peixin added a comment. In D109321#2985284 <https://reviews.llvm.org/D109321#2985284>, @jhuber6 wrote:
> In D109321#2985281 <https://reviews.llvm.org/D109321#2985281>, @lebedev.ri > wrote: > >> Aha. But i don't think this is the right fix, >> the fact that the inlining manifests the miscompile is a symptom. > > Preventing the outlined region from being inlined would also hurt OpenMP > performance considerably. Please notice that this remove is only inside `emitOutlinedOrderedFunction`, which is only used when `ordered simd` directive is there. According to OpenMP 5.0 Spec, the ordered construct either specifies a structured block in a worksharing-loop, simd, or worksharing-loop SIMD region that will be executed in the order of the loop iterations. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D109321/new/ https://reviews.llvm.org/D109321 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
