[llvm] MachineScheduler: Improve instruction clustering (PR #137784)

[Pengcheng Wang via llvm-commits]

================
@@ -26,26 +26,26 @@ define void @add_v4i32(ptr %x, ptr %y) {
 define void @add_v2i64(ptr %x, ptr %y) {
 ; RV32-LABEL: add_v2i64:
 ; RV32:       # %bb.0:
-; RV32-NEXT:    lw a2, 0(a1)
-; RV32-NEXT:    lw a3, 4(a1)
-; RV32-NEXT:    lw a4, 0(a0)
-; RV32-NEXT:    lw a5, 4(a0)
-; RV32-NEXT:    lw a6, 8(a0)
-; RV32-NEXT:    lw a7, 12(a0)
+; RV32-NEXT:    lw a2, 0(a0)
+; RV32-NEXT:    lw a3, 4(a0)
+; RV32-NEXT:    lw a4, 8(a0)
+; RV32-NEXT:    lw a5, 12(a0)
+; RV32-NEXT:    lw a6, 4(a1)
+; RV32-NEXT:    lw a7, 0(a1)
----------------
wangpc-pp wrote:

Ditto.

https://github.com/llvm/llvm-project/pull/137784