maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/Transforms/DFAJumpThreading/dfa-unfold-select.ll

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

294 lines
14 KiB
LLVM
Raw Normal View History

Add jump-threading optimization for deterministic finite automata The current JumpThreading pass does not jump thread loops since it can result in irreducible control flow that harms other optimizations. This prevents switch statements inside a loop from being optimized to use unconditional branches. This code pattern occurs in the core_state_transition function of Coremark. The state machine can be implemented manually with goto statements resulting in a large runtime improvement, and this transform makes the switch implementation match the goto version in performance. This patch specifically targets switch statements inside a loop that have the opportunity to be threaded. Once it identifies an opportunity, it creates new paths that branch directly to the correct code block. For example, the left CFG could be transformed to the right CFG: ``` sw.bb sw.bb / | \ / | \ case1 case2 case3 case1 case2 case3 \ | / / | \ latch.bb latch.2 latch.3 latch.1 br sw.bb / | \ sw.bb.2 sw.bb.3 sw.bb.1 br case2 br case3 br case1 ``` Co-author: Justin Kreiner @jkreiner Co-author: Ehsan Amiri @amehsan Reviewed By: SjoerdMeijer Differential Revision: https://reviews.llvm.org/D99205
2021-07-28 02:26:49 +08:00
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -S -dfa-jump-threading %s | FileCheck %s
; These tests check if selects are unfolded properly for jump threading
; opportunities. There are three different patterns to consider:
; 1) Both operands are constant and the false branch is unfolded by default
; 2) One operand is constant and the other is another select to be unfolded. In
; this case a single select is sunk to a new block to unfold.
; 3) Both operands are a select, and both should be sunk to new blocks.
define i32 @test1(i32 %num) {
; CHECK-LABEL: @test1(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[COUNT:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC:%.*]], [[FOR_INC:%.*]] ]
; CHECK-NEXT: [[STATE:%.*]] = phi i32 [ 1, [[ENTRY]] ], [ undef, [[FOR_INC]] ]
; CHECK-NEXT: switch i32 [[STATE]], label [[FOR_INC_JT1:%.*]] [
; CHECK-NEXT: i32 1, label [[CASE1:%.*]]
; CHECK-NEXT: i32 2, label [[CASE2:%.*]]
; CHECK-NEXT: ]
; CHECK: for.body.jt2:
; CHECK-NEXT: [[COUNT_JT2:%.*]] = phi i32 [ [[INC_JT2:%.*]], [[FOR_INC_JT2:%.*]] ]
; CHECK-NEXT: [[STATE_JT2:%.*]] = phi i32 [ [[STATE_NEXT_JT2:%.*]], [[FOR_INC_JT2]] ]
; CHECK-NEXT: br label [[CASE2]]
; CHECK: for.body.jt1:
; CHECK-NEXT: [[COUNT_JT1:%.*]] = phi i32 [ [[INC_JT1:%.*]], [[FOR_INC_JT1]] ]
; CHECK-NEXT: [[STATE_JT1:%.*]] = phi i32 [ [[STATE_NEXT_JT1:%.*]], [[FOR_INC_JT1]] ]
; CHECK-NEXT: br label [[CASE1]]
; CHECK: case1:
; CHECK-NEXT: [[COUNT2:%.*]] = phi i32 [ [[COUNT_JT1]], [[FOR_BODY_JT1:%.*]] ], [ [[COUNT]], [[FOR_BODY]] ]
; CHECK-NEXT: br label [[FOR_INC_JT2]]
; CHECK: case2:
; CHECK-NEXT: [[COUNT1:%.*]] = phi i32 [ [[COUNT_JT2]], [[FOR_BODY_JT2:%.*]] ], [ [[COUNT]], [[FOR_BODY]] ]
; CHECK-NEXT: [[CMP:%.*]] = icmp slt i32 [[COUNT1]], 50
; CHECK-NEXT: br i1 [[CMP]], label [[FOR_INC_JT1]], label [[SI_UNFOLD_FALSE:%.*]]
; CHECK: si.unfold.false:
; CHECK-NEXT: br label [[FOR_INC_JT2]]
; CHECK: for.inc:
; CHECK-NEXT: [[INC]] = add nsw i32 undef, 1
; CHECK-NEXT: [[CMP_EXIT:%.*]] = icmp slt i32 [[INC]], [[NUM:%.*]]
; CHECK-NEXT: br i1 [[CMP_EXIT]], label [[FOR_BODY]], label [[FOR_END:%.*]]
; CHECK: for.inc.jt2:
; CHECK-NEXT: [[COUNT4:%.*]] = phi i32 [ [[COUNT1]], [[SI_UNFOLD_FALSE]] ], [ [[COUNT2]], [[CASE1]] ]
; CHECK-NEXT: [[STATE_NEXT_JT2]] = phi i32 [ 2, [[CASE1]] ], [ 2, [[SI_UNFOLD_FALSE]] ]
; CHECK-NEXT: [[INC_JT2]] = add nsw i32 [[COUNT4]], 1
; CHECK-NEXT: [[CMP_EXIT_JT2:%.*]] = icmp slt i32 [[INC_JT2]], [[NUM]]
; CHECK-NEXT: br i1 [[CMP_EXIT_JT2]], label [[FOR_BODY_JT2]], label [[FOR_END]]
; CHECK: for.inc.jt1:
; CHECK-NEXT: [[COUNT3:%.*]] = phi i32 [ [[COUNT1]], [[CASE2]] ], [ [[COUNT]], [[FOR_BODY]] ]
; CHECK-NEXT: [[STATE_NEXT_JT1]] = phi i32 [ 1, [[CASE2]] ], [ 1, [[FOR_BODY]] ]
; CHECK-NEXT: [[INC_JT1]] = add nsw i32 [[COUNT3]], 1
; CHECK-NEXT: [[CMP_EXIT_JT1:%.*]] = icmp slt i32 [[INC_JT1]], [[NUM]]
; CHECK-NEXT: br i1 [[CMP_EXIT_JT1]], label [[FOR_BODY_JT1]], label [[FOR_END]]
; CHECK: for.end:
; CHECK-NEXT: ret i32 0
;
entry:
br label %for.body
for.body:
%count = phi i32 [ 0, %entry ], [ %inc, %for.inc ]
%state = phi i32 [ 1, %entry ], [ %state.next, %for.inc ]
switch i32 %state, label %for.inc [
i32 1, label %case1
i32 2, label %case2
]
case1:
br label %for.inc
case2:
%cmp = icmp slt i32 %count, 50
%sel = select i1 %cmp, i32 1, i32 2
br label %for.inc
for.inc:
%state.next = phi i32 [ %sel, %case2 ], [ 1, %for.body ], [ 2, %case1 ]
