llvm-project/llvm/lib/Transforms/Utils/LCSSA.cpp

373 lines
13 KiB
C++
Raw Normal View History

//===-- LCSSA.cpp - Convert loops into loop-closed SSA form ---------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This pass transforms loops by placing phi nodes at the end of the loops for
// all values that are live across the loop boundary. For example, it turns
// the left into the right code:
//
// for (...) for (...)
2007-05-12 05:10:54 +08:00
// if (c) if (c)
// X1 = ... X1 = ...
// else else
// X2 = ... X2 = ...
// X3 = phi(X1, X2) X3 = phi(X1, X2)
// ... = X3 + 4 X4 = phi(X3)
// ... = X4 + 4
//
// This is still valid LLVM; the extra phi nodes are purely redundant, and will
// be trivially eliminated by InstCombine. The major benefit of this
// transformation is that it makes many other loop optimizations, such as
// LoopUnswitching, simpler.
//
//===----------------------------------------------------------------------===//
#include "llvm/Transforms/Utils/LCSSA.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/Analysis/BasicAliasAnalysis.h"
[PM/AA] Rebuild LLVM's alias analysis infrastructure in a way compatible with the new pass manager, and no longer relying on analysis groups. This builds essentially a ground-up new AA infrastructure stack for LLVM. The core ideas are the same that are used throughout the new pass manager: type erased polymorphism and direct composition. The design is as follows: - FunctionAAResults is a type-erasing alias analysis results aggregation interface to walk a single query across a range of results from different alias analyses. Currently this is function-specific as we always assume that aliasing queries are *within* a function. - AAResultBase is a CRTP utility providing stub implementations of various parts of the alias analysis result concept, notably in several cases in terms of other more general parts of the interface. This can be used to implement only a narrow part of the interface rather than the entire interface. This isn't really ideal, this logic should be hoisted into FunctionAAResults as currently it will cause a significant amount of redundant work, but it faithfully models the behavior of the prior infrastructure. - All the alias analysis passes are ported to be wrapper passes for the legacy PM and new-style analysis passes for the new PM with a shared result object. In some cases (most notably CFL), this is an extremely naive approach that we should revisit when we can specialize for the new pass manager. - BasicAA has been restructured to reflect that it is much more fundamentally a function analysis because it uses dominator trees and loop info that need to be constructed for each function. All of the references to getting alias analysis results have been updated to use the new aggregation interface. All the preservation and other pass management code has been updated accordingly. The way the FunctionAAResultsWrapperPass works is to detect the available alias analyses when run, and add them to the results object. This means that we should be able to continue to respect when various passes are added to the pipeline, for example adding CFL or adding TBAA passes should just cause their results to be available and to get folded into this. The exception to this rule is BasicAA which really needs to be a function pass due to using dominator trees and loop info. As a consequence, the FunctionAAResultsWrapperPass directly depends on BasicAA and always includes it in the aggregation. This has significant implications for preserving analyses. Generally, most passes shouldn't bother preserving FunctionAAResultsWrapperPass because rebuilding the results just updates the set of known AA passes. The exception to this rule are LoopPass instances which need to preserve all the function analyses that the loop pass manager will end up needing. This means preserving both BasicAAWrapperPass and the aggregating FunctionAAResultsWrapperPass. Now, when preserving an alias analysis, you do so by directly preserving that analysis. This is only necessary for non-immutable-pass-provided alias analyses though, and there are only three of interest: BasicAA, GlobalsAA (formerly GlobalsModRef), and SCEVAA. Usually BasicAA is preserved when needed because it (like DominatorTree and LoopInfo) is marked as a CFG-only pass. I've expanded GlobalsAA into the preserved set everywhere we previously were preserving all of AliasAnalysis, and I've added SCEVAA in the intersection of that with where we preserve SCEV itself. One significant challenge to all of this is that the CGSCC passes were actually using the alias analysis implementations by taking advantage of a pretty amazing set of loop holes in the old pass manager's analysis management code which allowed analysis groups to slide through in many cases. Moving away from analysis groups makes this problem much more obvious. To fix it, I've leveraged the flexibility the design of the new PM components provides to just directly construct the relevant alias analyses for the relevant functions in the IPO passes that need them. This is a bit hacky, but should go away with the new pass manager, and is already in many ways cleaner than the prior state. Another significant challenge is that various facilities of the old alias analysis infrastructure just don't fit any more. The most significant of these is the alias analysis 'counter' pass. That pass relied on the ability to snoop on AA queries at different points in the analysis group chain. Instead, I'm planning to build printing functionality directly into the aggregation layer. I've not included that in this patch merely to keep it smaller. Note that all of this needs a nearly complete rewrite of the AA documentation. I'm planning to do that, but I'd like to make sure the new design settles, and to flesh out a bit more of what it looks like in the new pass manager first. Differential Revision: http://reviews.llvm.org/D12080 llvm-svn: 247167
2015-09-10 01:55:00 +08:00
#include "llvm/Analysis/GlobalsModRef.h"
#include "llvm/Analysis/LoopPass.h"
#include "llvm/Analysis/ScalarEvolution.h"
