[llvm] [AMDGPU] selecting v_sat_pk instruction, version 2 (PR #123297)

[Matt Arsenault via llvm-commits]

================
@@ -5498,6 +5498,7 @@ const char* AMDGPUTargetLowering::getTargetNodeName(unsigned Opcode) const {
   NODE_NAME_CASE(UMIN3)
   NODE_NAME_CASE(FMED3)
   NODE_NAME_CASE(SMED3)
+  NODE_NAME_CASE(SAT_PK_CAST)
----------------
arsenm wrote:

Because the type signature is different. This is forcing the pack to use a legal integer type instead of v2i8 

https://github.com/llvm/llvm-project/pull/123297