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

[Felix Buenemann]

On Thursday 11 July 2002 22:41, Per-Arne Holtmon Akø wrote:
> On Tue, Jul 09, 2002 at 23:18:48 +0000, Felix Buenemann wrote:
> > > Also attached the output from a 'cat /usr/compat/linux/proc/cpuinfo',
> > > at your request. Have fun with it :)
> >
> > Hmm, your /proc/cpuinfo advertises SSE support (the entry called "xmm"),
> > but if you have an Athlon MP, this is false. Also cpuMHz and bogomips are
> > the same, which I doubt is correct. Your should probably contact the
> > FreeBSD develoeprs about this.
> >
> > (All these assumptions are based on AMDs techdocs, so I hope they are
> > correct about this.)
>
> I'll answer in this mail only, since two threads about the same is kinda
> worthless. :)  Anyway, yes, I did compile with
> --enable-runtime-cpudetection, and it spat out that error/warning. Since
> my system is rather fast, I haven't seen any major slowdowns, though.
If you enable runtime cpu detection (which only makes sense if you are a 
binary packager and want to transfer the same mplayer binary to another 
system with different cpu or you swap mainboard+cpu in your pc regularly), 
than you miss the whole point the discussion was about, because then gcc will 
be instructed to compile generic code, that will even work on 486!

-- 
Best Regards,
        Atmos
____________________________________________
- MPlayer Developer - http://mplayerhq.hu/ -
____________________________________________