[llvm] d27a15f - [NFCI][LoopUnrollAndJam] Changing LoopUnrollAndJamPass to a function
Whitney Tsang via llvm-commits
llvm-commits at lists.llvm.org
Thu Jan 9 08:18:53 PST 2020
Author: Whitney Tsang
Date: 2020-01-09T16:18:36Z
New Revision: d27a15fed79c8e8484825ce6c3ce1682daeb1547
URL: https://github.com/llvm/llvm-project/commit/d27a15fed79c8e8484825ce6c3ce1682daeb1547
DIFF: https://github.com/llvm/llvm-project/commit/d27a15fed79c8e8484825ce6c3ce1682daeb1547.diff
LOG: [NFCI][LoopUnrollAndJam] Changing LoopUnrollAndJamPass to a function
pass.
Summary: This patch changes LoopUnrollAndJamPass to a function pass, and
keeps the loops traversal order same as defined in
FunctionToLoopPassAdaptor LoopPassManager.h.
The next patch will change the loop traversal to outer to inner order,
so more loops can be transform.
Discussion in llvm-dev mailing list:
https://groups.google.com/forum/#!topic/llvm-dev/LF4rUjkVI2g
Reviewer: dmgreen, jdoerfert, Meinersbur, kbarton, bmahjour, etiotto
Reviewed By: dmgreen
Subscribers: hiraditya, zzheng, llvm-commits
Tag: LLVM
Differential Revision: https://reviews.llvm.org/D72230
Added:
Modified:
llvm/include/llvm/Transforms/Scalar/LoopUnrollAndJamPass.h
llvm/lib/Passes/PassBuilder.cpp
llvm/lib/Passes/PassRegistry.def
llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
llvm/test/Transforms/LoopUnrollAndJam/dependencies.ll
llvm/test/Transforms/LoopUnrollAndJam/disable.ll
llvm/test/Transforms/LoopUnrollAndJam/pragma-explicit.ll
llvm/test/Transforms/LoopUnrollAndJam/unroll-and-jam.ll
Removed:
################################################################################
diff --git a/llvm/include/llvm/Transforms/Scalar/LoopUnrollAndJamPass.h b/llvm/include/llvm/Transforms/Scalar/LoopUnrollAndJamPass.h
index 7920269b0fb2..233963528595 100644
--- a/llvm/include/llvm/Transforms/Scalar/LoopUnrollAndJamPass.h
+++ b/llvm/include/llvm/Transforms/Scalar/LoopUnrollAndJamPass.h
@@ -15,9 +15,7 @@
namespace llvm {
-class Loop;
-struct LoopStandardAnalysisResults;
-class LPMUpdater;
+class Function;
/// A simple loop rotation transformation.
class LoopUnrollAndJamPass : public PassInfoMixin<LoopUnrollAndJamPass> {
@@ -25,8 +23,7 @@ class LoopUnrollAndJamPass : public PassInfoMixin<LoopUnrollAndJamPass> {
public:
explicit LoopUnrollAndJamPass(int OptLevel = 2) : OptLevel(OptLevel) {}
- PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
- LoopStandardAnalysisResults &AR, LPMUpdater &U);
+ PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM);
};
} // end namespace llvm
diff --git a/llvm/lib/Passes/PassBuilder.cpp b/llvm/lib/Passes/PassBuilder.cpp
index 64d748f94f14..646eb7d26cbd 100644
--- a/llvm/lib/Passes/PassBuilder.cpp
+++ b/llvm/lib/Passes/PassBuilder.cpp
@@ -970,8 +970,7 @@ ModulePassManager PassBuilder::buildModuleOptimizationPipeline(
// across the loop nests.
