This is an archive of the discontinued LLVM Phabricator instance.

[flang][openacc] Lower parallel loop combined construct
AbandonedPublic

Authored by clementval on Nov 24 2020, 4:31 PM.

Details

Summary

This patch lower the parallel loop combined construct to MLIR ops.

This code was part of the PR #558 on fir-dev
https://github.com/flang-compiler/f18-llvm-project/pull/558

Diff Detail

Event Timeline

clementval created this revision.Nov 24 2020, 4:31 PM
clementval requested review of this revision.Nov 24 2020, 4:31 PM
schweitz accepted this revision.Jan 4 2021, 5:04 PM

LGTM.

flang/lib/Lower/OpenACC.cpp
615

When we get the green-light to upstream, you may want to consider using fir::emitFatalError() here, as it will take mlir::Location information and produce a bit nicer output.

(On the other hand, the error seems like a general "this can never happen", but this was a chance to add a gratuitous plug.)

This revision is now accepted and ready to land.Jan 4 2021, 5:04 PM
mehdi_amini requested changes to this revision.Jan 4 2021, 5:12 PM

We're waiting on the ability to have tests for these I believe.

This revision now requires changes to proceed.Jan 4 2021, 5:12 PM

We're waiting on the ability to have tests for these I believe.

Yeah I don't mind waiting for the fir-dev RFC to go through and adding test for that.

clementval abandoned this revision.May 12 2021, 7:27 AM

Will send a new version when the general lowering is available upstream.