Fangrui Song 2208c97c1b [Hexagon,test] Change llc -march= to -mtriple=
Similar to 806761a7629df268c8aed49657aeccffa6bca449

-mtriple= specifies the full target triple while -march= merely sets the
architecture part of the default target triple, leaving a target triple which
may not make sense.

Therefore, -march= is error-prone and not recommended for tests without a target
triple. The issue has been benign as we recognize $unknown-apple-darwin as ELF instead
of rejecting it outrightly.
2024-12-15 10:20:22 -08:00

51 lines
1.6 KiB
LLVM

; RUN: llc -mtriple=hexagon -enable-pipeliner -stats -o /dev/null < %s 2>&1 -pipeliner-experimental-cg=true | FileCheck %s --check-prefix=STATS
; REQUIRES: asserts
; Test that we don't pipeline, incorrectly, the swap operation.
; STATS-NOT: 1 pipeliner - Number of loops software pipelined
@g0 = common global ptr null, align 4
; Function Attrs: nounwind
define void @f0(i32 %a0, i32 %a1, i32 %a2) #0 {
b0:
%v0 = icmp sgt i32 %a2, 0
br i1 %v0, label %b1, label %b4
b1: ; preds = %b0
%v1 = load ptr, ptr @g0, align 4, !tbaa !0
br label %b2
b2: ; preds = %b2, %b1
%v2 = phi i32 [ %a0, %b1 ], [ %v9, %b2 ]
%v3 = phi i32 [ %a2, %b1 ], [ %v11, %b2 ]
%v4 = phi i32 [ %a1, %b1 ], [ %v10, %b2 ]
%v5 = getelementptr inbounds i32, ptr %v1, i32 %v2
%v6 = load i32, ptr %v5, align 4, !tbaa !4
%v7 = getelementptr inbounds i32, ptr %v1, i32 %v4
%v8 = load i32, ptr %v7, align 4, !tbaa !4
store i32 %v8, ptr %v5, align 4, !tbaa !4
store i32 %v6, ptr %v7, align 4, !tbaa !4
%v9 = add nsw i32 %v2, 1
%v10 = add nsw i32 %v4, 1
%v11 = add nsw i32 %v3, -1
%v12 = icmp sgt i32 %v11, 0
br i1 %v12, label %b2, label %b3
b3: ; preds = %b2
br label %b4
b4: ; preds = %b3, %b0
ret void
}
attributes #0 = { nounwind "target-cpu"="hexagonv55" }
!0 = !{!1, !1, i64 0}
!1 = !{!"any pointer", !2, i64 0}
!2 = !{!"omnipotent char", !3, i64 0}
!3 = !{!"Simple C/C++ TBAA"}
!4 = !{!5, !5, i64 0}
!5 = !{!"int", !2, i64 0}