[PATCH] D127871: [RISCV] Optimize 2x SELECT for floating-point types

Wed Jul 6 20:28:13 PDT 2022

dcaballe added a comment.

Hi!

This commit is causing a correctness regression in one of the ML models in IREE. Please, find attach the LLVM IR after Codegen Prepare. Hopefully, you can easily identify where this change is triggering.
To reproduce:  `llc bug.ll -mcpu=generic-rv64 -mattr=+m,+a,+f,+d,+c -target-abi=lp64d`. Please, note that no `+v` is provided. Could we please consider a revert while this is being investigated?

Thanks,
Diego
F23703572: bug.ll <https://reviews.llvm.org/F23703572>

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D127871/new/

https://reviews.llvm.org/D127871