// We do UnrollAndJam in a separate LPM to ensure it happens before unroll
if (EnableUnrollAndJam && PTO.LoopUnrolling) {
- OptimizePM.addPass(
- createFunctionToLoopPassAdaptor(LoopUnrollAndJamPass(Level)));
+ OptimizePM.addPass(LoopUnrollAndJamPass(Level));
}
OptimizePM.addPass(LoopUnrollPass(
LoopUnrollOptions(Level, /*OnlyWhenForced=*/!PTO.LoopUnrolling,
diff --git a/llvm/lib/Passes/PassRegistry.def b/llvm/lib/Passes/PassRegistry.def
index 66b38872b386..3efb57cd3589 100644
--- a/llvm/lib/Passes/PassRegistry.def
+++ b/llvm/lib/Passes/PassRegistry.def
@@ -235,6 +235,7 @@ FUNCTION_PASS("spec-phis", SpeculateAroundPHIsPass())
FUNCTION_PASS("sroa", SROA())
FUNCTION_PASS("tailcallelim", TailCallElimPass())
FUNCTION_PASS("unreachableblockelim", UnreachableBlockElimPass())
+FUNCTION_PASS("unroll-and-jam", LoopUnrollAndJamPass())
FUNCTION_PASS("verify", VerifierPass())
FUNCTION_PASS("verify<domtree>", DominatorTreeVerifierPass())
FUNCTION_PASS("verify<loops>", LoopVerifierPass())
@@ -307,7 +308,6 @@ LOOP_PASS("simplify-cfg", LoopSimplifyCFGPass())
LOOP_PASS("strength-reduce", LoopStrengthReducePass())
LOOP_PASS("indvars", IndVarSimplifyPass())
LOOP_PASS("irce", IRCEPass())
-LOOP_PASS("unroll-and-jam", LoopUnrollAndJamPass())
LOOP_PASS("unroll-full", LoopFullUnrollPass())
LOOP_PASS("print-access-info", LoopAccessInfoPrinterPass(dbgs()))
LOOP_PASS("print<ddg>", DDGAnalysisPrinterPass(dbgs()))
diff --git a/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp b/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
index 558f63113db6..92ad8dafa5ab 100644
--- a/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
+++ b/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
@@ -427,51 +427,76 @@ tryToUnrollAndJamLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
return UnrollResult;
}
+static bool tryToUnrollAndJamLoop(Function &F, DominatorTree &DT, LoopInfo &LI,
+ ScalarEvolution &SE,
+ const TargetTransformInfo &TTI,
+ AssumptionCache &AC, DependenceInfo &DI,
+ OptimizationRemarkEmitter &ORE,
+ int OptLevel) {
+ bool DidSomething = false;
+
+ // The loop unroll and jam pass requires loops to be in simplified form, and also needs LCSSA.
+ // Since simplification may add new inner loops, it has to run before the
+ // legality and profitability checks. This means running the loop unroll and jam pass
+ // will simplify all loops, regardless of whether anything end up being
+ // unroll and jammed.
+ for (auto &L : LI) {
+ DidSomething |=
+ simplifyLoop(L, &DT, &LI, &SE, &AC, nullptr, false /* PreserveLCSSA */);
+ DidSomething |= formLCSSARecursively(*L, DT, &LI, &SE);
+ }
+
+ SmallPriorityWorklist<Loop *, 4> Worklist;
+ internal::appendLoopsToWorklist(reverse(LI), Worklist);
+ while (!Worklist.empty()) {
+ Loop *L = Worklist.pop_back_val();
+ formLCSSA(*L, DT, &LI, &SE);
+ LoopUnrollResult Result =
+ tryToUnrollAndJamLoop(L, DT, &LI, SE, TTI, AC, DI, ORE, OptLevel);
+ if (Result != LoopUnrollResult::Unmodified)
+ DidSomething = true;
+ }
+
+ return DidSomething;
+}
+
namespace {
-class LoopUnrollAndJam : public LoopPass {
+class LoopUnrollAndJam : public FunctionPass {
public:
static char ID; // Pass ID, replacement for typeid
unsigned OptLevel;
- LoopUnrollAndJam(int OptLevel = 2) : LoopPass(ID), OptLevel(OptLevel) {
+ LoopUnrollAndJam(int OptLevel = 2) : FunctionPass(ID), OptLevel(OptLevel) {
initializeLoopUnrollAndJamPass(*PassRegistry::getPassRegistry());
}
- bool runOnLoop(Loop *L, LPPassManager &LPM) override {
- if (skipLoop(L))
+ bool runOnFunction(Function &F) override {
+ if (skipFunction(F))
return false;
- Function &F = *L->getHeader()->getParent();
-
auto &DT = getAnalysis<DominatorTreeWrapperPass>().getDomTree();
- LoopInfo *LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
+ LoopInfo &LI = getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
ScalarEvolution &SE = getAnalysis<ScalarEvolutionWrapperPass>().getSE();
const TargetTransformInfo &TTI =
getAnalysis<TargetTransformInfoWrapperPass>().getTTI(F);
auto &AC = getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F);
auto &DI = getAnalysis<DependenceAnalysisWrapperPass>().getDI();
- // For the old PM, we can't use OptimizationRemarkEmitter as an analysis
- // pass. Function analyses need to be preserved across loop transformations
- // but ORE cannot be preserved (see comment before the pass definition).
