[llvm] CodeGen: Add -denormal-fp-math-bf16 flag (PR #90425)

[Freddy Ye via llvm-commits]

FreddyLeaf wrote:

> So all the instructions are just defective?

I'm afraid so. BF16 instructions belong to two CPUIDs of AVX512_BF16 and AVX_NE_CONVERT. E.g.

The first one explicitly mentioned it in SDM but the second one doesn't(ISE):
E.g.
![image](https://github.com/llvm/llvm-project/assets/64940021/42ade9cc-c3b1-47ee-8166-811fce2a3a7b)

But since both of them mentioned `does not generate floating-point exceptions and does not consult or update MXCSR.` I suppose AVX_NE_CONVERT instructions as well has such denormal handling.





https://github.com/llvm/llvm-project/pull/90425