On Monday, December 12, 2016 10:14:35 AM PST Eric Anholt wrote: > Kenneth Graunke <kenn...@whitecape.org> writes: > > > A 16384x16384 RGBA8888 buffer takes 1GB of RAM. If the driver allocates > > mipmap levels, that increases our storage requirements to 1.5GB. The > > test doesn't use mipmapping, so this seems like a waste. > > > > Disabling mipmap filtering before allocating the texture provides a hint > > to the driver, suggesting that it should only allocate space for the > > base level. Most Mesa drivers (Gallium and i965) follow this practice. > > > > Signed-off-by: Kenneth Graunke <kenn...@whitecape.org> > > It would probably be good for coverage if the fbo-maxsize test tested > that you could render all the way down the miptree of a maximum-sized > FBO, but given that it doesnt, this is: > > Reviewed-by: Eric Anholt <e...@anholt.net> >
Thanks! I screwed up and missed applying the Reviewed-by tag, and apparently Piglit doesn't allow non-fast-forward pushes (like Mesa), so I can't fix it. Sorry :(
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Piglit mailing list Piglit@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/piglit