[MPlayer-users] Possible bug [FreeBSD 5.0-CURRENT]

Felix Buenemann atmosfear at users.sourceforge.net
Tue Jul 9 23:30:02 CEST 2002


On Tuesday 09 July 2002 13:30, Per-Arne Holtmon Akø wrote:
> On Tue, Jul 09, 2002 at 01:14:11 +0000, Felix Buenemann wrote:
> > On Thursday 04 July 2002 09:40, Per-Arne Holtmon Akø wrote:
> > OK, I tracked down the problem.
> > Athlon XP/MP and some other Athlon models all share the same CPUID (666),
> > but only the XP supports SSE, this leads to miscompiled code for MP and
> > others. Your log also shows, that gcc uses SSE which is not possible on
> > your cpu.
> >
> > I have now commited a fix to configure, that will fallback to athlon-4
> > optimization, if no SSE is detected.
> >
> > However if you want to get the most out of your cpu, you should edit the
> > config.mak to use -march=athlon-mp -mcpu=athlon-mp, as the MP is not
> > currently auto-detected as MP but as Athlon-4.
> >
> > Please test if current configure fixes your prob and report back.
> >
> > And if you have linux /proc/cpuinfo emulation, it would be nice, if you
> > could paste the output here.
>
> Now configure will set -march=i486 and -mcpu=i686, which I suppose was
> your intention. It compiles and runs fine now.
>
> There is one thing though:
> "Compiled with RUNTIME CPU Detection - warning, it's not optimal! To get
> best performance, recompile mplayer from sources with
> --disable-runtime-cpudetection"
> Does this mean that my current configuration should not use
> runtime-detection, or is that a general hint now?
Heh? You have enabled runtime cpudetection, it will slow things down, you 
should compile with configure --disable-runtime-cpudetection, it should be 
default, did you enable it by --enable-runtime-cpudetection ?

> Also attached the output from a 'cat /usr/compat/linux/proc/cpuinfo', at
> your request. Have fun with it :)
thx
-- 
Best Regards,
        Atmos
____________________________________________
- MPlayer Developer - http://mplayerhq.hu/ -
____________________________________________




More information about the MPlayer-users mailing list