projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 96a1f00) | patch
AMDGPU/SI: Add VI patterns to select FLAT instructions for global memory ops
authorTom Stellard <thomas.stellard@amd.com>
Mon, 20 Jul 2015 14:28:41 +0000 (14:28 +0000)
committerTom Stellard <thomas.stellard@amd.com>
Mon, 20 Jul 2015 14:28:41 +0000 (14:28 +0000)
commitf799b25cfc7c1eb3d3d569fd263ff122834f2529
tree334d004171d8f4789793999743fc922526e2420atree | snapshot
parent96a1f00376e05ea16dd6671685affb92fc8dde94commit | diff
AMDGPU/SI: Add VI patterns to select FLAT instructions for global memory ops

Summary:
The MUBUF addr64 bit has been removed on VI, so we must use FLAT
instructions when the pointer is stored in VGPRs.

Reviewers: arsenm

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D11067

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@242673 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp diff | blob | history
lib/Target/AMDGPU/SIISelLowering.cpp diff | blob | history
lib/Target/AMDGPU/SIISelLowering.h diff | blob | history
lib/Target/AMDGPU/VIInstructions.td diff | blob | history
test/CodeGen/AMDGPU/cgp-addressing-modes.ll diff | blob | history
test/CodeGen/AMDGPU/global_atomics.ll diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors