Re: [FFmpeg-devel] [PATCH] SSE2 version of vf_idet's filter_line()

2014-09-04 Thread Michael Niedermayer
On Wed, Sep 03, 2014 at 02:04:43PM -0700, Pascal Massimino wrote: Clément On Wed, Sep 3, 2014 at 12:37 PM, Clément Bœsch u...@pkh.me wrote: On Wed, Sep 03, 2014 at 07:05:48PM +0200, Pascal Massimino wrote: [...] +punpcklbw m3, m_zero +punpckhbw m4, m_zero + +

Re: [FFmpeg-devel] [PATCH] SSE2 version of vf_idet's filter_line()

2014-09-03 Thread Pascal Massimino
Michael, On Wed, Sep 3, 2014 at 4:29 PM, Michael Niedermayer michae...@gmx.at wrote: On Wed, Sep 03, 2014 at 11:42:10AM +0200, Pascal Massimino wrote: On Wed, Sep 3, 2014 at 11:32 AM, Benoit Fouet benoit.fo...@free.fr wrote: Hi, - Mail original - Hi,

Re: [FFmpeg-devel] [PATCH] SSE2 version of vf_idet's filter_line()

2014-09-03 Thread Pascal Massimino
Clément, On Wed, Sep 3, 2014 at 6:19 PM, Clément Bœsch u...@pkh.me wrote: On Wed, Sep 03, 2014 at 05:50:32PM +0200, Pascal Massimino wrote: [...] removed this step in both mmx and sse2 version. - new patch attached. /skal From d2249b05b4a881ec3c9de8fc105b2a40c680a0ea Mon Sep 17

Re: [FFmpeg-devel] [PATCH] SSE2 version of vf_idet's filter_line()

2014-09-03 Thread Pascal Massimino
James, On Wed, Sep 3, 2014 at 10:14 AM, James Almer jamr...@gmail.com wrote: diff --git a/libavfilter/x86/vf_idet_init.c b/libavfilter/x86/vf_idet_init.c new file mode 100644 index 000..402d504 --- /dev/null +++ b/libavfilter/x86/vf_idet_init.c @@ -0,0 +1,70 @@ +/* + * This

Re: [FFmpeg-devel] [PATCH] SSE2 version of vf_idet's filter_line()

2014-09-03 Thread Clément Bœsch
On Wed, Sep 03, 2014 at 07:05:48PM +0200, Pascal Massimino wrote: [...] +punpcklbw m3, m_zero +punpckhbw m4, m_zero + +paddswm0, m3 +paddswm1, m4 + +movq m3, [bq+indexq*1] +movq m4, m3 +punpcklbw m3, m_zero +punpckhbw

Re: [FFmpeg-devel] [PATCH] SSE2 version of vf_idet's filter_line()

2014-09-02 Thread Michael Niedermayer
On Tue, Sep 02, 2014 at 05:13:08PM +0200, Pascal Massimino wrote: Hi, Much faster. An example (time ffmpeg -i ... -vf idet -f null /dev/null) Raw C: user 25.007s MMX:user 16.818s MMXEXT: user 16.191s SSE2: user 15.481s no idet filter: user 15.025s YMMV. skal

Re: [FFmpeg-devel] [PATCH] SSE2 version of vf_idet's filter_line()

2014-09-02 Thread Pascal Massimino
Michael, On Tue, Sep 2, 2014 at 9:39 AM, Michael Niedermayer michae...@gmx.at wrote: On Tue, Sep 02, 2014 at 05:13:08PM +0200, Pascal Massimino wrote: Hi, Much faster. An example (time ffmpeg -i ... -vf idet -f null /dev/null) Raw C: user 25.007s MMX:user 16.818s