aboutsummaryrefslogtreecommitdiff
path: root/clang/lib/FrontendTool/ExecuteCompilerInvocation.cpp
diff options
context:
space:
mode:
authorJeffrey Byrnes <jeffrey.byrnes@amd.com>2024-10-09 09:54:11 -0700
committerGitHub <noreply@github.com>2024-10-09 09:54:11 -0700
commit17bc959961aa5792821516b547100316fc886ab4 (patch)
treee08b2acb780cc8daa9a0d0946cd9506914c28f1c /clang/lib/FrontendTool/ExecuteCompilerInvocation.cpp
parent1553cb5d3b14a0516c2796c295a3b32d147d13d0 (diff)
downloadllvm-17bc959961aa5792821516b547100316fc886ab4.zip
llvm-17bc959961aa5792821516b547100316fc886ab4.tar.gz
llvm-17bc959961aa5792821516b547100316fc886ab4.tar.bz2
[AMDGPU] Optionally Use GCNRPTrackers during scheduling (#93090)
This adds the ability to use the GCNRPTrackers during scheduling. These trackers have several advantages over the generic trackers: 1. global live-thru trackers, 2. subregister based RP deltas, and 3. flexible vreg -> PressureSet mappings. This feature is off-by-default to ease with the roll-out process. In particular, when using the optional trackers, the scheduler will still maintain the generic trackers leading to unnecessary compile time.
Diffstat (limited to 'clang/lib/FrontendTool/ExecuteCompilerInvocation.cpp')
0 files changed, 0 insertions, 0 deletions