[llvm] [X86][BF16] Do not lower to VCVTNEPS2BF16 without AVX512VL (PR #86395)

[Simon Pilgrim via llvm-commits]

================
@@ -0,0 +1,74 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 4
+; RUN: llc < %s -mtriple=x86_64-linux-gnu -mattr=avx512bf16 | FileCheck %s
----------------
RKSimon wrote:

worth adding `-mattr=avx512bf16,-mattr=avx512vl` and `-mattr=+avxneconvert` RUN lines as well?

https://github.com/llvm/llvm-project/pull/86395