[compiler-rt] [llvm] [SelectionDAG] Add `STRICT_BF16_TO_FP` and `STRICT_FP_TO_BF16` (PR #80056)

[Phoebe Wang via llvm-commits]

================
@@ -0,0 +1,245 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=i686-linux-gnu | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-linux-gnu | FileCheck %s --check-prefix=X64
+; RUN: llc < %s -mtriple=x86_64-linux-gnu -mattr=+avx512bf16,+avx512vl | FileCheck %s --check-prefix=X64-AVX512
----------------
phoebewang wrote:

I see the only different between them are vaddss vs. addss. I think we can add a avx2 to the first one:
```
; RUN: llc < %s -mtriple=x86_64-linux-gnu -mattr=+avx2 | FileCheck %s --check-prefix=X64
; RUN: llc < %s -mtriple=x86_64-linux-gnu -mattr=+avx512bf16,+avx512vl | FileCheck %s --check-prefix=X64-AVX512
```
or
```
; RUN: llc < %s -mtriple=x86_64-linux-gnu | FileCheck %s --check-prefixes=X64,X64-SSE
; RUN: llc < %s -mtriple=x86_64-linux-gnu -mattr=+avx512bf16,+avx512vl | FileCheck %s --check-prefixes=X64,X64-AVX512
```

https://github.com/llvm/llvm-project/pull/80056