summaryrefslogtreecommitdiff
path: root/libavfilter/x86
diff options
context:
space:
mode:
authorMichael Niedermayer <michaelni@gmx.at>2014-11-15 04:07:08 +0100
committerVittorio Giovara <vittorio.giovara@gmail.com>2014-11-25 02:00:06 +0000
commitca5c3ff90972a5c97aabda2ace57ba72dcd7d83b (patch)
tree2777b0510afbf53d4eab7e2d8df5cb621d0a6468 /libavfilter/x86
parent57ed5a64feec4af1f16f9a74c63cfa9aa8147242 (diff)
vf_interlace: x86: improve asm performance
4775 decicycles -> 3688 decicycles
Diffstat (limited to 'libavfilter/x86')
-rw-r--r--libavfilter/x86/vf_interlace.asm46
1 files changed, 19 insertions, 27 deletions
diff --git a/libavfilter/x86/vf_interlace.asm b/libavfilter/x86/vf_interlace.asm
index 8c2e9b099c..b8d86165bf 100644
--- a/libavfilter/x86/vf_interlace.asm
+++ b/libavfilter/x86/vf_interlace.asm
@@ -2,6 +2,7 @@
;* x86-optimized functions for interlace filter
;*
;* Copyright (C) 2014 Kieran Kunhya <kierank@obe.tv>
+;* Copyright (c) 2014 Michael Niedermayer <michaelni@gmx.at>
;*
;* This file is part of Libav.
;*
@@ -34,36 +35,27 @@ cglobal lowpass_line, 5, 5, 7
add r4, r1
neg r1
- pxor m6, m6
+ pcmpeqb m6, m6
.loop
- mova m0, [r2+r1]
- punpcklbw m1, m0, m6
- punpckhbw m0, m6
- paddw m0, m0
- paddw m1, m1
+ mova m0, [r3+r1]
+ mova m1, [r3+r1+mmsize]
+ pavgb m0, [r4+r1]
+ pavgb m1, [r4+r1+mmsize]
+ mova m2, [r2+r1]
+ mova m3, [r2+r1+mmsize]
+ pxor m0, m6
+ pxor m1, m6
+ pxor m2, m6, [r2+r1]
+ pxor m3, m6, [r2+r1+mmsize]
+ pavgb m0, m2
+ pavgb m1, m3
+ pxor m0, m6
+ pxor m1, m6
+ mova [r0+r1], m0
+ mova [r0+r1+mmsize], m1
- mova m2, [r3+r1]
- punpcklbw m3, m2, m6
- punpckhbw m2, m6
-
- mova m4, [r4+r1]
- punpcklbw m5, m4, m6
- punpckhbw m4, m6
-
- paddw m1, m3
- pavgw m1, m5
-
- paddw m0, m2
- pavgw m0, m4
-
- psrlw m0, 1
- psrlw m1, 1
-
- packuswb m1, m0
- mova [r0+r1], m1
-
- add r1, mmsize
+ add r1, 2*mmsize
jl .loop
REP_RET
%endmacro