[llvm] [SDAG[[X86] Added method to scalarize `STRICT_FSETCC` (PR #154486)

[Abhishek Kaushik via llvm-commits]

abhishek-kaushik22 wrote:

> Wait, vectors of fp80 are allowed? Can we make that a verifier error? Can you test with a real element type?

It also happens for other floating-point types but only when the vector is `<1 x fp>`.

https://godbolt.org/z/beee7s77x

https://github.com/llvm/llvm-project/pull/154486