[PM/AA] Rebuild LLVM's alias analysis infrastructure in a way compatible with the new pass manager, and no longer relying on analysis groups. This builds essentially a ground-up new AA infrastructure stack for LLVM. The core ideas are the same that are used throughout the new pass manager: type erased polymorphism and direct composition. The design is as follows: - FunctionAAResults is a type-erasing alias analysis results aggregation interface to walk a single query across a range of results from different alias analyses. Currently this is function-specific as we always assume that aliasing queries are *within* a function. - AAResultBase is a CRTP utility providing stub implementations of various parts of the alias analysis result concept, notably in several cases in terms of other more general parts of the interface. This can be used to implement only a narrow part of the interface rather than the entire interface. This isn't really ideal, this logic should be hoisted into FunctionAAResults as currently it will cause a significant amount of redundant work, but it faithfully models the behavior of the prior infrastructure. - All the alias analysis passes are ported to be wrapper passes for the legacy PM and new-style analysis passes for the new PM with a shared result object. In some cases (most notably CFL), this is an extremely naive approach that we should revisit when we can specialize for the new pass manager. - BasicAA has been restructured to reflect that it is much more fundamentally a function analysis because it uses dominator trees and loop info that need to be constructed for each function. All of the references to getting alias analysis results have been updated to use the new aggregation interface. All the preservation and other pass management code has been updated accordingly. The way the FunctionAAResultsWrapperPass works is to detect the available alias analyses when run, and add them to the results object. This means that we should be able to continue to respect when various passes are added to the pipeline, for example adding CFL or adding TBAA passes should just cause their results to be available and to get folded into this. The exception to this rule is BasicAA which really needs to be a function pass due to using dominator trees and loop info. As a consequence, the FunctionAAResultsWrapperPass directly depends on BasicAA and always includes it in the aggregation. This has significant implications for preserving analyses. Generally, most passes shouldn't bother preserving FunctionAAResultsWrapperPass because rebuilding the results just updates the set of known AA passes. The exception to this rule are LoopPass instances which need to preserve all the function analyses that the loop pass manager will end up needing. This means preserving both BasicAAWrapperPass and the aggregating FunctionAAResultsWrapperPass. Now, when preserving an alias analysis, you do so by directly preserving that analysis. This is only necessary for non-immutable-pass-provided alias analyses though, and there are only three of interest: BasicAA, GlobalsAA (formerly GlobalsModRef), and SCEVAA. Usually BasicAA is preserved when needed because it (like DominatorTree and LoopInfo) is marked as a CFG-only pass. I've expanded GlobalsAA into the preserved set everywhere we previously were preserving all of AliasAnalysis, and I've added SCEVAA in the intersection of that with where we preserve SCEV itself. One significant challenge to all of this is that the CGSCC passes were actually using the alias analysis implementations by taking advantage of a pretty amazing set of loop holes in the old pass manager's analysis management code which allowed analysis groups to slide through in many cases. Moving away from analysis groups makes this problem much more obvious. To fix it, I've leveraged the flexibility the design of the new PM components provides to just directly construct the relevant alias analyses for the relevant functions in the IPO passes that need them. This is a bit hacky, but should go away with the new pass manager, and is already in many ways cleaner than the prior state. Another significant challenge is that various facilities of the old alias analysis infrastructure just don't fit any more. The most significant of these is the alias analysis 'counter' pass. That pass relied on the ability to snoop on AA queries at different points in the analysis group chain. Instead, I'm planning to build printing functionality directly into the aggregation layer. I've not included that in this patch merely to keep it smaller. Note that all of this needs a nearly complete rewrite of the AA documentation. I'm planning to do that, but I'd like to make sure the new design settles, and to flesh out a bit more of what it looks like in the new pass manager first. Differential Revision: http://reviews.llvm.org/D12080 llvm-svn: 247167
2015-09-10 01:55:00 +08:00
#include "llvm/Analysis/ScalarEvolutionAliasAnalysis.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/PredIteratorCache.h"
#include "llvm/Pass.h"
#include "llvm/Transforms/Scalar.h"
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
#include "llvm/Transforms/Utils/LoopUtils.h"
#include "llvm/Transforms/Utils/SSAUpdater.h"
using namespace llvm;
#define DEBUG_TYPE "lcssa"
STATISTIC(NumLCSSA, "Number of live out of a loop variables");
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
/// Return true if the specified block is in the list.
static bool isExitBlock(BasicBlock *BB,
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
const SmallVectorImpl<BasicBlock *> &ExitBlocks) {
return find(ExitBlocks, BB) != ExitBlocks.end();
}
/// For every instruction from the worklist, check to see if it has any uses
/// that are outside the current loop. If so, insert LCSSA PHI nodes and
/// rewrite the uses.
bool llvm::formLCSSAForInstructions(SmallVectorImpl<Instruction *> &Worklist,
DominatorTree &DT, LoopInfo &LI) {
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
SmallVector<Use *, 16> UsesToRewrite;
SmallVector<BasicBlock *, 8> ExitBlocks;
SmallSetVector<PHINode *, 16> PHIsToRemove;
PredIteratorCache PredCache;
bool Changed = false;
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
while (!Worklist.empty()) {
UsesToRewrite.clear();
ExitBlocks.clear();
Instruction *I = Worklist.pop_back_val();
BasicBlock *InstBB = I->getParent();
Loop *L = LI.getLoopFor(InstBB);
L->getExitBlocks(ExitBlocks);
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
if (ExitBlocks.empty())
continue;
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
// Tokens cannot be used in PHI nodes, so we skip over them.
