[clang] [llvm] Implement operand bundles for floating-point operations (PR #109798)

[Joshua Cranmer via cfe-commits]

jcranmer-intel wrote:

> We will need someone to go through and audit the libm-style intrinsics and make sure all optimizations of them are safe before we can switch over. Those are "unsafe by default" since the optimizers know about them already.

I've already been looking at all of the optimizations for fast-math flags as part of my work there.

Actually, most of the optimizations these days seem to be using the libcalls rather than the intrinsics (yay excessive duplication), so I don't think there's too much work with all of the libm intrinsics for optimizations.

https://github.com/llvm/llvm-project/pull/109798