%inc = add nsw i32 %count, 1
%cmp.exit = icmp slt i32 %inc, %num
br i1 %cmp.exit, label %for.body, label %for.end
for.end:
ret i32 0
}
define i32 @test2(i32 %num) {
; CHECK-LABEL: @test2(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[COUNT:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC:%.*]], [[FOR_INC:%.*]] ]
; CHECK-NEXT: [[STATE:%.*]] = phi i32 [ 1, [[ENTRY]] ], [ undef, [[FOR_INC]] ]
; CHECK-NEXT: switch i32 [[STATE]], label [[FOR_INC_JT1:%.*]] [
; CHECK-NEXT: i32 1, label [[CASE1:%.*]]
; CHECK-NEXT: i32 2, label [[CASE2:%.*]]
; CHECK-NEXT: ]
; CHECK: for.body.jt3:
; CHECK-NEXT: [[COUNT_JT3:%.*]] = phi i32 [ [[INC_JT3:%.*]], [[FOR_INC_JT3:%.*]] ]
; CHECK-NEXT: [[STATE_JT3:%.*]] = phi i32 [ [[STATE_NEXT_JT3:%.*]], [[FOR_INC_JT3]] ]
; CHECK-NEXT: br label [[FOR_INC_JT1]]
; CHECK: for.body.jt2:
; CHECK-NEXT: [[COUNT_JT2:%.*]] = phi i32 [ [[INC_JT2:%.*]], [[FOR_INC_JT2:%.*]] ]
; CHECK-NEXT: [[STATE_JT2:%.*]] = phi i32 [ [[STATE_NEXT_JT2:%.*]], [[FOR_INC_JT2]] ]
; CHECK-NEXT: br label [[CASE2]]
; CHECK: for.body.jt1:
; CHECK-NEXT: [[COUNT_JT1:%.*]] = phi i32 [ [[INC_JT1:%.*]], [[FOR_INC_JT1]] ]
; CHECK-NEXT: [[STATE_JT1:%.*]] = phi i32 [ [[STATE_NEXT_JT1:%.*]], [[FOR_INC_JT1]] ]
; CHECK-NEXT: br label [[CASE1]]
; CHECK: case1:
; CHECK-NEXT: [[COUNT5:%.*]] = phi i32 [ [[COUNT_JT1]], [[FOR_BODY_JT1:%.*]] ], [ [[COUNT]], [[FOR_BODY]] ]
; CHECK-NEXT: [[CMP_C1:%.*]] = icmp slt i32 [[COUNT5]], 50
; CHECK-NEXT: [[CMP2_C1:%.*]] = icmp slt i32 [[COUNT5]], 100
; CHECK-NEXT: br i1 [[CMP2_C1]], label [[SI_UNFOLD_TRUE:%.*]], label [[FOR_INC_JT3]]
; CHECK: case2:
; CHECK-NEXT: [[COUNT3:%.*]] = phi i32 [ [[COUNT_JT2]], [[FOR_BODY_JT2:%.*]] ], [ [[COUNT]], [[FOR_BODY]] ]
; CHECK-NEXT: [[CMP_C2:%.*]] = icmp slt i32 [[COUNT3]], 50
; CHECK-NEXT: [[CMP2_C2:%.*]] = icmp sgt i32 [[COUNT3]], 100
; CHECK-NEXT: br i1 [[CMP2_C2]], label [[FOR_INC_JT3]], label [[SI_UNFOLD_FALSE:%.*]]
; CHECK: si.unfold.false:
; CHECK-NEXT: br i1 [[CMP_C2]], label [[FOR_INC_JT1]], label [[SI_UNFOLD_FALSE1:%.*]]
; CHECK: si.unfold.false1:
; CHECK-NEXT: br label [[FOR_INC_JT2]]
; CHECK: si.unfold.true:
; CHECK-NEXT: br i1 [[CMP_C1]], label [[FOR_INC_JT1]], label [[SI_UNFOLD_FALSE2:%.*]]
; CHECK: si.unfold.false2:
; CHECK-NEXT: br label [[FOR_INC_JT2]]
; CHECK: for.inc:
; CHECK-NEXT: [[INC]] = add nsw i32 undef, 1
; CHECK-NEXT: [[CMP_EXIT:%.*]] = icmp slt i32 [[INC]], [[NUM:%.*]]
