mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2025-02-01 13:11:39 +01:00
e37295579e
Software pipelining is an optimization for improving ILP by overlapping loop iterations. Swing Modulo Scheduling (SMS) is an implementation of software pipelining that attempts to reduce register pressure and generate efficient pipelines with a low compile-time cost. This implementaion of SMS is a target-independent back-end pass. When enabled, the pass should run just prior to the register allocation pass, while the machine IR is in SSA form. If the pass is successful, then the original loop is replaced by the optimized loop. The optimized loop contains one or more prolog blocks, the pipelined kernel, and one or more epilog blocks. This pass is enabled for Hexagon only. To enable for other targets, a couple of target specific hooks must be implemented, and the pass needs to be called from the target's TargetMachine implementation. Differential Review: http://reviews.llvm.org/D16829 llvm-svn: 277169
162 lines
4.5 KiB
LLVM
162 lines
4.5 KiB
LLVM
; RUN: llc -march=hexagon -enable-pipeliner=false < %s | FileCheck %s
|
|
; Check that we generate hardware loop instructions.
|
|
|
|
; Case 1 : Loop with a constant number of iterations.
|
|
; CHECK-LABEL: @hwloop1
|
|
; CHECK: loop0(.LBB{{.}}_{{.}}, #10)
|
|
; CHECK: endloop0
|
|
|
|
@a = common global [10 x i32] zeroinitializer, align 4
|
|
define i32 @hwloop1() nounwind {
|
|
entry:
|
|
br label %for.body
|
|
for.body:
|
|
%i.01 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
|
|
%arrayidx = getelementptr inbounds [10 x i32], [10 x i32]* @a, i32 0, i32 %i.01
|
|
store i32 %i.01, i32* %arrayidx, align 4
|
|
%inc = add nsw i32 %i.01, 1
|
|
%exitcond = icmp eq i32 %inc, 10
|
|
br i1 %exitcond, label %for.end, label %for.body
|
|
for.end:
|
|
ret i32 0
|
|
}
|
|
|
|
; Case 2 : Loop with a run-time number of iterations.
|
|
; CHECK-LABEL: @hwloop2
|
|
; CHECK: loop0(.LBB{{.}}_{{.}}, r{{[0-9]+}})
|
|
; CHECK: endloop0
|
|
|
|
define i32 @hwloop2(i32 %n, i32* nocapture %b) nounwind {
|
|
entry:
|
|
%cmp1 = icmp sgt i32 %n, 0
|
|
br i1 %cmp1, label %for.body.preheader, label %for.end
|
|
|
|
for.body.preheader:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%a.03 = phi i32 [ %add, %for.body ], [ 0, %for.body.preheader ]
|
|
%i.02 = phi i32 [ %inc, %for.body ], [ 0, %for.body.preheader ]
|
|
%arrayidx = getelementptr inbounds i32, i32* %b, i32 %i.02
|
|
%0 = load i32, i32* %arrayidx, align 4
|
|
%add = add nsw i32 %0, %a.03
|
|
%inc = add nsw i32 %i.02, 1
|
|
%exitcond = icmp eq i32 %inc, %n
|
|
br i1 %exitcond, label %for.end.loopexit, label %for.body
|
|
|
|
for.end.loopexit:
|
|
br label %for.end
|
|
|
|
for.end:
|
|
%a.0.lcssa = phi i32 [ 0, %entry ], [ %add, %for.end.loopexit ]
|
|
ret i32 %a.0.lcssa
|
|
}
|
|
|
|
; Case 3 : Induction variable increment more than 1.