- OptimizationRemarkEmitter ORE(&F);
-
- LoopUnrollResult Result =
- tryToUnrollAndJamLoop(L, DT, LI, SE, TTI, AC, DI, ORE, OptLevel);
+ auto &ORE = getAnalysis<OptimizationRemarkEmitterWrapperPass>().getORE();
- if (Result == LoopUnrollResult::FullyUnrolled)
- LPM.markLoopAsDeleted(*L);
-
- return Result != LoopUnrollResult::Unmodified;
+ return tryToUnrollAndJamLoop(F, DT, LI, SE, TTI, AC, DI, ORE, OptLevel);
}
/// This transformation requires natural loop information & requires that
/// loop preheaders be inserted into the CFG...
void getAnalysisUsage(AnalysisUsage &AU) const override {
- AU.addRequired<AssumptionCacheTracker>();
+ AU.addRequired<DominatorTreeWrapperPass>();
+ AU.addRequired<LoopInfoWrapperPass>();
+ AU.addRequired<ScalarEvolutionWrapperPass>();
AU.addRequired<TargetTransformInfoWrapperPass>();
+ AU.addRequired<AssumptionCacheTracker>();
AU.addRequired<DependenceAnalysisWrapperPass>();
- getLoopAnalysisUsage(AU);
+ AU.addRequired<OptimizationRemarkEmitterWrapperPass>();
}
};
@@ -481,10 +506,13 @@ char LoopUnrollAndJam::ID = 0;
INITIALIZE_PASS_BEGIN(LoopUnrollAndJam, "loop-unroll-and-jam",
"Unroll and Jam loops", false, false)
-INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)
-INITIALIZE_PASS_DEPENDENCY(LoopPass)
+INITIALIZE_PASS_DEPENDENCY(DominatorTreeWrapperPass)
+INITIALIZE_PASS_DEPENDENCY(LoopInfoWrapperPass)
+INITIALIZE_PASS_DEPENDENCY(ScalarEvolutionWrapperPass)
INITIALIZE_PASS_DEPENDENCY(TargetTransformInfoWrapperPass)
+INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)
INITIALIZE_PASS_DEPENDENCY(DependenceAnalysisWrapperPass)
+INITIALIZE_PASS_DEPENDENCY(OptimizationRemarkEmitterWrapperPass)
INITIALIZE_PASS_END(LoopUnrollAndJam, "loop-unroll-and-jam",
"Unroll and Jam loops", false, false)
@@ -492,26 +520,18 @@ Pass *llvm::createLoopUnrollAndJamPass(int OptLevel) {
return new LoopUnrollAndJam(OptLevel);
}
-PreservedAnalyses LoopUnrollAndJamPass::run(Loop &L, LoopAnalysisManager &AM,
- LoopStandardAnalysisResults &AR,
- LPMUpdater &) {
- const auto &FAM =
- AM.getResult<FunctionAnalysisManagerLoopProxy>(L, AR).getManager();
- Function *F = L.getHeader()->getParent();
-
- auto *ORE = FAM.getCachedResult<OptimizationRemarkEmitterAnalysis>(*F);
- // FIXME: This should probably be optional rather than required.