// We can run into tokens which are live out of a loop with catchswitch
// instructions in Windows EH if the catchswitch has one catchpad which
// is inside the loop and another which is not.
if (I->getType()->isTokenTy())
continue;
for (Use &U : I->uses()) {
Instruction *User = cast<Instruction>(U.getUser());
BasicBlock *UserBB = User->getParent();
if (PHINode *PN = dyn_cast<PHINode>(User))
UserBB = PN->getIncomingBlock(U);
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
if (InstBB != UserBB && !L->contains(UserBB))
UsesToRewrite.push_back(&U);
}
// If there are no uses outside the loop, exit with no change.
if (UsesToRewrite.empty())
continue;
++NumLCSSA; // We are applying the transformation
// Invoke instructions are special in that their result value is not
// available along their unwind edge. The code below tests to see whether
// DomBB dominates the value, so adjust DomBB to the normal destination
// block, which is effectively where the value is first usable.
BasicBlock *DomBB = InstBB;
if (InvokeInst *Inv = dyn_cast<InvokeInst>(I))
DomBB = Inv->getNormalDest();
DomTreeNode *DomNode = DT.getNode(DomBB);
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
SmallVector<PHINode *, 16> AddedPHIs;
SmallVector<PHINode *, 8> PostProcessPHIs;
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
SmallVector<PHINode *, 4> InsertedPHIs;
SSAUpdater SSAUpdate(&InsertedPHIs);
SSAUpdate.Initialize(I->getType(), I->getName());
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
// Insert the LCSSA phi's into all of the exit blocks dominated by the
// value, and add them to the Phi's map.
for (BasicBlock *ExitBB : ExitBlocks) {
if (!DT.dominates(DomNode, DT.getNode(ExitBB)))
continue;
// If we already inserted something for this BB, don't reprocess it.
if (SSAUpdate.HasValueForBlock(ExitBB))
continue;
PHINode *PN = PHINode::Create(I->getType(), PredCache.size(ExitBB),
I->getName() + ".lcssa", &ExitBB->front());
// Add inputs from inside the loop for this PHI.
for (BasicBlock *Pred : PredCache.get(ExitBB)) {
PN->addIncoming(I, Pred);
// If the exit block has a predecessor not within the loop, arrange for
// the incoming value use corresponding to that predecessor to be
// rewritten in terms of a different LCSSA PHI.
if (!L->contains(Pred))
UsesToRewrite.push_back(
&PN->getOperandUse(PN->getOperandNumForIncomingValue(
PN->getNumIncomingValues() - 1)));
}
AddedPHIs.push_back(PN);
// Remember that this phi makes the value alive in this block.
SSAUpdate.AddAvailableValue(ExitBB, PN);
// LoopSimplify might fail to simplify some loops (e.g. when indirect
// branches are involved). In such situations, it might happen that an
// exit for Loop L1 is the header of a disjoint Loop L2. Thus, when we
// create PHIs in such an exit block, we are also inserting PHIs into L2's
// header. This could break LCSSA form for L2 because these inserted PHIs
// can also have uses outside of L2. Remember all PHIs in such situation
// as to revisit than later on. FIXME: Remove this if indirectbr support
// into LoopSimplify gets improved.
if (auto *OtherLoop = LI.getLoopFor(ExitBB))
if (!L->contains(OtherLoop))
PostProcessPHIs.push_back(PN);
}
// Rewrite all uses outside the loop in terms of the new PHIs we just
// inserted.
for (Use *UseToRewrite : UsesToRewrite) {
// If this use is in an exit block, rewrite to use the newly inserted PHI.
// This is required for correctness because SSAUpdate doesn't handle uses
// in the same block. It assumes the PHI we inserted is at the end of the
// block.
Instruction *User = cast<Instruction>(UseToRewrite->getUser());
BasicBlock *UserBB = User->getParent();
if (PHINode *PN = dyn_cast<PHINode>(User))
UserBB = PN->getIncomingBlock(*UseToRewrite);
if (isa<PHINode>(UserBB->begin()) && isExitBlock(UserBB, ExitBlocks)) {
// Tell the VHs that the uses changed. This updates SCEV's caches.
if (UseToRewrite->get()->hasValueHandle())
ValueHandleBase::ValueIsRAUWd(*UseToRewrite, &UserBB->front());
UseToRewrite->set(&UserBB->front());
continue;
}
// Otherwise, do full PHI insertion.
