search

plaidml / tpp-mlir

TPP experimentation on MLIR for linear algebra
https://arxiv.org/abs/2404.15204
Other
111 stars 31 forks source link

Retire downstream scf.forall to scf.parallel conversion #930

Closed adam-smnk closed 5 months ago

adam-smnk commented 5 months ago

Removes custom conversion logic and reuses the existing pass as a simple wrapper around upstream transform API.

adam-smnk commented 5 months ago

No regressions.