[llvm] [X86] Lower `minimum`/`maximum`/`minimumnum`/`maximumnum` using bitwise operations (PR #170069)

[via llvm-commits]

================
@@ -1782,4 +1228,5 @@ declare double @llvm.vector.reduce.fmaximum.v4f64(<4 x double>)
 declare double @llvm.vector.reduce.fmaximum.v8f64(<8 x double>)
 declare double @llvm.vector.reduce.fmaximum.v16f64(<16 x double>)
 ;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; AVX512: {{.*}}
----------------
valadaptive wrote:

There are some other issues with the RUN lines; the `SSE` prefix is also unused and there's a new warning about conflicting output for the two `AVX` lines (I think the `+avx2` one was meant to use the `AVX2` prefix.) Should I fix those in a separate PR?

https://github.com/llvm/llvm-project/pull/170069