[llvm] [AMDGPU] Flush denormal results when constant folding amdgcn.rcp (PR #203566)

Jay Foad via llvm-commits llvm-commits at lists.llvm.org
Mon Jun 15 06:45:06 PDT 2026

Previous message: [llvm] [GlobalISel] Add known-bits for G_INSERT_VECTOR_ELT. (PR #200930)
Next message: [llvm] [VPlan] Simplify reverse(reverse(x)) -> x (PR #199057)
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

jayfoad wrote:

> We probably should stop constant folding this if we can't guarantee bit identical results. Which this still can't.

Agreed.

https://github.com/llvm/llvm-project/pull/203566

Previous message: [llvm] [GlobalISel] Add known-bits for G_INSERT_VECTOR_ELT. (PR #200930)
Next message: [llvm] [VPlan] Simplify reverse(reverse(x)) -> x (PR #199057)
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

More information about the llvm-commits mailing list