mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-24 11:42:57 +01:00
db3c93e416
This will remove suboptimal branching from the generated ll/sc loops. The extra simplification pass affects a lot of testcases, which have been modified to accommodate this change: either by modifying the test to become immune to the CFG simplification, or (less preferablt) by adding option -hexagon-initial-cfg-clenaup=0. llvm-svn: 338774
56 lines
1.9 KiB
LLVM
56 lines
1.9 KiB
LLVM
; RUN: llc -march=hexagon -enable-pipeliner -stats -o /dev/null < %s 2>&1 | FileCheck %s --check-prefix=STATS
|
|
; REQUIRES: asserts
|
|
|
|
; STATS: 1 pipeliner - Number of loops software pipelined
|
|
|
|
; Function Attrs: nounwind
|
|
define i64 @f0(i32 %a0, i32* %a1) #0 {
|
|
b0:
|
|
%v0 = icmp slt i32 %a0, 123469
|
|
br i1 %v0, label %b1, label %b4
|
|
|
|
b1: ; preds = %b0
|
|
br label %b2
|
|
|
|
b2: ; preds = %b2, %b1
|
|
%v1 = phi i64 [ undef, %b1 ], [ %v12, %b2 ]
|
|
%v2 = phi i64 [ undef, %b1 ], [ %v10, %b2 ]
|
|
%v3 = phi i32 [ 0, %b1 ], [ %v13, %b2 ]
|
|
%v4 = phi i32 [ undef, %b1 ], [ %v9, %b2 ]
|
|
%v5 = phi i64 [ undef, %b1 ], [ %v7, %b2 ]
|
|
%v6 = phi i64 [ undef, %b1 ], [ %v11, %b2 ]
|
|
%v7 = tail call i64 @llvm.hexagon.M2.vdmacs.s0(i64 %v5, i64 %v6, i64 %v6)
|
|
%v8 = tail call i64 @llvm.hexagon.S2.packhl(i32 undef, i32 %v4)
|
|
%v9 = load i32, i32* %a1, align 4, !tbaa !0
|
|
%v10 = tail call i64 @llvm.hexagon.M2.vdmacs.s0(i64 %v2, i64 %v6, i64 %v8)
|
|
%v11 = tail call i64 @llvm.hexagon.S2.packhl(i32 %v9, i32 undef)
|
|
%v12 = tail call i64 @llvm.hexagon.M2.vdmacs.s0(i64 %v1, i64 %v6, i64 %v11)
|
|
%v13 = add nsw i32 %v3, 1
|
|
%v14 = icmp eq i32 %v13, undef
|
|
br i1 %v14, label %b3, label %b2
|
|
|
|
b3: ; preds = %b2
|
|
%v15 = lshr i64 %v12, 32
|
|
br label %b4
|
|
|
|
b4: ; preds = %b3, %b0
|
|
%v16 = phi i64 [ %v10, %b3 ], [ undef, %b0 ]
|
|
%v17 = phi i64 [ %v7, %b3 ], [ undef, %b0 ]
|
|
%v18 = add i64 %v16, %v17
|
|
ret i64 %v18
|
|
}
|
|
|
|
; Function Attrs: nounwind readnone
|
|
declare i64 @llvm.hexagon.M2.vdmacs.s0(i64, i64, i64) #1
|
|
|
|
; Function Attrs: nounwind readnone
|
|
declare i64 @llvm.hexagon.S2.packhl(i32, i32) #1
|
|
|
|
attributes #0 = { nounwind "target-cpu"="hexagonv55" }
|
|
attributes #1 = { nounwind readnone }
|
|
|
|
!0 = !{!1, !1, i64 0}
|
|
!1 = !{!"int", !2}
|
|
!2 = !{!"omnipotent char", !3}
|
|
!3 = !{!"Simple C/C++ TBAA"}
|