br>
+Â Â Â Â p.pu[LUMA_32x64].luma_hps = x265_interp_8tap_horiz_ps_32x64_avx2;<br>
+<br>
     p.chroma[X265_CSP_I420].pu[CHROMA_420_8x8].filter_hpp = x265_interp_4tap_horiz_pp_8x8_avx2;<br>
     p.chroma[X265_CSP_I420].pu[CHROMA_420_4x4].filter_hpp = x265_interp_4tap_horiz_pp_4x4_avx2;<br>
     p.chroma[X265_CSP_I420].pu[CHROMA_420_32x32].filter_hpp = x265_interp_4tap_horiz_pp_32x32_avx2;<br>
diff -r 19f6ba4dd661 -r b71432c67419 source/common/x86/ipfilter8.asm<br>
--- a/source/common/x86/ipfilter8.asm  Mon Mar 16 10:47:09 2015 +0530<br>
+++ b/source/common/x86/ipfilter8.asm  Mon Mar 16 11:03:09 2015 +0530<br>
@@ -2038,7 +2038,101 @@<br>
 IPFILTER_LUMA_PS_16x_AVX2 16 , 32<br>
 IPFILTER_LUMA_PS_16x_AVX2 16 , 64<br>
<br>
-<br>
+%macro IPFILTER_LUMA_PS_32xN_AVX2 2<br>
+INIT_YMM avx2<br>
+cglobal interp_8tap_horiz_ps_%1x%2, 6, 7, 8<br>
+%ifdef PIC<br>
+  lea             r6,        [tab_LumaCoeff]<br>
+  vpbroadcastq        m0,        [r6 + r4 * 8]<br>
+%else<br>
+  vpbroadcastq        m0,        [tab_LumaCoeff + r4 * 8]<br>
+%endif<br>
+  mova            m6,        [tab_Lm + 32]<br>
+  mova            m1,        [tab_Lm]<br>
+  mov             r4d,        %2              ;height<br>
+  add             r3d,        r3d<br>
+  vbroadcasti128       m2,        [pw_1]<br>
+  mova            m7,        [interp8_hps_shuf]<br>
+<br>
+Â Â ; register map<br>
+Â Â ; m0Â Â Â - interpolate coeff<br>
+Â Â ; m1 , m6 - shuffle order table<br>
+Â Â ; m2Â Â Â - pw_1<br>
+<br>
+  sub             r0,        3<br>
+  test            r5d,        r5d<br>
+  jz             .loop<br>
+  lea             r6,        [r1 * 3]           ; r8 = (N / 2 - 1) * srcStride<br>
+  sub             r0,        r6<br>
+  add             r4d,        7<br>
+<br>
+.loop<br>
+Â Â ; Row 0<br>
+  vbroadcasti128       m3,        [r0]             ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]<br>
+  pshufb           m4,        m3,       m6      ; row 0 (col 4 to 7)<br>
+  pshufb           m3,        m1              ; shuffled based on the col order tab_Lm row 0 (col 0 to 3)<br>
+  pmaddubsw          m3,        m0<br>
+  pmaddubsw          m4,        m0<br>
+  pmaddwd           m3,        m2<br>
+  pmaddwd           m4,        m2<br>
+  packssdw          m3,        m4<br>
+<br>
+<br>
+  vbroadcasti128       m4,        [r0 + 8]<br>
+  pshufb           m5,        m4,       m6      ;row 0 (col 12 to 15)<br>
+  pshufb           m4,        m1              ;row 0 (col 8 to 11)<br>
+  pmaddubsw          m4,        m0<br>
+  pmaddubsw          m5,        m0<br>
+  pmaddwd           m4,        m2<br>
+  pmaddwd           m5,        m2<br>
+  packssdw          m4,        m5<br>
+<br>
+  pmaddwd           m3,        m2<br>
+  pmaddwd           m4,        m2<br>
+  packssdw          m3,        m4<br>
+  vpermd           m3,        m7,       m3<br>
+  psubw            m3,        [pw_2000]<br>
+<br>
+  movu            [r2],       m3             ;row 0<br>
+<br>
+  vbroadcasti128       m3,        [r0 + 16]<br>
+  pshufb           m4,        m3,       m6      ; row 0 (col 20 to 23)<br>
+  pshufb           m3,        m1              ; row 0 (col 16 to 19)<br>
+  pmaddubsw          m3,        m0<br>
+  pmaddubsw          m4,        m0<br>
+  pmaddwd           m3,        m2<br>
+  pmaddwd           m4,        m2<br>
+  packssdw          m3,        m4<br>
+<br>
+  vbroadcasti128       m4,        [r0 + 24]<br>
+  pshufb           m5,        m4,       m6      ;row 0 (col 28 to 31)<br>
+  pshufb           m4,        m1              ;row 0 (col 24 to 27)<br>
+  pmaddubsw          m4,        m0<br>
+  pmaddubsw          m5,        m0<br>
+  pmaddwd           m4,        m2<br>
+  pmaddwd           m5,        m2<br>
+  packssdw          m4,        m5<br>
+<br>
+  pmaddwd           m3,        m2<br>
+  pmaddwd           m4,        m2<br>
+  packssdw          m3,        m4<br>
+  vpermd           m3,        m7,       m3<br>
+  psubw            m3,        [pw_2000]<br>
+<br>
+  movu            [r2 + 32],     m3             ;row 0<br>
+<br>
+  add             r0,        r1<br>
+  add             r2,        r3<br>
+  dec             r4d<br>
+  jnz             .loop<br>
+Â Â RET<br>
+%endmacro<br>
+<br>
+IPFILTER_LUMA_PS_32xN_AVX2 32 , 32<br>
+IPFILTER_LUMA_PS_32xN_AVX2 32 , 16<br>
+IPFILTER_LUMA_PS_32xN_AVX2 32 , 24<br>
+IPFILTER_LUMA_PS_32xN_AVX2 32 , 8<br>
+IPFILTER_LUMA_PS_32xN_AVX2 32 , 64<br>
 ;--------------------------------------------------------------------------------------------------------------<br>
 ; void interp_8tap_horiz_pp_%1x%2(pixel *src, intptr_t srcStride, pixel *dst, intptr_t dstStride, int coeffIdx)<br>
 ;--------------------------------------------------------------------------------------------------------------<br></div></div>
_______________________________________________<br>
x265-devel mailing list<br>
<a href="mailto:x265-devel@videolan.org" target="_blank">x265-devel@videolan.org</a><br>
<a href="https://mailman.videolan.org/listinfo/x265-devel" target="_blank">https://mailman.videolan.org/listinfo/x265-devel</a><br>
</blockquote></div><br></div>
<br>_______________________________________________<br>
x265-devel mailing list<br>
<a href="mailto:x265-devel@videolan.org">x265-devel@videolan.org</a><br>
<a href="https://mailman.videolan.org/listinfo/x265-devel" target="_blank">https://mailman.videolan.org/listinfo/x265-devel</a><br>
<br></blockquote></div><br></div>