[llvm] Intrinsic: introduce minimumnum and maximumnum for IR and SelectionDAG (PR #96649)

[Matt Arsenault via llvm-commits]

================
@@ -0,0 +1,2121 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc --mtriple=x86_64 < %s | FileCheck %s --check-prefix=X86-64
+; RUN: llc --mtriple=i386 < %s | FileCheck %s --check-prefix=I386
+; RUN: llc --mtriple=i386 -mattr=+sse2 < %s | FileCheck %s --check-prefix=SSE2
+; RUN: llc --mtriple=i386 -mattr=+avx < %s | FileCheck %s --check-prefix=AVX
----------------
arsenm wrote:

I'd rather just leave x86_32 untested until #100967 has a separate fix 

https://github.com/llvm/llvm-project/pull/96649