[FFmpeg-devel] [PATCH 3/3] h264: new assembly version of get_cabac for x86_64 with PIC
michaelni at gmx.at
Fri Apr 27 22:02:03 CEST 2012
On Fri, Apr 27, 2012 at 09:54:27PM +0200, Roland Scheidegger wrote:
> Am 27.04.2012 21:06, schrieb Michael Niedermayer:
> > On Fri, Apr 27, 2012 at 03:45:26AM +0200, Roland Scheidegger wrote:
> >> This adds a hand-optimized assembly version for get_cabac much like the
> >> existing one, but it works if the table offsets are RIP-relative.
> >> Compared to the non-RIP-relative version this adds 2 lea instructions
> >> and it needs one extra register.
> >> There is a surprisingly large performance improvement over the c version (more
> >> so than the generated assembly seems to suggest) just in get_cabac, I measured
> >> roughly 40% faster for get_cabac on a K8. However, overall the difference is
> >> not that big, I measured roughly 5% on a test clip on a K8 and a Core2.
> >> Hopefully it still compiles on x86 32bit...
> >> Now that only one table is used, there's some chance even darwin as compiles
> >> this (apparently the label arithmetic used previously doesn't work if it
> >> involves symbols defined in a different file, thanks to Ronald S. Bultje for
> >> helping me with this).
> > with the 3 patches applied i get
> > In file included from libavcodec/cabac_functions.h:36:0,
> > from libavcodec/cabac.c:32:
> > libavcodec/x86/cabac.h: In function ‘get_cabac_inline_x86’:
> > libavcodec/x86/cabac.h:169:79: error: macro "BRANCHLESS_GET_CABAC_UPDATE" passed 5 arguments, but takes just 4
> > libavcodec/x86/cabac.h:167:9: error: expected ‘;’ before string constant
> > libavcodec/x86/cabac.h:156:14: warning: unused variable ‘tmp’ [-Wunused-variable]
> Ahh that must be non-pic case.
> That was an accident happening when splitting up the patch into multiple
> ones, had to trim macro arguments and inadvertently trimmed arguments
> for BRANCHLESS_GET_CABAC_UPDATE too (which indeed did have a unnecessary
> I can respin the series either way (with the argument removed or just
> left as is as that's not really a related change).
if its unrelated then its better not to be in this patch series
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
The worst form of inequality is to try to make unequal things equal.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 198 bytes
Desc: Digital signature
More information about the ffmpeg-devel