SSAUpdate.RewriteUse(*UseToRewrite);
// SSAUpdater might have inserted phi-nodes inside other loops. We'll need
// to post-process them to keep LCSSA form.
for (PHINode *InsertedPN : InsertedPHIs) {
if (auto *OtherLoop = LI.getLoopFor(InsertedPN->getParent()))
if (!L->contains(OtherLoop))
PostProcessPHIs.push_back(InsertedPN);
}
}
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
// Post process PHI instructions that were inserted into another disjoint
// loop and update their exits properly.
for (auto *PostProcessPN : PostProcessPHIs) {
if (PostProcessPN->use_empty())
continue;
// Reprocess each PHI instruction.
Worklist.push_back(PostProcessPN);
}
// Keep track of PHI nodes that we want to remove because they did not have
// any uses rewritten.
for (PHINode *PN : AddedPHIs)
if (PN->use_empty())
PHIsToRemove.insert(PN);
Changed = true;
}
// Remove PHI nodes that did not have any uses rewritten.
for (PHINode *PN : PHIsToRemove) {
assert (PN->use_empty() && "Trying to remove a phi with uses.");
PN->eraseFromParent();
}
return Changed;
}
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
/// Return true if the specified block dominates at least
/// one of the blocks in the specified list.
static bool
blockDominatesAnExit(BasicBlock *BB,
DominatorTree &DT,
const SmallVectorImpl<BasicBlock *> &ExitBlocks) {
DomTreeNode *DomNode = DT.getNode(BB);
return llvm::any_of(ExitBlocks, [&](BasicBlock * EB) {
return DT.dominates(DomNode, DT.getNode(EB));
});
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
}
bool llvm::formLCSSA(Loop &L, DominatorTree &DT, LoopInfo *LI,
ScalarEvolution *SE) {
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
bool Changed = false;
// Get the set of exiting blocks.
SmallVector<BasicBlock *, 8> ExitBlocks;
L.getExitBlocks(ExitBlocks);
if (ExitBlocks.empty())
return false;
SmallVector<Instruction *, 8> Worklist;
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
// Look at all the instructions in the loop, checking to see if they have uses
// outside the loop. If so, put them into the worklist to rewrite those uses.
for (BasicBlock *BB : L.blocks()) {
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
// For large loops, avoid use-scanning by using dominance information: In
// particular, if a block does not dominate any of the loop exits, then none
// of the values defined in the block could be used outside the loop.
if (!blockDominatesAnExit(BB, DT, ExitBlocks))
continue;
for (Instruction &I : *BB) {
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
// Reject two common cases fast: instructions with no uses (like stores)
// and instructions with one use that is in the same block as this.
if (I.use_empty() ||
(I.hasOneUse() && I.user_back()->getParent() == BB &&
!isa<PHINode>(I.user_back())))
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
continue;
Worklist.push_back(&I);
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
}
}
Changed = formLCSSAForInstructions(Worklist, DT, *LI);
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
// If we modified the code, remove any caches about the loop from SCEV to
// avoid dangling entries.
// FIXME: This is a big hammer, can we clear the cache more selectively?
if (SE && Changed)
SE->forgetLoop(&L);
assert(L.isLCSSAForm(DT));
return Changed;
}
/// Process a loop nest depth first.
bool llvm::formLCSSARecursively(Loop &L, DominatorTree &DT, LoopInfo *LI,
ScalarEvolution *SE) {
bool Changed = false;
// Recurse depth-first through inner loops.
for (Loop *SubLoop : L.getSubLoops())
Changed |= formLCSSARecursively(*SubLoop, DT, LI, SE);
Changed |= formLCSSA(L, DT, LI, SE);
return Changed;
}
/// Process all loops in the function, inner-most out.
static bool formLCSSAOnAllLoops(LoopInfo *LI, DominatorTree &DT,
ScalarEvolution *SE) {
bool Changed = false;
for (auto &L : *LI)
Changed |= formLCSSARecursively(*L, DT, LI, SE);
return Changed;
}
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
namespace {
struct LCSSAWrapperPass : public FunctionPass {
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
static char ID; // Pass identification, replacement for typeid
LCSSAWrapperPass() : FunctionPass(ID) {
initializeLCSSAWrapperPassPass(*PassRegistry::getPassRegistry());
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
}
// Cached analysis information for the current function.
DominatorTree *DT;
LoopInfo *LI;
ScalarEvolution *SE;
bool runOnFunction(Function &F) override;
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
/// This transformation requires natural loop information & requires that
/// loop preheaders be inserted into the CFG. It maintains both of these,
/// as well as the CFG. It also requires dominator information.
