Mailing Lists: Apple Mailing Lists
Image of Mac OS face in stamp
 
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

ATI_array_rev_comps_in_4_bytes bug?

I've been playing with ATI_array_rev_comps_in_4_bytes recently, and found that it gives me up to a 40% speedup in my VAR benchmark test on Radeon 9000 and 9200 systems. Good. However, on Radeon 7500 systems I have problems:

Under 10.3.x (tested 10.3.2, 10.3.4, 10.3.6) I see no improvement (it is slightly slower due to CPU byteswapping, I use NXSwapLong to flip 4ub colors.)

Under 10.2.8, the extension is exported, but doesn't work. My flipped colors are used onscreen!


So:
* can anyone corroborate this behavior (broken extension)? I suppose there is no point in logging a bug against Jaguar.

* can anyone explain how to get better VAR performance on Radeon 7500 systems? My vertex data is interleaved, tightly packed: 4ub color, 2f coord, 2f tex. That gives a 20 byte stride, is there an alignment issue here? The data is used by both glDrawArrays and glDrawElements, using GL_QUADS and passing at least 4096 vertices at a time via multiple fenced VARs. Is this an unoptimized primitive type? My indices are unsigned shorts but they aren't in an element_array and aren't being byteswapped.


thanks,
-alex

_______________________________________________
Do not post admin requests to the list. They will be ignored.
Mac-opengl mailing list      (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/mac-opengl/email@hidden

This email sent to email@hidden


Visit the Apple Store online or at retail locations.
1-800-MY-APPLE

Contact Apple | Terms of Use | Privacy Policy

Copyright © 2007 Apple Inc. All rights reserved.