John Tobin wrote:
>
> I have been monitoring the DRI since it first began and I have just
>
> one question regarding the MGA driver.
>
> I was wondering if you consider it to be functioning at the highest
> performance that you can get out of it?
>
> I ask this because I have been doing benchmarking and comparison between
> Windows and the various Linux drivers that are available. I just built a
> CVS snapshot from May 22 and there have been very little if any
> performance improvements since last November when multitexturing was
> finally enabled. Most notably the area in which performance is lacking
> that of windows is in 32 bit rendering. 32 bit rendering is generall 33
> to 25% slower than in windows regardless of AGP mode. I am however
> testing this on a 16 MB card so that may be affecting the performance to
> some degree.
>
> Even though performance hasn't increased considerably in some time I
> must say that the driver stability has increased 100 fold from when I
> first began using the driver. I remember pulling hair out trying to get
> UT working last July and August.
>
> Anyhow great job and keep up the great work!
We haven't done any real work on the MGA driver in a long time.
Nobody's funding us to do so.
You may get a bit of a performance improvement when we move
to Mesa 3.5-based driver. (still underway)
-Brian
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/dri-devel