On Fri, 16 May 2008 10:41:39 +0200 Thomas Hellström <[EMAIL PROTECTED]> wrote:
> Keith Packard wrote: > > > > Certainly moving from object management to page management would > > simplify handling of the GPU memory space, but I'm not sure how making > > these objects demand-paged would help all that much. > > > > > Me neither. I also think that we can get IRQ on page fault from GPU so you can upload only page the GPU actualy use, of course you likely want to avoid irq for each page so a tradeof would be to update predictively on irq a number of adjacent page a bit like cache predictively fetch ram. I think such things might really help reducing bandwith pressure by for instance helping to only upload a part of texture (one mipmap size for instance). Also one of the biggest win i see with virtual address space is that for radeon hw i no longer need to check command stream to catch any illegal memory access. This would simplify drm a lot for me. Cheers, Jerome Glisse ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ -- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
