To find the candidates to merge stores we iterate over all nodes in a chain
for each store, which leads to quadratic compile times for large basic blocks
with a large number of stores.
Details
Details
- Reviewers
niravd spatel craig.topper - Commits
- rZORGfcaa96bf378c: [DAGCombiner] Limit number of nodes explored as store candidates.
rZORG1d04fbbd8eed: [DAGCombiner] Limit number of nodes explored as store candidates.
rGfcaa96bf378c: [DAGCombiner] Limit number of nodes explored as store candidates.
rG1d04fbbd8eed: [DAGCombiner] Limit number of nodes explored as store candidates.
rGbe10bc71f9af: [DAGCombiner] Limit number of nodes explored as store candidates.
rL360357: [DAGCombiner] Limit number of nodes explored as store candidates.
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
- Build Status
Buildable 31369 Build 31368: arc lint + arc unit
Event Timeline
Comment Actions
Updated increments.
Thanks, the token factor size limiting patch is at D61397, please let me know what you think and
if there is a better place to do the limiting somewhere else.
llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp | ||
---|---|---|
14845 | Done,thanks! |
nit: use prefix increment to be consistent with other uses.