void getAnalysisUsage(AnalysisUsage &AU) const override {
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
AU.setPreservesCFG();
AU.addRequired<DominatorTreeWrapperPass>();
AU.addRequired<LoopInfoWrapperPass>();
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
AU.addPreservedID(LoopSimplifyID);
[PM/AA] Rebuild LLVM's alias analysis infrastructure in a way compatible with the new pass manager, and no longer relying on analysis groups. This builds essentially a ground-up new AA infrastructure stack for LLVM. The core ideas are the same that are used throughout the new pass manager: type erased polymorphism and direct composition. The design is as follows: - FunctionAAResults is a type-erasing alias analysis results aggregation interface to walk a single query across a range of results from different alias analyses. Currently this is function-specific as we always assume that aliasing queries are *within* a function. - AAResultBase is a CRTP utility providing stub implementations of various parts of the alias analysis result concept, notably in several cases in terms of other more general parts of the interface. This can be used to implement only a narrow part of the interface rather than the entire interface. This isn't really ideal, this logic should be hoisted into FunctionAAResults as currently it will cause a significant amount of redundant work, but it faithfully models the behavior of the prior infrastructure. - All the alias analysis passes are ported to be wrapper passes for the legacy PM and new-style analysis passes for the new PM with a shared result object. In some cases (most notably CFL), this is an extremely naive approach that we should revisit when we can specialize for the new pass manager. - BasicAA has been restructured to reflect that it is much more fundamentally a function analysis because it uses dominator trees and loop info that need to be constructed for each function. All of the references to getting alias analysis results have been updated to use the new aggregation interface. All the preservation and other pass management code has been updated accordingly. The way the FunctionAAResultsWrapperPass works is to detect the available alias analyses when run, and add them to the results object. This means that we should be able to continue to respect when various passes are added to the pipeline, for example adding CFL or adding TBAA passes should just cause their results to be available and to get folded into this. The exception to this rule is BasicAA which really needs to be a function pass due to using dominator trees and loop info. As a consequence, the FunctionAAResultsWrapperPass directly depends on BasicAA and always includes it in the aggregation. This has significant implications for preserving analyses. Generally, most passes shouldn't bother preserving FunctionAAResultsWrapperPass because rebuilding the results just updates the set of known AA passes. The exception to this rule are LoopPass instances which need to preserve all the function analyses that the loop pass manager will end up needing. This means preserving both BasicAAWrapperPass and the aggregating FunctionAAResultsWrapperPass. Now, when preserving an alias analysis, you do so by directly preserving that analysis. This is only necessary for non-immutable-pass-provided alias analyses though, and there are only three of interest: BasicAA, GlobalsAA (formerly GlobalsModRef), and SCEVAA. Usually BasicAA is preserved when needed because it (like DominatorTree and LoopInfo) is marked as a CFG-only pass. I've expanded GlobalsAA into the preserved set everywhere we previously were preserving all of AliasAnalysis, and I've added SCEVAA in the intersection of that with where we preserve SCEV itself. One significant challenge to all of this is that the CGSCC passes were actually using the alias analysis implementations by taking advantage of a pretty amazing set of loop holes in the old pass manager's analysis management code which allowed analysis groups to slide through in many cases. Moving away from analysis groups makes this problem much more obvious. To fix it, I've leveraged the flexibility the design of the new PM components provides to just directly construct the relevant alias analyses for the relevant functions in the IPO passes that need them. This is a bit hacky, but should go away with the new pass manager, and is already in many ways cleaner than the prior state. Another significant challenge is that various facilities of the old alias analysis infrastructure just don't fit any more. The most significant of these is the alias analysis 'counter' pass. That pass relied on the ability to snoop on AA queries at different points in the analysis group chain. Instead, I'm planning to build printing functionality directly into the aggregation layer. I've not included that in this patch merely to keep it smaller. Note that all of this needs a nearly complete rewrite of the AA documentation. I'm planning to do that, but I'd like to make sure the new design settles, and to flesh out a bit more of what it looks like in the new pass manager first. Differential Revision: http://reviews.llvm.org/D12080 llvm-svn: 247167
2015-09-10 01:55:00 +08:00
AU.addPreserved<AAResultsWrapperPass>();
AU.addPreserved<BasicAAWrapperPass>();
[PM/AA] Rebuild LLVM's alias analysis infrastructure in a way compatible with the new pass manager, and no longer relying on analysis groups. This builds essentially a ground-up new AA infrastructure stack for LLVM. The core ideas are the same that are used throughout the new pass manager: type erased polymorphism and direct composition. The design is as follows: - FunctionAAResults is a type-erasing alias analysis results aggregation interface to walk a single query across a range of results from different alias analyses. Currently this is function-specific as we always assume that aliasing queries are *within* a function. - AAResultBase is a CRTP utility providing stub implementations of various parts of the alias analysis result concept, notably in several cases in terms of other more general parts of the interface. This can be used to implement only a narrow part of the interface rather than the entire interface. This isn't really ideal, this logic should be hoisted into FunctionAAResults as currently it will cause a significant amount of redundant work, but it faithfully