Hi,
On 08-09-15 09:53, Hans de Goede wrote:
Hi,
On 07-09-15 21:55, Ilia Mirkin wrote:
May I ask why you're doing 512x512 instead of 1024x1024? These are
already scaled up coordinates, so 1024x1024 should work no? Or is it
because of the seams on the edges? Do those not also appear with
512x512 or does it sample outside of the box?
This is my bad because of the bug fixed by patch 1/3 I had 512x512 in there
for a while, I've moved back and forth between 512 and 1024 a couple of times.
I've also tried 2048 but the hardware does not like that.
I will retest with 1024 and submit a fixed version.
I've not noticed any seams on the edges, even though I've been actively looking
for them.
Separately, why not use this approach on nv40 as well? I can't imagine
the blitter would be faster... does this result in lower quality?
I've the feeling the sifm bilinear filtering is more "blurry" then the
blitter one. I will do some more objective (ahem) tests on nv4x and get
back to you on this.
Ok I've run some tests comparing the rendering between using the blitter
and the sifm and the results seem identical, so one v2 using 1024x1024
blocks and doing so on all nv3x/nv4x cards coming up.
Regards,
Hans
_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau