4 commits

Author	SHA1	Message	Date
Rhys Perry	e4d75c22be	nir/opt_shrink_vectors: shrink image stores using the format ... fossil-db (Navi): Totals from 657 (0.48% of 135946) affected shaders: VGPRs: 26076 -> 25520 (-2.13%); split: -2.15%, +0.02% CodeSize: 3033016 -> 3014472 (-0.61%); split: -0.64%, +0.03% MaxWaves: 9386 -> 9420 (+0.36%) Instrs: 590109 -> 585502 (-0.78%); split: -0.82%, +0.04% Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Reviewed-by: Eric Anholt <eric@anholt.net> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5772>	2020-09-07 18:06:50 +00:00
Jason Ekstrand	ff124e3fe3	nir: Add a load_global_constant intrinsic ... This has the same semantics as load_global except the memory it reads is known to be constant so load_global_constant intrinsics can be CSEd rather than relying on more complex copy-propagation. Reviewed-by: Karol Herbst <kherbst@redhat.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6379>	2020-09-01 20:50:03 +00:00
Eric Anholt	b440c28b78	nir: Shrink store intrinsic num_components to the size used by the writemask. ... This cuts a bunch of vector setup for undef components in the i965 vec4 backend. Noticed while looking into codegen regressions in nir-to-tgsi. brw results: total instructions in shared programs: 3893221 -> 3881461 (-0.30%) total cycles in shared programs: 113792154 -> 113810288 (0.02%) Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6054>	2020-08-20 16:44:08 +00:00
Eric Anholt	1c9906d5ff	nir: Add a pass to cut the trailing ends of vectors. ... Ideally we'd also handle unused middles of vectors and reswizzle ALU-only uses of it so we could write fewer channels, but that's future work/ Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Reviewed-by: Kenneth Graunke <kenneth@whitecape.org> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6050>	2020-08-03 21:26:45 +00:00