[FFmpeg-devel] [PATCH] Fix configure to disable MMX for Pentium Pro

Dominik 'Rathann' Mierzejewski dominik
Thu Aug 26 16:59:46 CEST 2010


On Thursday, 26 August 2010 at 16:48, Jason Garrett-Glaser wrote:
> 2010/8/26 M?ns Rullg?rd <mans at mansr.com>:
> > "Ronald S. Bultje" <rsbultje at gmail.com> writes:
> >
> >> Hi,
> >>
> >> On Thu, Aug 26, 2010 at 9:34 AM, Alex Converse <alex.converse at gmail.com> wrote:
> >>> ./configure --cpu=prentiumpro says build me an FFmpeg for a
> >>> PentiumPro. The result will no longer run on a PentiumPro.
> >>
> >> A) It should require --disable-mmx (and thus the configure patch to
> >> disable mmx should be changed to error out unless mmx is disabled),
> >
> > That's a stupid idea. ?Specifying a CPU without MMX should of course
> > disable MMX. ?That stands to reason. ?It is also how all the other
> > architectures work.
> >
> >> B) we should revert the mmx-requiring commit and only do unconditional
> >> emms on x86-64 (I think this is acceptable. Mans?).
> >
> > The main point of the change was to get rid of the stupid global
> > mm_flags variable. ?We discussed it at length some time ago (it nearly
> > went full bikeshed), and nobody objected then, not even the distro people.
> > Why the sudden outrage?
> 
> Whether we like it or not, a huge number of people compile ffmpeg with
> --cpu=i686 or similar for generic builds.  With these changes,
> --cpu=i686 will actually SLOW DOWN ffmpeg as opposed to not specifying
> CPU at all.  This is not intuitive behavior.

Not in my opinion. Do you expect gcc -march=i686 to generate code
that won't run on a Pentium Pro? No. If you want MMX/SSE/whatever,
you add -mmmx/-msse/whatever, and then you know that it won't,
because you specified it. Perfectly intuitive, if you ask me.

Just my two cents.

Regards,
Dominik

-- 
MPlayer http://mplayerhq.hu | RPMFusion http://rpmfusion.org
There should be a science of discontent. People need hard times and
oppression to develop psychic muscles.
	-- from "Collected Sayings of Muad'Dib" by the Princess Irulan



More information about the ffmpeg-devel mailing list