erichkeane wrote: > > In the future, we'll have to teach the OpenACC dialect how to derive the > > upperbound, lowerbound, and step from the cir.for loop > > Yea, even better if the `cir.for` representation can itself be improved to > expose that information (either by transforming it into a more refined loop > op or by redesigning the existing one).
Agreed! I suspect this will be a bit of a 'meet in the middle' for some of it, but it is hopefully clear that the the CIR and ACC dialect folks can/will work together to modify the loops to work together better. https://github.com/llvm/llvm-project/pull/137972 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
