nv50/ir: use levelZero for non-frag tex/txp ops
authorIlia Mirkin <imirkin@alum.mit.edu>
Tue, 18 Oct 2016 04:25:38 +0000 (00:25 -0400)
committerIlia Mirkin <imirkin@alum.mit.edu>
Sat, 22 Oct 2016 16:02:35 +0000 (12:02 -0400)
commitadad576bfc606694323fcbda3113ea339a3e935d
tree7f3c2ac6f03e8cff0f264ff3bcbb5992cc68ee0b
parent3fdeb7c983d5a6e8704024dc841ff217377ba284
nv50/ir: use levelZero for non-frag tex/txp ops

radeonsi also does the same thing. I suspect that this is likely to be a
no-op in reality, but it brings nouveau code closer to what the blob
produces. Plus it makes sense to not try to do auto-derivatives on this.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
src/gallium/drivers/nouveau/codegen/nv50_ir_from_tgsi.cpp