Page MenuHomePhabricator

[mlir][VectorOps] Generalized vector.print to i32/i64
ClosedPublic

Authored by aartbik on Thu, Feb 6, 4:13 PM.

Details

Summary

Lowering to LLVM IR was restricted to float/double.
This CL also adds the integral values.

Diff Detail

Event Timeline

aartbik created this revision.Thu, Feb 6, 4:13 PM
Herald added a project: Restricted Project. · View Herald TranscriptThu, Feb 6, 4:13 PM
ftynse accepted this revision.Fri, Feb 7, 4:59 AM
ftynse marked an inline comment as done.
ftynse added inline comments.
mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir
403

Nit: let's drop the arg[0-9]+ pattern and just use .*, we don't guarantee anything on SSA names

mlir/test/mlir-cpu-runner/mlir_runner_utils.cpp
81

Cool, I did not know these macros existed!

This revision is now accepted and ready to land.Fri, Feb 7, 4:59 AM
nicolasvasilache accepted this revision.Fri, Feb 7, 8:07 AM

Thanks Aart!

aartbik updated this revision to Diff 243189.Fri, Feb 7, 9:12 AM
aartbik marked 3 inline comments as done.

addressed comments

aartbik added inline comments.Fri, Feb 7, 9:14 AM
mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir
403

Done, also for the other occurrences for consistency.

mlir/test/mlir-cpu-runner/mlir_runner_utils.cpp
81

A great way to keep our code portable!

This revision was automatically updated to reflect the committed changes.