models the behavior of the prior infrastructure. - All the alias analysis passes are ported to be wrapper passes for the legacy PM and new-style analysis passes for the new PM with a shared result object. In some cases (most notably CFL), this is an extremely naive approach that we should revisit when we can specialize for the new pass manager. - BasicAA has been restructured to reflect that it is much more fundamentally a function analysis because it uses dominator trees and loop info that need to be constructed for each function. All of the references to getting alias analysis results have been updated to use the new aggregation interface. All the preservation and other pass management code has been updated accordingly. The way the FunctionAAResultsWrapperPass works is to detect the available alias analyses when run, and add them to the results object. This means that we should be able to continue to respect when various passes are added to the pipeline, for example adding CFL or adding TBAA passes should just cause their results to be available and to get folded into this. The exception to this rule is BasicAA which really needs to be a function pass due to using dominator trees and loop info. As a consequence, the FunctionAAResultsWrapperPass directly depends on BasicAA and always includes it in the aggregation. This has significant implications for preserving analyses. Generally, most passes shouldn't bother preserving FunctionAAResultsWrapperPass because rebuilding the results just updates the set of known AA passes. The exception to this rule are LoopPass instances which need to preserve all the function analyses that the loop pass manager will end up needing. This means preserving both BasicAAWrapperPass and the aggregating FunctionAAResultsWrapperPass. Now, when preserving an alias analysis, you do so by directly preserving that analysis. This is only necessary for non-immutable-pass-provided alias analyses though, and there are only three of interest: BasicAA, GlobalsAA (formerly GlobalsModRef), and SCEVAA. Usually BasicAA is preserved when needed because it (like DominatorTree and LoopInfo) is marked as a CFG-only pass. I've expanded GlobalsAA into the preserved set everywhere we previously were preserving all of AliasAnalysis, and I've added SCEVAA in the intersection of that with where we preserve SCEV itself. One significant challenge to all of this is that the CGSCC passes were actually using the alias analysis implementations by taking advantage of a pretty amazing set of loop holes in the old pass manager's analysis management code which allowed analysis groups to slide through in many cases. Moving away from analysis groups makes this problem much more obvious. To fix it, I've leveraged the flexibility the design of the new PM components provides to just directly construct the relevant alias analyses for the relevant functions in the IPO passes that need them. This is a bit hacky, but should go away with the new pass manager, and is already in many ways cleaner than the prior state. Another significant challenge is that various facilities of the old alias analysis infrastructure just don't fit any more. The most significant of these is the alias analysis 'counter' pass. That pass relied on the ability to snoop on AA queries at different points in the analysis group chain. Instead, I'm planning to build printing functionality directly into the aggregation layer. I've not included that in this patch merely to keep it smaller. Note that all of this needs a nearly complete rewrite of the AA documentation. I'm planning to do that, but I'd like to make sure the new design settles, and to flesh out a bit more of what it looks like in the new pass manager first. Differential Revision: http://reviews.llvm.org/D12080 llvm-svn: 247167
2015-09-10 01:55:00 +08:00
AU.addPreserved<GlobalsAAWrapperPass>();
[PM] Port ScalarEvolution to the new pass manager. This change makes ScalarEvolution a stand-alone object and just produces one from a pass as needed. Making this work well requires making the object movable, using references instead of overwritten pointers in a number of places, and other refactorings. I've also wired it up to the new pass manager and added a RUN line to a test to exercise it under the new pass manager. This includes basic printing support much like with other analyses. But there is a big and somewhat scary change here. Prior to this patch ScalarEvolution was never *actually* invalidated!!! Re-running the pass just re-wired up the various other analyses and didn't remove any of the existing entries in the SCEV caches or clear out anything at all. This might seem OK as everything in SCEV that can uses ValueHandles to track updates to the values that serve as SCEV keys. However, this still means that as we ran SCEV over each function in the module, we kept accumulating more and more SCEVs into the cache. At the end, we would have a SCEV cache with every value that we ever needed a SCEV for in the entire module!!! Yowzers. The releaseMemory routine would dump all of this, but that isn't realy called during normal runs of the pipeline as far as I can see. To make matters worse, there *is* actually a key that we don't update with value handles -- there is a map keyed off of Loop*s. Because LoopInfo *does* release its memory from run to run, it is entirely possible to run SCEV over one function, then over another function, and then lookup a Loop* from the second function but find an entry inserted for the first function! Ouch. To make matters still worse, there are plenty of updates that *don't* trip a value handle. It seems incredibly unlikely that today GVN or another pass that invalidates SCEV can update values in *just* such a way that a subsequent run of SCEV will incorrectly find lookups in a cache, but it is theoretically possible and would be a nightmare to debug. With this refactoring, I've fixed all this by actually destroying and recreating the ScalarEvolution object from run to run. Technically, this could increase the amount of malloc traffic we see, but then again it is also technically correct. ;] I don't actually think we're suffering from tons of malloc traffic from SCEV because if we were, the fact that we never clear the memory would seem more likely to have come up as an actual problem before now. So, I've made the simple fix here. If in fact there are serious issues with too much allocation and deallocation, I can work on a clever fix that preserves the allocations (while clearing the data) between each run, but I'd prefer to do that kind of optimization with a test case / benchmark that shows why we need such cleverness (and that can test that we actually make it faster). It's possible that this will make some things faster by making the SCEV caches have higher locality (due to being significantly smaller) so until there is a clear benchmark, I think the simple change is best. Differential Revision: http://reviews.llvm.org/D12063 llvm-svn: 245193
