[llvm] [LV][EVL] Support call instruction with EVL-vectorization (PR #110412)

Mel Chen via llvm-commits llvm-commits at lists.llvm.org
Mon Nov 4 03:14:52 PST 2024


https://github.com/Mel-Chen commented:

I noticed that we need to pass not only the debug location, but possibly also the fast math flags (FMF). This might require modifying the constructor. 
Possible solutions could be: 
1. Pass FMF directly, similar to how we pass the debug location. 
2. Pass the underlying instruction as `Instruction &`. 
3. Pass `VPWidenIntrinsicRecipe &`.
@fhahn What do you think?

https://github.com/llvm/llvm-project/pull/110412


More information about the llvm-commits mailing list