; CHECK-NEXT: br i1 [[CMP_EXIT]], label [[FOR_BODY]], label [[FOR_END:%.*]]
; CHECK: for.inc.jt3:
; CHECK-NEXT: [[COUNT6:%.*]] = phi i32 [ [[COUNT3]], [[CASE2]] ], [ [[COUNT5]], [[CASE1]] ]
; CHECK-NEXT: [[STATE_NEXT_JT3]] = phi i32 [ 3, [[CASE1]] ], [ 3, [[CASE2]] ]
; CHECK-NEXT: [[INC_JT3]] = add nsw i32 [[COUNT6]], 1
; CHECK-NEXT: [[CMP_EXIT_JT3:%.*]] = icmp slt i32 [[INC_JT3]], [[NUM]]
; CHECK-NEXT: br i1 [[CMP_EXIT_JT3]], label [[FOR_BODY_JT3:%.*]], label [[FOR_END]]
; CHECK: for.inc.jt2:
; CHECK-NEXT: [[COUNT7:%.*]] = phi i32 [ [[COUNT3]], [[SI_UNFOLD_FALSE1]] ], [ [[COUNT5]], [[SI_UNFOLD_FALSE2]] ]
; CHECK-NEXT: [[STATE_NEXT_JT2]] = phi i32 [ 2, [[SI_UNFOLD_FALSE1]] ], [ 2, [[SI_UNFOLD_FALSE2]] ]
; CHECK-NEXT: [[INC_JT2]] = add nsw i32 [[COUNT7]], 1
; CHECK-NEXT: [[CMP_EXIT_JT2:%.*]] = icmp slt i32 [[INC_JT2]], [[NUM]]
; CHECK-NEXT: br i1 [[CMP_EXIT_JT2]], label [[FOR_BODY_JT2]], label [[FOR_END]]
; CHECK: for.inc.jt1:
; CHECK-NEXT: [[COUNT4:%.*]] = phi i32 [ [[COUNT_JT3]], [[FOR_BODY_JT3]] ], [ [[COUNT3]], [[SI_UNFOLD_FALSE]] ], [ [[COUNT5]], [[SI_UNFOLD_TRUE]] ], [ [[COUNT]], [[FOR_BODY]] ]
; CHECK-NEXT: [[STATE_NEXT_JT1]] = phi i32 [ 1, [[FOR_BODY]] ], [ 1, [[SI_UNFOLD_FALSE]] ], [ 1, [[SI_UNFOLD_TRUE]] ], [ 1, [[FOR_BODY_JT3]] ]
; CHECK-NEXT: [[INC_JT1]] = add nsw i32 [[COUNT4]], 1
; CHECK-NEXT: [[CMP_EXIT_JT1:%.*]] = icmp slt i32 [[INC_JT1]], [[NUM]]
; CHECK-NEXT: br i1 [[CMP_EXIT_JT1]], label [[FOR_BODY_JT1]], label [[FOR_END]]
; CHECK: for.end:
; CHECK-NEXT: ret i32 0
;
entry:
br label %for.body
for.body:
%count = phi i32 [ 0, %entry ], [ %inc, %for.inc ]
%state = phi i32 [ 1, %entry ], [ %state.next, %for.inc ]
switch i32 %state, label %for.inc [
i32 1, label %case1
i32 2, label %case2
]
case1:
%cmp.c1 = icmp slt i32 %count, 50
%cmp2.c1 = icmp slt i32 %count, 100
%state1.1 = select i1 %cmp.c1, i32 1, i32 2
%state1.2 = select i1 %cmp2.c1, i32 %state1.1, i32 3
br label %for.inc
case2:
%cmp.c2 = icmp slt i32 %count, 50
%cmp2.c2 = icmp sgt i32 %count, 100
%state2.1 = select i1 %cmp.c2, i32 1, i32 2
%state2.2 = select i1 %cmp2.c2, i32 3, i32 %state2.1
br label %for.inc
for.inc:
%state.next = phi i32 [ %state1.2, %case1 ], [ %state2.2, %case2 ], [ 1, %for.body ]
%inc = add nsw i32 %count, 1
%cmp.exit = icmp slt i32 %inc, %num
br i1 %cmp.exit, label %for.body, label %for.end
for.end:
ret i32 0
}
define i32 @test3(i32 %num) {
