xine-lib/src/xine-utils, branch 1.2.5 xine-lib git mirror Fix a short zero-init. 2014-03-11T14:59:27+00:00 Darren Salt devspam@moreofthesa.me.uk 2014-03-11T14:59:27+00:00 c4eafc62007b47d443d53321c71e94239ab028e3 






Silence compiler warnings on register_enum () arg #4.
2014-03-10T15:36:55+00:00

Torsten Jager
t.jager@gmx.de

2014-03-10T15:36:55+00:00

9dc01446b018803415e0fe02b0f060b037e0738a

Making them all "const char * const *" did work too
(even with Kaffeine build/run), but that would be
an API change.





Making them all "const char * const *" did work too
(even with Kaffeine build/run), but that would be
an API change.







Silence some compiler warnings.
2014-03-08T14:11:20+00:00

Torsten Jager
t.jager@gmx.de

2014-03-08T14:11:20+00:00

51f379ba834dbc5bca0007ef4f3793a7d650ae78












Copyright year update by hg log #4.
2014-01-04T14:44:35+00:00

Torsten Jager
t.jager@gmx.de

2014-01-04T14:44:35+00:00

552bc88536783305bc78664f1226d4e028935634

BTW. Happy 2014 to you!
     Didnt think xine will survive this far :-)
     And yes, I finally got an ffmpeg patch through. That kind of rounds the circle :-)





BTW. Happy 2014 to you!
     Didnt think xine will survive this far :-)
     And yes, I finally got an ffmpeg patch through. That kind of rounds the circle :-)







Copyright year update by hg log info #2.
2013-10-14T15:44:10+00:00

Torsten Jager
t.jager@gmx.de

2013-10-14T15:44:10+00:00

c2a1e087ac54a165dbf632650143f6a90ba51784












Clean up vaapi µs delay code.
2013-10-01T20:25:36+00:00

Darren Salt
devspam@moreofthesa.me.uk

2013-10-01T20:25:36+00:00

b158fdb88389b33de8e1447f77d29ff7a8975463












Put vzeroupper in the right place, ref. patch as sent to xine-devel.
2013-09-18T17:01:02+00:00

Darren Salt
devspam@moreofthesa.me.uk

2013-09-18T17:01:02+00:00

14d684005aa248bb0fca6c383cb1122158ada3a4












Emit vzeroupper after avx memcpy
2013-09-17T22:58:37+00:00

Roland Scheidegger
rscheidegger_lists@hispeed.ch

2013-09-17T22:58:37+00:00

b0fa6cab5e64886d3f47156a049c5d1b85dcabd9

Emitting vzeroupper is necessary to avoid avx<->sse transition penalties (when
using avx-256 instructions).
This didn't really matter much in the past, since other code wasn't using avx,
hence there was just a penalty once afterwards when sse code was executed.
However, there's code in ffmpeg which mixes avx-128 and sse a lot, and each
time this happens there's a huge penalty. This causes in particular
ff_deblock_v_luma_8_avx to slow down by a factor of 50 or so which makes the
whole decoding about twice as slow (might be dependent on the h264 stream or
maybe ffmpeg version too, since ffmpeg will also emit vzeroupper when using
avx-256 hence not doing it here might not always be an issue, but in the case
I was seeing nothing else used avx-256).





Emitting vzeroupper is necessary to avoid avx<->sse transition penalties (when
using avx-256 instructions).
This didn't really matter much in the past, since other code wasn't using avx,
hence there was just a penalty once afterwards when sse code was executed.
However, there's code in ffmpeg which mixes avx-128 and sse a lot, and each
time this happens there's a huge penalty. This causes in particular
ff_deblock_v_luma_8_avx to slow down by a factor of 50 or so which makes the
whole decoding about twice as slow (might be dependent on the h264 stream or
maybe ffmpeg version too, since ffmpeg will also emit vzeroupper when using
avx-256 hence not doing it here might not always be an issue, but in the case
I was seeing nothing else used avx-256).







Copyright year update by hg log info
2013-09-13T10:50:25+00:00

Torsten Jager
t.jager@gmx.de

2013-09-13T10:50:25+00:00

50caa6c2122224df42c6d8a5c88ffb5b2d4afb8b

script execution time: 55"





script execution time: 55"







xine_mmx.h: make sure alignment macros are defined. Make some noise if alignment requirements are not met.
2013-04-15T20:05:40+00:00

Petri Hintukainen
phintuka@users.sourceforge.net

2013-04-15T20:05:40+00:00

6b571cf5da3a4639400b2845e8a90a033d675290