search

opencompl / lean-mlir

A minimal development of SSA theory
Other
68 stars 6 forks source link

Chore: Update InstCombine Benchmark #585

Closed AtticusKuhn closed 1 week ago

AtticusKuhn commented 1 week ago

Ever since LLVM.intMin was removed in https://github.com/opencompl/lean-mlir/pull/550 The InstCombine benchmark was actually broken. I have updated the InstCombine benchmark to use BitVec.intMin instead of LLVM.intMin. Now, the InstCombine benchmark is no longer broken.

If we want to avoid problems like this in the future, we could add all.lean to the GitHub workflow to make sure that no regressions happen (although this would be expensive).

github-actions[bot] commented 1 week ago

Alive Statistics: 64 / 93 (29 failed)