; CHECK-LABEL: @test3(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[COUNT:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC:%.*]], [[FOR_INC:%.*]] ]
; CHECK-NEXT: [[STATE:%.*]] = phi i32 [ 1, [[ENTRY]] ], [ undef, [[FOR_INC]] ]
; CHECK-NEXT: switch i32 [[STATE]], label [[FOR_INC_JT1:%.*]] [
; CHECK-NEXT: i32 1, label [[CASE1:%.*]]
; CHECK-NEXT: i32 2, label [[CASE2:%.*]]
; CHECK-NEXT: ]
; CHECK: for.body.jt4:
; CHECK-NEXT: [[COUNT_JT4:%.*]] = phi i32 [ [[INC_JT4:%.*]], [[FOR_INC_JT4:%.*]] ]
; CHECK-NEXT: [[STATE_JT4:%.*]] = phi i32 [ [[STATE_NEXT_JT4:%.*]], [[FOR_INC_JT4]] ]
; CHECK-NEXT: br label [[FOR_INC_JT1]]
; CHECK: for.body.jt3:
; CHECK-NEXT: [[COUNT_JT3:%.*]] = phi i32 [ [[INC_JT3:%.*]], [[FOR_INC_JT3:%.*]] ]
; CHECK-NEXT: [[STATE_JT3:%.*]] = phi i32 [ [[STATE_NEXT_JT3:%.*]], [[FOR_INC_JT3]] ]
; CHECK-NEXT: br label [[FOR_INC_JT1]]
; CHECK: for.body.jt2:
; CHECK-NEXT: [[COUNT_JT2:%.*]] = phi i32 [ [[INC_JT2:%.*]], [[FOR_INC_JT2:%.*]] ]
; CHECK-NEXT: [[STATE_JT2:%.*]] = phi i32 [ [[STATE_NEXT_JT2:%.*]], [[FOR_INC_JT2]] ]
; CHECK-NEXT: br label [[CASE2]]
; CHECK: for.body.jt1:
; CHECK-NEXT: [[COUNT_JT1:%.*]] = phi i32 [ [[INC_JT1:%.*]], [[FOR_INC_JT1]] ]
; CHECK-NEXT: [[STATE_JT1:%.*]] = phi i32 [ [[STATE_NEXT_JT1:%.*]], [[FOR_INC_JT1]] ]
; CHECK-NEXT: br label [[CASE1]]
; CHECK: case1:
; CHECK-NEXT: [[COUNT5:%.*]] = phi i32 [ [[COUNT_JT1]], [[FOR_BODY_JT1:%.*]] ], [ [[COUNT]], [[FOR_BODY]] ]
; CHECK-NEXT: br label [[FOR_INC_JT2]]
; CHECK: case2:
; CHECK-NEXT: [[COUNT4:%.*]] = phi i32 [ [[COUNT_JT2]], [[FOR_BODY_JT2:%.*]] ], [ [[COUNT]], [[FOR_BODY]] ]
; CHECK-NEXT: [[CMP_1:%.*]] = icmp slt i32 [[COUNT4]], 50
; CHECK-NEXT: [[CMP_2:%.*]] = icmp slt i32 [[COUNT4]], 100
; CHECK-NEXT: [[TMP0:%.*]] = and i32 [[COUNT4]], 1
; CHECK-NEXT: [[CMP_3:%.*]] = icmp eq i32 [[TMP0]], 0
; CHECK-NEXT: br i1 [[CMP_3]], label [[SI_UNFOLD_TRUE:%.*]], label [[SI_UNFOLD_FALSE:%.*]]
; CHECK: si.unfold.true:
; CHECK-NEXT: br i1 [[CMP_1]], label [[FOR_INC_JT1]], label [[SI_UNFOLD_FALSE2:%.*]]
; CHECK: si.unfold.false:
; CHECK-NEXT: br i1 [[CMP_2]], label [[FOR_INC_JT3]], label [[SI_UNFOLD_FALSE1:%.*]]
; CHECK: si.unfold.false1:
; CHECK-NEXT: br label [[FOR_INC_JT4]]
; CHECK: si.unfold.false2:
; CHECK-NEXT: br label [[FOR_INC_JT2]]
; CHECK: for.inc:
; CHECK-NEXT: [[INC]] = add nsw i32 undef, 1
; CHECK-NEXT: [[CMP_EXIT:%.*]] = icmp slt i32 [[INC]], [[NUM:%.*]]