|
|
; CHECK-LABEL: @hwloop3
|
|
; CHECK: lsr(r{{[0-9]+}}, #2)
|
|
; CHECK: loop0(.LBB{{.}}_{{.}}, r{{[0-9]+}})
|
|
; CHECK: endloop0
|
|
|
|
define i32 @hwloop3(i32 %n, i32* nocapture %b) nounwind {
|
|
entry:
|
|
%cmp1 = icmp sgt i32 %n, 0
|
|
br i1 %cmp1, label %for.body.preheader, label %for.end
|
|
|
|
for.body.preheader:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%a.03 = phi i32 [ %add, %for.body ], [ 0, %for.body.preheader ]
|
|
%i.02 = phi i32 [ %inc, %for.body ], [ 0, %for.body.preheader ]
|
|
%arrayidx = getelementptr inbounds i32, i32* %b, i32 %i.02
|
|
%0 = load i32, i32* %arrayidx, align 4
|
|
%add = add nsw i32 %0, %a.03
|
|
%inc = add nsw i32 %i.02, 4
|
|
%exitcond = icmp eq i32 %inc, %n
|
|
br i1 %exitcond, label %for.end.loopexit, label %for.body
|
|
|
|
for.end.loopexit:
|
|
br label %for.end
|
|
|
|
for.end:
|
|
%a.0.lcssa = phi i32 [ 0, %entry ], [ %add, %for.end.loopexit ]
|
|
ret i32 %a.0.lcssa
|
|
}
|
|
|
|
; Case 4 : Loop exit compare uses register instead of immediate value.
|
|
; CHECK-LABEL: @hwloop4
|
|
; CHECK: loop0(.LBB{{.}}_{{.}}, r{{[0-9]+}})
|
|
; CHECK: endloop0
|
|
|
|
define i32 @hwloop4(i32 %n, i32* nocapture %b) nounwind {
|
|
entry:
|
|
%cmp1 = icmp sgt i32 %n, 0
|
|
br i1 %cmp1, label %for.body.preheader, label %for.end
|
|
|
|
for.body.preheader:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%i.02 = phi i32 [ %inc, %for.body ], [ 0, %for.body.preheader ]
|
|
%arrayidx = getelementptr inbounds i32, i32* %b, i32 %i.02
|
|
store i32 %i.02, i32* %arrayidx, align 4
|
|
%inc = add nsw i32 %i.02, 1
|
|
%exitcond = icmp eq i32 %inc, %n
|
|
br i1 %exitcond, label %for.end.loopexit, label %for.body
|
|
|
|
for.end.loopexit:
|
|
br label %for.end
|
|
|
|
for.end:
|
|
ret i32 0
|
|
}
|
|
|
|
; Case 5: After LSR, the initial value is 100 and the iv decrements to 0.
|
|
; CHECK-LABEL: @hwloop5
|
|
; CHECK: loop0(.LBB{{.}}_{{.}}, #100)
|
|
; CHECK: endloop0
|
|
|
|
define void @hwloop5(i32* nocapture %a, i32* nocapture %res) nounwind {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%i.03 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
|
|
%arrayidx = getelementptr inbounds i32, i32* %a, i32 %i.03
|
|
%0 = load i32, i32* %arrayidx, align 4
|
|
%mul = mul nsw i32 %0, %0
|
|
%arrayidx2 = getelementptr inbounds i32, i32* %res, i32 %i.03
|
|
store i32 %mul, i32* %arrayidx2, align 4
|
|
%inc = add nsw i32 %i.03, 1
|
|
%exitcond = icmp eq i32 %inc, 100
|
|
br i1 %exitcond, label %for.end, label %for.body
|
|
|
|
for.end:
|
|
ret void
|
|
}
|
|
|
|
; Case 6: Large immediate offset
|
|
; CHECK-LABEL: @hwloop6
|
|
; CHECK-NOT: loop0(.LBB{{.}}_{{.}}, #1024)
|
|
; CHECK: loop0(.LBB{{.}}_{{.}}, r{{[0-9]+}})
|
|
; CHECK: endloop0
|
|
|
|
define void @hwloop6(i32* nocapture %a, i32* nocapture %res) nounwind {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%i.02 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
|
|
%arrayidx = getelementptr inbounds i32, i32* %a, i32 %i.02
|
|
%0 = load i32, i32* %arrayidx, align 4
|
|
%arrayidx1 = getelementptr inbounds i32, i32* %res, i32 %i.02
|
|
store i32 %0, i32* %arrayidx1, align 4
|
|
%inc = add nsw i32 %i.02, 1
|
|
%exitcond = icmp eq i32 %inc, 1024
|
|
br i1 %exitcond, label %for.end, label %for.body
|
|
|
|
for.end:
|
|
ret void
|
|
}
|