[llvm] [AMDGPU] Code clean up and denorm check removal for bf16 omod folding (PR #219042)

[Jay Foad via llvm-commits]

================
@@ -2505,8 +2505,9 @@ SIFoldOperandsImpl::isOMod(const MachineInstr &MI) const {
   }
   case AMDGPU::V_PK_MUL_BF16: {
     // OMOD folding for BF16 packed multiply
-    if (MFI->getMode().FP32Denormals.Output != DenormalMode::PreserveSign ||
-        MI.mayRaiseFPException())
+    // PK_*_BF16 ignores DENORM controls, and always preserves denorms
----------------
jayfoad wrote:

> If I understand it correctly omod simply adds to the exponential and does nothing else. So, it almost guaranteed to produce a denorm. Correct me if I am wrong.

The docs say that OMOD flushes output denorms, so either you're wrong or the docs are wrnog :)

https://github.com/llvm/llvm-project/pull/219042