[x265] [PATCH] pixel_add_ps_16x4, asm code
praveen at multicorewareinc.com
praveen at multicorewareinc.com
Thu Nov 21 09:10:15 CET 2013
# HG changeset patch
# User Praveen Tiwari
# Date 1385021399 -19800
# Node ID fe475a8af3408576f07ece8dab7b7814bc393b50
# Parent db1151bb4974f1288745ba39dfd6e1838113feb7
pixel_add_ps_16x4, asm code
diff -r db1151bb4974 -r fe475a8af340 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp Wed Nov 20 18:36:04 2013 -0600
+++ b/source/common/x86/asm-primitives.cpp Thu Nov 21 13:39:59 2013 +0530
@@ -645,6 +645,7 @@
p.chroma[X265_CSP_I420].add_ps[CHROMA_8x8] = x265_pixel_add_ps_8x8_sse4;
p.chroma[X265_CSP_I420].add_ps[CHROMA_8x16] = x265_pixel_add_ps_8x16_sse4;
p.chroma[X265_CSP_I420].add_ps[CHROMA_8x32] = x265_pixel_add_ps_8x32_sse4;
+ p.chroma[X265_CSP_I420].add_ps[CHROMA_16x4] = x265_pixel_add_ps_16x4_sse4;
p.chroma[X265_CSP_I420].filter_vsp[CHROMA_2x4] = x265_interp_4tap_vert_sp_2x4_sse4;
p.chroma[X265_CSP_I420].filter_vsp[CHROMA_2x8] = x265_interp_4tap_vert_sp_2x8_sse4;
diff -r db1151bb4974 -r fe475a8af340 source/common/x86/pixeladd8.asm
--- a/source/common/x86/pixeladd8.asm Wed Nov 20 18:36:04 2013 -0600
+++ b/source/common/x86/pixeladd8.asm Thu Nov 21 13:39:59 2013 +0530
@@ -361,3 +361,69 @@
PIXEL_ADD_PS_W8_H4 8, 8
PIXEL_ADD_PS_W8_H4 8, 16
PIXEL_ADD_PS_W8_H4 8, 32
+
+;-----------------------------------------------------------------------------
+; void pixel_add_ps_16x4(pixel *dest, intptr_t destride, pixel *src0, int16_t *scr1, intptr_t srcStride0, intptr_t srcStride1)
+;-----------------------------------------------------------------------------
+INIT_XMM sse4
+cglobal pixel_add_ps_16x4, 6, 6, 4, dest, destride, src0, scr1, srcStride0, srcStride1
+
+add r5, r5
+
+pmovzxbw m0, [r2]
+pmovzxbw m1, [r2 + 8]
+
+movu m2, [r3]
+movu m3, [r3 + 16]
+
+paddw m0, m2
+paddw m1, m3
+
+packuswb m0, m1
+
+movu [r0], m0
+
+pmovzxbw m0, [r2 + r4]
+pmovzxbw m1, [r2 + r4 + 8]
+
+movu m2, [r3 + r5]
+movu m3, [r3 + r5 + 16]
+
+paddw m0, m2
+paddw m1, m3
+
+packuswb m0, m1
+
+movu [r0 + r1], m0
+
+pmovzxbw m0, [r2 + 2 * r4]
+pmovzxbw m1, [r2 + 2 * r4 + 8]
+
+movu m2, [r3 + 2 * r5]
+movu m3, [r3 + 2 * r5 + 16]
+
+paddw m0, m2
+paddw m1, m3
+
+packuswb m0, m1
+
+movu [r0 + 2 * r1], m0
+
+lea r0, [r0 + 2 * r1]
+lea r2, [r2 + 2 * r4]
+lea r3, [r3 + 2 * r5]
+
+pmovzxbw m0, [r2 + r4]
+pmovzxbw m1, [r2 + r4 + 8]
+
+movu m2, [r3 + r5]
+movu m3, [r3 + r5 + 16]
+
+paddw m0, m2
+paddw m1, m3
+
+packuswb m0, m1
+
+movu [r0 + r1], m0
+
+RET
More information about the x265-devel
mailing list