On Tue, Feb 7, 2012 at 4:29 PM, Jakob Bornecrantz <ja...@vmware.com> wrote: > ----- Original Message ----- >> From: Dave Airlie <airl...@redhat.com> >> >> If we had no vertex textures previously and we have none now, >> don't bother doing the enables dance. >> >> Signed-off-by: Dave Airlie <airl...@redhat.com> >> --- >> src/mesa/state_tracker/st_atom_texture.c | 3 +++ >> 1 files changed, 3 insertions(+), 0 deletions(-) > > Nice find. Does this help performance? Can't we do the same for > fragment or does it just cost more to check (when this is rarely > the case) then just doing the rest of the function?
Well I think we generally have a fragment texture so it doesn't save much there, but I was just playing with noop and nexuiz and saw 0.46 CPU in vertex texture updates when I knew I had none. I think we could improve the frag texture updates a fair bit, but I'd have to a bit more testing on it. Dave, (oh I'll fix whitespace in that before pushing). _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
