2208c97c1b
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.
36 lines
1.1 KiB
LLVM
36 lines
1.1 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
|
|
|
|
; We're unable to pipeline a loop with a subreg as an operand of a Phi.
|
|
|
|
; STATS-NOT: 1 pipeliner - Number of loops software pipelined
|
|
|
|
; Function Attrs: nounwind readnone
|
|
define void @f0(i32 %a0) #0 {
|
|
b0:
|
|
%v0 = sext i32 %a0 to i64
|
|
br label %b1
|
|
|
|
b1: ; preds = %b1, %b0
|
|
%v1 = phi i32 [ 805306368, %b0 ], [ %v12, %b1 ]
|
|
%v2 = phi i32 [ 5, %b0 ], [ %v13, %b1 ]
|
|
%v3 = sext i32 %v1 to i64
|
|
%v4 = mul nsw i64 %v3, %v0
|
|
%v5 = lshr i64 %v4, 32
|
|
%v6 = trunc i64 %v5 to i32
|
|
%v7 = sub nsw i32 536870912, %v6
|
|
%v8 = sext i32 %v7 to i64
|
|
%v9 = mul nsw i64 %v8, %v3
|
|
%v10 = lshr i64 %v9, 32
|
|
%v11 = shl nuw nsw i64 %v10, 4
|
|
%v12 = trunc i64 %v11 to i32
|
|
%v13 = add nsw i32 %v2, -1
|
|
%v14 = icmp eq i32 %v13, 0
|
|
br i1 %v14, label %b2, label %b1
|
|
|
|
b2: ; preds = %b1
|
|
ret void
|
|
}
|
|
|
|
attributes #0 = { nounwind readnone "target-cpu"="hexagonv55" }
|