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

[Pengcheng Wang via llvm-commits]

================
@@ -14204,23 +14204,23 @@ define <8 x i16> @mgather_strided_unaligned(ptr %base) {
 ; RV64ZVE32F-NEXT:    slli t2, t2, 8
 ; RV64ZVE32F-NEXT:    or a6, t0, a7
 ; RV64ZVE32F-NEXT:    or a2, a4, a2
-; RV64ZVE32F-NEXT:    lbu a4, 24(a0)
+; RV64ZVE32F-NEXT:    or a4, t2, t1
 ; RV64ZVE32F-NEXT:    lbu a7, 25(a0)
-; RV64ZVE32F-NEXT:    or t0, t2, t1
+; RV64ZVE32F-NEXT:    lbu t0, 24(a0)
----------------
wangpc-pp wrote:

Yeah, I agree. My concern is that it may have some impacts on macro fusion, but I don't if it is a real issue. We can revisit this if we encounter some problems in the future.

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