2015-08-17 10:08:17 +08:00
AU.addPreserved<ScalarEvolutionWrapperPass>();
[PM/AA] Rebuild LLVM's alias analysis infrastructure in a way compatible with the new pass manager, and no longer relying on analysis groups. This builds essentially a ground-up new AA infrastructure stack for LLVM. The core ideas are the same that are used throughout the new pass manager: type erased polymorphism and direct composition. The design is as follows: - FunctionAAResults is a type-erasing alias analysis results aggregation interface to walk a single query across a range of results from different alias analyses. Currently this is function-specific as we always assume that aliasing queries are *within* a function. - AAResultBase is a CRTP utility providing stub implementations of various parts of the alias analysis result concept, notably in several cases in terms of other more general parts of the interface. This can be used to implement only a narrow part of the interface rather than the entire interface. This isn't really ideal, this logic should be hoisted into FunctionAAResults as currently it will cause a significant amount of redundant work, but it faithfully models the behavior of the prior infrastructure. - All the alias analysis passes are ported to be wrapper passes for the legacy PM and new-style analysis passes for the new PM with a shared result object. In some cases (most notably CFL), this is an extremely naive approach that we should revisit when we can specialize for the new pass manager. - BasicAA has been restructured to reflect that it is much more fundamentally a function analysis because it uses dominator trees and loop info that need to be constructed for each function. All of the references to getting alias analysis results have been updated to use the new aggregation interface. All the preservation and other pass management code has been updated accordingly. The way the FunctionAAResultsWrapperPass works is to detect the available alias analyses when run, and add them to the results object. This means that we should be able to continue to respect when various passes are added to the pipeline, for example adding CFL or adding TBAA passes should just cause their results to be available and to get folded into this. The exception to this rule is BasicAA which really needs to be a function pass due to using dominator trees and loop info. As a consequence, the FunctionAAResultsWrapperPass directly depends on BasicAA and always includes it in the aggregation. This has significant implications for preserving analyses. Generally, most passes shouldn't bother preserving FunctionAAResultsWrapperPass because rebuilding the results just updates the set of known AA passes. The exception to this rule are LoopPass instances which need to preserve all the function analyses that the loop pass manager will end up needing. This means preserving both BasicAAWrapperPass and the aggregating FunctionAAResultsWrapperPass. Now, when preserving an alias analysis, you do so by directly preserving that analysis. This is only necessary for non-immutable-pass-provided alias analyses though, and there are only three of interest: BasicAA, GlobalsAA (formerly GlobalsModRef), and SCEVAA. Usually BasicAA is preserved when needed because it (like DominatorTree and LoopInfo) is marked as a CFG-only pass. I've expanded GlobalsAA into the preserved set everywhere we previously were preserving all of AliasAnalysis, and I've added SCEVAA in the intersection of that with where we preserve SCEV itself. One significant challenge to all of this is that the CGSCC passes were actually using the alias analysis implementations by taking advantage of a pretty amazing set of loop holes in the old pass manager's analysis management code which allowed analysis groups to slide through in many cases. Moving away from analysis groups makes this problem much more obvious. To fix it, I've leveraged the flexibility the design of the new PM components provides to just directly construct the relevant alias analyses for the relevant functions in the IPO passes that need them. This is a bit hacky, but should go away with the new pass manager, and is already in many ways cleaner than the prior state. Another significant challenge is that various facilities of the old alias analysis infrastructure just don't fit any more. The most significant of these is the alias analysis 'counter' pass. That pass relied on the ability to snoop on AA queries at different points in the analysis group chain. Instead, I'm planning to build printing functionality directly into the aggregation layer. I've not included that in this patch merely to keep it smaller. Note that all of this needs a nearly complete rewrite of the AA documentation. I'm planning to do that, but I'd like to make sure the new design settles, and to flesh out a bit more of what it looks like in the new pass manager first. Differential Revision: http://reviews.llvm.org/D12080 llvm-svn: 247167
2015-09-10 01:55:00 +08:00
AU.addPreserved<SCEVAAWrapperPass>();
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
}
};
}
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
char LCSSAWrapperPass::ID = 0;
INITIALIZE_PASS_BEGIN(LCSSAWrapperPass, "lcssa", "Loop-Closed SSA Form Pass",
false, false)
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
INITIALIZE_PASS_DEPENDENCY(DominatorTreeWrapperPass)
INITIALIZE_PASS_DEPENDENCY(LoopInfoWrapperPass)
INITIALIZE_PASS_END(LCSSAWrapperPass, "lcssa", "Loop-Closed SSA Form Pass",
false, false)
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
Pass *llvm::createLCSSAPass() { return new LCSSAWrapperPass(); }
char &llvm::LCSSAID = LCSSAWrapperPass::ID;
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
/// Transform \p F into loop-closed SSA form.
