[llvm] [WebAssembly] Lower fmuladd to fma (PR #161355)

[Jasmine Tang via llvm-commits]

badumbatish wrote:

> Performance wise, this looks really good for ML. Running NCNN, via V8 on AArch64, I observe a mean execution time reduction of ~15% over >30 workloads.

is there some sort of generic automated test suite that you're running that you can share with me? for llvm-test-suite, it's a bit hard since cmake can't build the targets (missing libs) so it requires some manual scripting

https://github.com/llvm/llvm-project/pull/161355