XRender acceleration

[Adam Jackson]

On Thursday 16 March 2006 15:52, Zack Rusin wrote:
> On Thursday 16 March 2006 15:26, you wrote:
> > On Thursday 16 March 2006 14:48, Zack Rusin wrote:
> > > Hey, [*1]
> > >
> > > we're doing better and better at accelerating XRender. Muchas problemos
> > > that we have on the toolkit level are caused by the fact that we (*2)
> > > have no idea what is accelerated.
> >
> > So why don't we have a useful Render performance benchmark?
> >
> > There's no useful way to report this information from the server that
> > doesn't lead to polynomial code explosion on both sides.  What you
> > actually need is a performance profile of the ops that matter and some
> > user checkboxes that say "yes gradients are slow but I want them anyway".
>
> That sounds like a good idea...
>
> http://www.freedesktop.org/~zack/code/xrenderbenchmark.git/
>
> Now can you provide me with a way of querying server/exa/driver
> version/vendor number that doesn't involve me scanning log files? (aka. is
> reasonable)

Grnh, I always blanch at this suggestion, the tuple you'd need for accurate 
identification is unpleasantly large.  I suppose the loader could attach some 
atoms to the root window of each screen; would that be close enough?

- ajax

> Zack
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.x.org/archives/xorg/attachments/20060316/d81689e4/attachment.pgp>