bool LCSSAWrapperPass::runOnFunction(Function &F) {
LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
DT = &getAnalysis<DominatorTreeWrapperPass>().getDomTree();
[PM] Port ScalarEvolution to the new pass manager. This change makes ScalarEvolution a stand-alone object and just produces one from a pass as needed. Making this work well requires making the object movable, using references instead of overwritten pointers in a number of places, and other refactorings. I've also wired it up to the new pass manager and added a RUN line to a test to exercise it under the new pass manager. This includes basic printing support much like with other analyses. But there is a big and somewhat scary change here. Prior to this patch ScalarEvolution was never *actually* invalidated!!! Re-running the pass just re-wired up the various other analyses and didn't remove any of the existing entries in the SCEV caches or clear out anything at all. This might seem OK as everything in SCEV that can uses ValueHandles to track updates to the values that serve as SCEV keys. However, this still means that as we ran SCEV over each function in the module, we kept accumulating more and more SCEVs into the cache. At the end, we would have a SCEV cache with every value that we ever needed a SCEV for in the entire module!!! Yowzers. The releaseMemory routine would dump all of this, but that isn't realy called during normal runs of the pipeline as far as I can see. To make matters worse, there *is* actually a key that we don't update with value handles -- there is a map keyed off of Loop*s. Because LoopInfo *does* release its memory from run to run, it is entirely possible to run SCEV over one function, then over another function, and then lookup a Loop* from the second function but find an entry inserted for the first function! Ouch. To make matters still worse, there are plenty of updates that *don't* trip a value handle. It seems incredibly unlikely that today GVN or another pass that invalidates SCEV can update values in *just* such a way that a subsequent run of SCEV will incorrectly find lookups in a cache, but it is theoretically possible and would be a nightmare to debug. With this refactoring, I've fixed all this by actually destroying and recreating the ScalarEvolution object from run to run. Technically, this could increase the amount of malloc traffic we see, but then again it is also technically correct. ;] I don't actually think we're suffering from tons of malloc traffic from SCEV because if we were, the fact that we never clear the memory would seem more likely to have come up as an actual problem before now. So, I've made the simple fix here. If in fact there are serious issues with too much allocation and deallocation, I can work on a clever fix that preserves the allocations (while clearing the data) between each run, but I'd prefer to do that kind of optimization with a test case / benchmark that shows why we need such cleverness (and that can test that we actually make it faster). It's possible that this will make some things faster by making the SCEV caches have higher locality (due to being significantly smaller) so until there is a clear benchmark, I think the simple change is best. Differential Revision: http://reviews.llvm.org/D12063 llvm-svn: 245193
2015-08-17 10:08:17 +08:00
auto *SEWP = getAnalysisIfAvailable<ScalarEvolutionWrapperPass>();
SE = SEWP ? &SEWP->getSE() : nullptr;
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
return formLCSSAOnAllLoops(LI, *DT, SE);
[LPM] Make LCSSA a utility with a FunctionPass that applies it to all the loops in a function, and teach LICM to work in the presance of LCSSA. Previously, LCSSA was a loop pass. That made passes requiring it also be loop passes and unable to depend on function analysis passes easily. It also caused outer loops to have a different "canonical" form from inner loops during analysis. Instead, we go into LCSSA form and preserve it through the loop pass manager run. Note that this has the same problem as LoopSimplify that prevents enabling its verification -- loop passes which run at the end of the loop pass manager and don't preserve these are valid, but the subsequent loop pass runs of outer loops that do preserve this pass trigger too much verification and fail because the inner loop no longer verifies. The other problem this exposed is that LICM was completely unable to handle LCSSA form. It didn't preserve it and it actually would give up on moving instructions in many cases when they were used by an LCSSA phi node. I've taught LICM to support detecting LCSSA-form PHI nodes and to hoist and sink around them. This may actually let LICM fire significantly more because we put everything into LCSSA form to rotate the loop before running LICM. =/ Now LICM should handle that fine and preserve it correctly. The down side is that LICM has to require LCSSA in order to preserve it. This is just a fact of life for LCSSA. It's entirely possible we should completely remove LCSSA from the optimizer. The test updates are essentially accomodating LCSSA phi nodes in the output of LICM, and the fact that we now completely sink every instruction in ashr-crash below the loop bodies prior to unrolling. With this change, LCSSA is computed only three times in the pass pipeline. One of them could be removed (and potentially a SCEV run and a separate LoopPassManager entirely!) if we had a LoopPass variant of InstCombine that ran InstCombine on the loop body but refused to combine away LCSSA PHI nodes. Currently, this also prevents loop unrolling from being in the same loop pass manager is rotate, LICM, and unswitch. There is one thing that I *really* don't like -- preserving LCSSA in LICM is quite expensive. We end up having to re-run LCSSA twice for some loops after LICM runs because LICM can undo LCSSA both in the current loop and the parent loop. I don't really see good solutions to this other than to completely move away from LCSSA and using tools like SSAUpdater instead. llvm-svn: 200067
2014-01-25 12:07:24 +08:00
}
PreservedAnalyses LCSSAPass::run(Function &F, AnalysisManager<Function> &AM) {
auto &LI = AM.getResult<LoopAnalysis>(F);
auto &DT = AM.getResult<DominatorTreeAnalysis>(F);
auto *SE = AM.getCachedResult<ScalarEvolutionAnalysis>(F);
if (!formLCSSAOnAllLoops(&LI, DT, SE))
return PreservedAnalyses::all();
// FIXME: This should also 'preserve the CFG'.
PreservedAnalyses PA;
PA.preserve<BasicAA>();
PA.preserve<GlobalsAA>();
PA.preserve<SCEVAA>();
PA.preserve<ScalarEvolutionAnalysis>();
return PA;
}