; CHECK-NEXT: br i1 [[CMP_EXIT]], label [[FOR_BODY]], label [[FOR_END:%.*]]
; CHECK: for.inc.jt4:
; CHECK-NEXT: [[STATE_NEXT_JT4]] = phi i32 [ 4, [[SI_UNFOLD_FALSE1]] ]
; CHECK-NEXT: [[INC_JT4]] = add nsw i32 [[COUNT4]], 1
; CHECK-NEXT: [[CMP_EXIT_JT4:%.*]] = icmp slt i32 [[INC_JT4]], [[NUM]]
; CHECK-NEXT: br i1 [[CMP_EXIT_JT4]], label [[FOR_BODY_JT4:%.*]], label [[FOR_END]]
; CHECK: for.inc.jt3:
; CHECK-NEXT: [[STATE_NEXT_JT3]] = phi i32 [ 3, [[SI_UNFOLD_FALSE]] ]
; CHECK-NEXT: [[INC_JT3]] = add nsw i32 [[COUNT4]], 1
; CHECK-NEXT: [[CMP_EXIT_JT3:%.*]] = icmp slt i32 [[INC_JT3]], [[NUM]]
; CHECK-NEXT: br i1 [[CMP_EXIT_JT3]], label [[FOR_BODY_JT3:%.*]], label [[FOR_END]]
; CHECK: for.inc.jt2:
; CHECK-NEXT: [[COUNT6:%.*]] = phi i32 [ [[COUNT4]], [[SI_UNFOLD_FALSE2]] ], [ [[COUNT5]], [[CASE1]] ]
; CHECK-NEXT: [[STATE_NEXT_JT2]] = phi i32 [ 2, [[CASE1]] ], [ 2, [[SI_UNFOLD_FALSE2]] ]
; CHECK-NEXT: [[INC_JT2]] = add nsw i32 [[COUNT6]], 1
; CHECK-NEXT: [[CMP_EXIT_JT2:%.*]] = icmp slt i32 [[INC_JT2]], [[NUM]]
; CHECK-NEXT: br i1 [[CMP_EXIT_JT2]], label [[FOR_BODY_JT2]], label [[FOR_END]]
; CHECK: for.inc.jt1:
; CHECK-NEXT: [[COUNT3:%.*]] = phi i32 [ [[COUNT_JT4]], [[FOR_BODY_JT4]] ], [ [[COUNT_JT3]], [[FOR_BODY_JT3]] ], [ [[COUNT4]], [[SI_UNFOLD_TRUE]] ], [ [[COUNT]], [[FOR_BODY]] ]
; CHECK-NEXT: [[STATE_NEXT_JT1]] = phi i32 [ 1, [[FOR_BODY]] ], [ 1, [[SI_UNFOLD_TRUE]] ], [ 1, [[FOR_BODY_JT3]] ], [ 1, [[FOR_BODY_JT4]] ]
; CHECK-NEXT: [[INC_JT1]] = add nsw i32 [[COUNT3]], 1
; CHECK-NEXT: [[CMP_EXIT_JT1:%.*]] = icmp slt i32 [[INC_JT1]], [[NUM]]
; CHECK-NEXT: br i1 [[CMP_EXIT_JT1]], label [[FOR_BODY_JT1]], label [[FOR_END]]
; CHECK: for.end:
; CHECK-NEXT: ret i32 0
;
entry:
br label %for.body
for.body:
%count = phi i32 [ 0, %entry ], [ %inc, %for.inc ]
%state = phi i32 [ 1, %entry ], [ %state.next, %for.inc ]
switch i32 %state, label %for.inc [
i32 1, label %case1
i32 2, label %case2
]
case1:
br label %for.inc
case2:
%cmp.1 = icmp slt i32 %count, 50
%cmp.2 = icmp slt i32 %count, 100
%0 = and i32 %count, 1
%cmp.3 = icmp eq i32 %0, 0
%sel.1 = select i1 %cmp.1, i32 1, i32 2
%sel.2 = select i1 %cmp.2, i32 3, i32 4
%sel.3 = select i1 %cmp.3, i32 %sel.1, i32 %sel.2
br label %for.inc
for.inc:
%state.next = phi i32 [ %sel.3, %case2 ], [ 1, %for.body ], [ 2, %case1 ]
%inc = add nsw i32 %count, 1
%cmp.exit = icmp slt i32 %inc, %num
br i1 %cmp.exit, label %for.body, label %for.end
for.end:
ret i32 0
}