[x265] [PATCH 3 of 6] asm: 10bpp sse4 code for saoCuOrgE1_2Rows, improved 10095c->900c, over C code
dnyaneshwar at multicorewareinc.com
dnyaneshwar at multicorewareinc.com
Mon Jun 22 14:50:35 CEST 2015
# HG changeset patch
# User Dnyaneshwar G <dnyaneshwar at multicorewareinc.com>
# Date 1434977140 -19800
# Mon Jun 22 18:15:40 2015 +0530
# Node ID 558ffdc4e832061d99f1ec688fe1ae64db48642f
# Parent 1c02df66f093b5b1bacc8a1bbf9be2ef81591ad5
asm: 10bpp sse4 code for saoCuOrgE1_2Rows, improved 10095c->900c, over C code
diff -r 1c02df66f093 -r 558ffdc4e832 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp Mon Jun 22 10:18:14 2015 +0530
+++ b/source/common/x86/asm-primitives.cpp Mon Jun 22 18:15:40 2015 +0530
@@ -1091,6 +1091,7 @@
{
p.saoCuOrgE0 = PFX(saoCuOrgE0_sse4);
p.saoCuOrgE1 = PFX(saoCuOrgE1_sse4);
+ p.saoCuOrgE1_2Rows = PFX(saoCuOrgE1_2Rows_sse4);
LUMA_ADDAVG(sse4);
CHROMA_420_ADDAVG(sse4);
diff -r 1c02df66f093 -r 558ffdc4e832 source/common/x86/loopfilter.asm
--- a/source/common/x86/loopfilter.asm Mon Jun 22 10:18:14 2015 +0530
+++ b/source/common/x86/loopfilter.asm Mon Jun 22 18:15:40 2015 +0530
@@ -442,6 +442,111 @@
; void saoCuOrgE1_2Rows(pixel *pRec, int8_t *m_iUpBuff1, int8_t *m_iOffsetEo, Int iStride, Int iLcuWidth)
;========================================================================================================
INIT_XMM sse4
+%if HIGH_BIT_DEPTH
+cglobal saoCuOrgE1_2Rows, 4,7,8
+ add r3d, r3d
+ mov r4d, r4m
+ pxor m0, m0 ; m0 = 0
+ mova m6, [pw_1023]
+ mov r5d, r4d
+ shr r4d, 4
+ mov r6, r0
+.loop
+ movu m7, [r0]
+ movu m5, [r0 + 16]
+ movu m3, [r0 + r3]
+ movu m1, [r0 + r3 + 16]
+
+ pcmpgtw m2, m7, m3
+ pcmpgtw m3, m7
+ pcmpgtw m4, m5, m1
+ pcmpgtw m1, m5
+ packsswb m2, m4
+ packsswb m3, m1
+ pand m2, [pb_1]
+ por m2, m3
+
+ movu m3, [r1] ; m3 = m_iUpBuff1
+
+ paddb m3, m2
+ paddb m3, [pb_2]
+
+ movu m4, [r2] ; m4 = m_iOffsetEo
+ pshufb m1, m4, m3
+
+ psubb m3, m0, m2
+ movu [r1], m3
+
+ pmovsxbw m3, m1
+ punpckhbw m1, m1
+ psraw m1, 8
+
+ paddw m7, m3
+ paddw m5, m1
+
+ pmaxsw m7, m0
+ pmaxsw m5, m0
+ pminsw m7, m6
+ pminsw m5, m6
+
+ movu [r0], m7
+ movu [r0 + 16], m5
+
+ add r0, 32
+ add r1, 16
+ dec r4d
+ jnz .loop
+
+ sub r1, r5
+ shr r5d, 4
+ lea r0, [r6 + r3]
+.loopH:
+ movu m7, [r0]
+ movu m5, [r0 + 16]
+ movu m3, [r0 + r3]
+ movu m1, [r0 + r3 + 16]
+
+ pcmpgtw m2, m7, m3
+ pcmpgtw m3, m7
+ pcmpgtw m4, m5, m1
+ pcmpgtw m1, m5
+ packsswb m2, m4
+ packsswb m3, m1
+ pand m2, [pb_1]
+ por m2, m3
+
+ movu m3, [r1] ; m3 = m_iUpBuff1
+
+ paddb m3, m2
+ paddb m3, [pb_2]
+
+ movu m4, [r2] ; m4 = m_iOffsetEo
+ pshufb m1, m4, m3
+
+ psubb m3, m0, m2
+ movu [r1], m3
+
+ pmovsxbw m3, m1
+ punpckhbw m1, m1
+ psraw m1, 8
+
+ paddw m7, m3
+ paddw m5, m1
+
+ pmaxsw m7, m0
+ pmaxsw m5, m0
+ pminsw m7, m6
+ pminsw m5, m6
+
+ movu [r0], m7
+ movu [r0 + 16], m5
+
+ add r0, 32
+ add r1, 16
+ dec r5d
+ jnz .loopH
+ RET
+%else ; HIGH_BIT_DEPTH
cglobal saoCuOrgE1_2Rows, 3, 5, 8, pRec, m_iUpBuff1, m_iOffsetEo, iStride, iLcuWidth
mov r3d, r3m
mov r4d, r4m
@@ -512,6 +617,7 @@
dec r4d
jnz .loop
RET
+%endif
INIT_YMM avx2
cglobal saoCuOrgE1_2Rows, 3, 5, 7, pRec, m_iUpBuff1, m_iOffsetEo, iStride, iLcuWidth
More information about the x265-devel
mailing list