projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: ba77c3e) | patch
R600/SI: add sample and image intrinsics exposing all instruction fields
authorMarek Olsak <marek.olsak@amd.com>
Fri, 11 Jul 2014 17:11:46 +0000 (17:11 +0000)
committerMarek Olsak <marek.olsak@amd.com>
Fri, 11 Jul 2014 17:11:46 +0000 (17:11 +0000)
commitd8ecaeec02bddd18ddf2593140998c3fa4cf3b40
tree42b592078a5db168be48b6d4c7a8024fd24ece79tree | snapshot
parentba77c3e4ed214a7ea669bd8ccbc997fde55b8424commit | diff
R600/SI: add sample and image intrinsics exposing all instruction fields

We need the intrinsics with offsets, so why not just add them all.
The R128 parameter will also be useful for reducing SGPR usage.
GL_ARB_image_load_store also adds some image GLSL modifiers like "coherent",
so Mesa will probably translate those to slc, glc, etc.

When LLVM 3.5 is released, I'll switch Mesa to these new intrinsics.

llvm-svn: 212830
llvm/lib/Target/R600/SIInstructions.td diff | blob | history
llvm/lib/Target/R600/SIIntrinsics.td diff | blob | history
llvm/test/CodeGen/R600/llvm.SI.image.ll [new file with mode: 0644] blob
llvm/test/CodeGen/R600/llvm.SI.image.sample.ll [new file with mode: 0644] blob
llvm/test/CodeGen/R600/llvm.SI.image.sample.o.ll [new file with mode: 0644] blob
Domain: System / Toolchain;