- if (!ORE)
- report_fatal_error(
- "LoopUnrollAndJamPass: OptimizationRemarkEmitterAnalysis not cached at "
- "a higher level");
-
- DependenceInfo DI(F, &AR.AA, &AR.SE, &AR.LI);
-
- LoopUnrollResult Result = tryToUnrollAndJamLoop(
- &L, AR.DT, &AR.LI, AR.SE, AR.TTI, AR.AC, DI, *ORE, OptLevel);
-
- if (Result == LoopUnrollResult::Unmodified)
+PreservedAnalyses LoopUnrollAndJamPass::run(Function &F,
+ FunctionAnalysisManager &AM) {
+ ScalarEvolution &SE = AM.getResult<ScalarEvolutionAnalysis>(F);
+ LoopInfo &LI = AM.getResult<LoopAnalysis>(F);
+ TargetTransformInfo &TTI = AM.getResult<TargetIRAnalysis>(F);
+ AssumptionCache &AC = AM.getResult<AssumptionAnalysis>(F);
+ DominatorTree &DT = AM.getResult<DominatorTreeAnalysis>(F);
+ DependenceInfo &DI = AM.getResult<DependenceAnalysis>(F);
+ OptimizationRemarkEmitter &ORE =
+ AM.getResult<OptimizationRemarkEmitterAnalysis>(F);
+
+ if (!tryToUnrollAndJamLoop(F, DT, LI, SE, TTI, AC, DI, ORE, OptLevel))
return PreservedAnalyses::all();
return getLoopPassPreservedAnalyses();
diff --git a/llvm/test/Transforms/LoopUnrollAndJam/dependencies.ll b/llvm/test/Transforms/LoopUnrollAndJam/dependencies.ll
index 890683043b4b..d584238745cf 100644
--- a/llvm/test/Transforms/LoopUnrollAndJam/dependencies.ll
+++ b/llvm/test/Transforms/LoopUnrollAndJam/dependencies.ll
@@ -1,4 +1,5 @@
; RUN: opt -basicaa -loop-unroll-and-jam -allow-unroll-and-jam -unroll-and-jam-count=4 < %s -S | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='unroll-and-jam' -allow-unroll-and-jam -unroll-and-jam-count=4 < %s -S | FileCheck %s
target datalayout = "e-m:e-p:32:32-i64:64-v128:64:128-a:0:32-n32-S64"
diff --git a/llvm/test/Transforms/LoopUnrollAndJam/disable.ll b/llvm/test/Transforms/LoopUnrollAndJam/disable.ll
index 4a00937b9c58..6e879896c55f 100644
--- a/llvm/test/Transforms/LoopUnrollAndJam/disable.ll
+++ b/llvm/test/Transforms/LoopUnrollAndJam/disable.ll
@@ -1,4 +1,5 @@
; RUN: opt -loop-unroll-and-jam -allow-unroll-and-jam -unroll-and-jam-count=4 -pass-remarks=loop-unroll-and-jam < %s -S 2>&1 | FileCheck %s
+; RUN: opt -passes='unroll-and-jam' -allow-unroll-and-jam -unroll-and-jam-count=4 -pass-remarks=loop-unroll-and-jam < %s -S 2>&1 | FileCheck %s
target datalayout = "e-m:e-p:32:32-i64:64-v128:64:128-a:0:32-n32-S64"
diff --git a/llvm/test/Transforms/LoopUnrollAndJam/pragma-explicit.ll b/llvm/test/Transforms/LoopUnrollAndJam/pragma-explicit.ll
index 5254c779d0f8..f2c12702f9e8 100644
--- a/llvm/test/Transforms/LoopUnrollAndJam/pragma-explicit.ll
+++ b/llvm/test/Transforms/LoopUnrollAndJam/pragma-explicit.ll
@@ -1,4 +1,5 @@
; RUN: opt -loop-unroll-and-jam -allow-unroll-and-jam -unroll-runtime -unroll-partial-threshold=60 < %s -S | FileCheck %s
+; RUN: opt -aa-pipeline=type-based-aa,basic-aa -passes='unroll-and-jam' -allow-unroll-and-jam -unroll-runtime -unroll-partial-threshold=60 < %s -S | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
diff --git a/llvm/test/Transforms/LoopUnrollAndJam/unroll-and-jam.ll b/llvm/test/Transforms/LoopUnrollAndJam/unroll-and-jam.ll
index bdb47c27f1cf..7580b50f8703 100644
--- a/llvm/test/Transforms/LoopUnrollAndJam/unroll-and-jam.ll
+++ b/llvm/test/Transforms/LoopUnrollAndJam/unroll-and-jam.ll
@@ -1,5 +1,6 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -basicaa -tbaa -loop-unroll-and-jam -allow-unroll-and-jam -unroll-and-jam-count=4 -unroll-remainder < %s -S | FileCheck %s
+; RUN: opt -aa-pipeline=type-based-aa,basic-aa -passes='unroll-and-jam' -allow-unroll-and-jam -unroll-and-jam-count=4 -unroll-remainder < %s -S | FileCheck %s
target datalayout = "e-m:e-p:32:32-i64:64-v128:64:128-a:0:32-n32-S64"
More information about the llvm-commits
mailing list