Stella Laurenzo comments

Results 204 comments of


                                            Stella Laurenzo

Upstream improvements for scf.for folding on loop ranges.

Any value in keeping it open with breadcrumbs for upstream fixes needed?

Upstream improvements for scf.for folding on loop ranges.

Sent out a flag to de-specialize FuncOp SCF stuff: https://reviews.llvm.org/D128614

Upstream improvements for scf.for folding on loop ranges.

Also de-specialized Affine passes: https://reviews.llvm.org/D128616

Upstream improvements for scf.for folding on loop ranges.

Iterating on a repro but can't see the illegal thing. I did bisect the other optimization passes and was able to repro only with this one, so I don't believe...

Upstream improvements for scf.for folding on loop ranges.

Ugh: https://github.com/llvm/llvm-project/issues/56235

Figure out how to build and test efficiently across multiple CI machines

Have you tried stripping all executables? Some kind of find piped to xargs strip.

Figure out how to build and test efficiently across multiple CI machines

Will, then, it is going to be huge and there isn't much to be done about it. Although some savings can be had by tweaking flags to only have line...

Figure out how to build and test efficiently across multiple CI machines

Switching the compiler to shared library builds would help with the big, duplicated binary problem a lot. It does come with a measurable performance cost though and should not be...

[RFC] Contributing an ONNX to TOSA legalization pass to onnx-mlir

I'm really excited to see this as well. Better connections here have long been on my wish list and seemed like gaps we could solve and make things better connect...

Experiment: coordinate LLVM commits between torch-mlir and onnx-mlir

Just in case if people aren't aware, in order to have a submodule dependency, it is sufficient to have the commit reachable by any branch in the source repo. On...