projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 7894ffd) | patch
R600: Treat CONSTANT_ADDRESS loads like GLOBAL_ADDRESS loads when necessary
authorTom Stellard <thomas.stellard@amd.com>
Tue, 23 Jul 2013 23:54:56 +0000 (23:54 +0000)
committerTom Stellard <thomas.stellard@amd.com>
Tue, 23 Jul 2013 23:54:56 +0000 (23:54 +0000)
commit8ea83d499935f406f9fc190de783710c2293ab63
treeb95a57b1b8f241a424a9930e64322ebda0c957aatree | snapshot
parent7894ffd006bfabd45c75696c9b7cfdda7c7bad9dcommit | diff
R600: Treat CONSTANT_ADDRESS loads like GLOBAL_ADDRESS loads when necessary

These are really the same address space in hardware.  The only
difference is that CONSTANT_ADDRESS uses a special cache for faster
access.  When we are unable to use the constant kcache for some reason
(e.g. smaller types or lack of indirect addressing) then the instruction
selector must use GLOBAL_ADDRESS loads instead.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187006 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/R600/AMDGPUISelDAGToDAG.cpp diff | blob | history
lib/Target/R600/R600Instructions.td diff | blob | history
test/CodeGen/R600/load.ll diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors