> > > > I am interested in getting in touch with people who could help me with 
> > > > the
> > > > following question: what is the minimal subset of features necessary 
> > > > for vdr
> > > > (or mplayer)?
> > >
> > > In regard of OGP? Just a framebuffer. Everything else can be done
> > > in software.
> >
> > This has been discussed before.  CPUs aren't fast enough (see the archives
> > for examples).  Therefore we MUST have some sort of hardware assist.
> 
> I don't remember that.  I think people figured out that on not too old
> hardware, they could do the full decoding in software with no
> problems.

I definitely remember.  For one example, Peter wrote:

>       My system can't handle HD MPEG4 decoding and playback.  This a Radeon 
> 9800X=
>       T=20
>       on AGP8x, with a Athlon 64 3200+ and a gig of RAM.

An AMD64 3200+ is not a wimpy CPU.

In many cases, people decode SD or less and think there is no problem.  But HD
needs a LOT more CPU to decode than SD.  And some HD needs more than others.
HD decoding is not some optional luxury.  These days a lot of material is only
available in HD, so you have to decode the HD even if you only have a SD 
display.  

> > OGD1 is one thing, but OGC needs mpeg decoding or why bother?
> > We already have hw that can't do video, don't need another.
> 
> One of the main reasons to make OGD1 is so that we can full test the
> GPU design before it goes into an ASIC.  We can't afford to have an
> expensive mistake.  So, if we can avoid it, no logic will go into
> TRV10 that wasn't fully tested and integrated into the OGD1
> implementation of OGA.

If the logic doesn't fit into OGD1 all at once, is it modular enough
to test in stages?
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to