[llvm] [LV][AArch64] Prefer Fixed over Scalable if cost-model is equal (Neoverse V2) (PR #95819)

[Sjoerd Meijer via llvm-commits]

================
@@ -0,0 +1,86 @@
+; RUN: opt -S < %s -passes=loop-vectorize -force-target-instruction-cost=1 | FileCheck %s
+
+target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128-Fn32"
+target triple = "aarch64-unknown-linux-gnu"
+
+ at a = dso_local local_unnamed_addr global [32000 x float] zeroinitializer, align 64
+ at b = dso_local local_unnamed_addr global [32000 x float] zeroinitializer, align 64
+
+define void @NeoverseV2() #0 {
+; CHECK-LABEL: define void @NeoverseV2(
+; CHECK:       store <4 x float>
+;
+entry:
+  br label %for.body
+
+for.cond.cleanup:
+  ret void
+
+for.body:
+  %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
+  %arrayidx = getelementptr inbounds [32000 x float], ptr @a, i64 0, i64 %indvars.iv
+  %0 = load float, ptr %arrayidx, align 4
+  %arrayidx2 = getelementptr inbounds [32000 x float], ptr @b, i64 0, i64 %indvars.iv
+  %1 = load float, ptr %arrayidx2, align 4
+  %add = fadd fast float %1, %0
+  %2 = add nuw nsw i64 %indvars.iv, 16000
+  %arrayidx5 = getelementptr inbounds [32000 x float], ptr @a, i64 0, i64 %2
+  store float %add, ptr %arrayidx5, align 4
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %exitcond.not = icmp eq i64 %indvars.iv.next, 16000
+  br i1 %exitcond.not, label %for.cond.cleanup, label %for.body
+}
+
+define void @GenericCPU() #1 {
+; CHECK-LABEL: define void @GenericCPU(
+; CHECK:       store <vscale x 4 x float>
+;
+entry:
+  br label %for.body
+
+for.cond.cleanup:
+  ret void
+
+for.body:
+  %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
+  %arrayidx = getelementptr inbounds [32000 x float], ptr @a, i64 0, i64 %indvars.iv
+  %0 = load float, ptr %arrayidx, align 4
+  %arrayidx2 = getelementptr inbounds [32000 x float], ptr @b, i64 0, i64 %indvars.iv
+  %1 = load float, ptr %arrayidx2, align 4
+  %add = fadd fast float %1, %0
+  %2 = add nuw nsw i64 %indvars.iv, 16000
+  %arrayidx5 = getelementptr inbounds [32000 x float], ptr @a, i64 0, i64 %2
+  store float %add, ptr %arrayidx5, align 4
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %exitcond.not = icmp eq i64 %indvars.iv.next, 16000
+  br i1 %exitcond.not, label %for.cond.cleanup, label %for.body
+}
+
+define void @GenericCPUPreferFixed() #2 {
+; CHECK-LABEL: define void @GenericCPUPreferFixed(
+; CHECK:       store <vscale x 4 x float>
----------------
sjoerdmeijer wrote:

Thanks, I have removed this test, and I don't think we lose (much) testing/coverage.

https://github.com/llvm/llvm-project/pull/95819