1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-26 04:32:44 +01:00
llvm-mirror/lib/CodeGen/PostRASchedulerList.cpp

702 lines
24 KiB
C++
Raw Normal View History

//===----- SchedulePostRAList.cpp - list scheduler ------------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This implements a top-down list scheduler, using standard algorithms.
// The basic approach uses a priority queue of available nodes to schedule.
// One at a time, nodes are taken from the priority queue (thus in priority
// order), checked for legality to schedule, and emitted if legal.
//
// Nodes may not be legal to schedule either due to structural hazards (e.g.
// pipeline or resource constraints) or because an input to the instruction has
// not completed execution.
//
//===----------------------------------------------------------------------===//
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/CodeGen/AntiDepBreaker.h"
#include "llvm/CodeGen/LatencyPriorityQueue.h"
#include "llvm/CodeGen/MachineDominators.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/CodeGen/MachineLoopInfo.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/Passes.h"
#include "llvm/CodeGen/RegisterClassInfo.h"
#include "llvm/CodeGen/ScheduleDAGInstrs.h"
#include "llvm/CodeGen/ScheduleHazardRecognizer.h"
#include "llvm/CodeGen/SchedulerRegistry.h"
#include "llvm/CodeGen/TargetInstrInfo.h"
#include "llvm/CodeGen/TargetLowering.h"
#include "llvm/CodeGen/TargetPassConfig.h"
#include "llvm/CodeGen/TargetRegisterInfo.h"
#include "llvm/CodeGen/TargetSubtargetInfo.h"
#include "llvm/Config/llvm-config.h"
#include "llvm/InitializePasses.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
2009-08-11 03:44:26 +02:00
#include "llvm/Support/raw_ostream.h"
using namespace llvm;
#define DEBUG_TYPE "post-RA-sched"
STATISTIC(NumNoops, "Number of noops inserted");
STATISTIC(NumStalls, "Number of pipeline stalls");
STATISTIC(NumFixedAnti, "Number of fixed anti-dependencies");
// Post-RA scheduling is enabled with
// TargetSubtargetInfo.enablePostRAScheduler(). This flag can be used to
// override the target.
static cl::opt<bool>
EnablePostRAScheduler("post-RA-scheduler",
cl::desc("Enable scheduling after register allocation"),
cl::init(false), cl::Hidden);
static cl::opt<std::string>
EnableAntiDepBreaking("break-anti-dependencies",
cl::desc("Break post-RA scheduling anti-dependencies: "
"\"critical\", \"all\", or \"none\""),
cl::init("none"), cl::Hidden);
// If DebugDiv > 0 then only schedule MBB with (ID % DebugDiv) == DebugMod
static cl::opt<int>
DebugDiv("postra-sched-debugdiv",
cl::desc("Debug control MBBs that are scheduled"),
cl::init(0), cl::Hidden);
static cl::opt<int>
DebugMod("postra-sched-debugmod",
cl::desc("Debug control MBBs that are scheduled"),
cl::init(0), cl::Hidden);
AntiDepBreaker::~AntiDepBreaker() { }
namespace {
class PostRAScheduler : public MachineFunctionPass {
const TargetInstrInfo *TII = nullptr;
RegisterClassInfo RegClassInfo;
public:
static char ID;
PostRAScheduler() : MachineFunctionPass(ID) {}
void getAnalysisUsage(AnalysisUsage &AU) const override {
AU.setPreservesCFG();
[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
2015-09-09 19:55:00 +02:00
AU.addRequired<AAResultsWrapperPass>();
AU.addRequired<TargetPassConfig>();
AU.addRequired<MachineDominatorTree>();
AU.addPreserved<MachineDominatorTree>();
AU.addRequired<MachineLoopInfo>();
AU.addPreserved<MachineLoopInfo>();
MachineFunctionPass::getAnalysisUsage(AU);
}
MachineFunctionProperties getRequiredProperties() const override {
return MachineFunctionProperties().set(
MachineFunctionProperties::Property::NoVRegs);
}
bool runOnMachineFunction(MachineFunction &Fn) override;
2014-10-29 16:23:11 +01:00
private:
bool enablePostRAScheduler(
const TargetSubtargetInfo &ST, CodeGenOpt::Level OptLevel,
TargetSubtargetInfo::AntiDepBreakMode &Mode,
TargetSubtargetInfo::RegClassVector &CriticalPathRCs) const;
};
char PostRAScheduler::ID = 0;
class SchedulePostRATDList : public ScheduleDAGInstrs {
/// AvailableQueue - The priority queue to use for the available SUnits.
///
LatencyPriorityQueue AvailableQueue;
/// PendingQueue - This contains all of the instructions whose operands have
/// been issued, but their results are not ready yet (due to the latency of
/// the operation). Once the operands becomes available, the instruction is
/// added to the AvailableQueue.
std::vector<SUnit*> PendingQueue;
/// HazardRec - The hazard recognizer to use.
ScheduleHazardRecognizer *HazardRec;
/// AntiDepBreak - Anti-dependence breaking object, or NULL if none
AntiDepBreaker *AntiDepBreak;
/// AA - AliasAnalysis for making memory reference queries.
AliasAnalysis *AA;
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
/// The schedule. Null SUnit*'s represent noop instructions.
std::vector<SUnit*> Sequence;
/// Ordered list of DAG postprocessing steps.
std::vector<std::unique_ptr<ScheduleDAGMutation>> Mutations;
/// The index in BB of RegionEnd.
///
/// This is the instruction number from the top of the current block, not
/// the SlotIndex. It is only used by the AntiDepBreaker.
unsigned EndIndex;
public:
SchedulePostRATDList(
MachineFunction &MF, MachineLoopInfo &MLI, AliasAnalysis *AA,
const RegisterClassInfo &,
TargetSubtargetInfo::AntiDepBreakMode AntiDepMode,
SmallVectorImpl<const TargetRegisterClass *> &CriticalPathRCs);
~SchedulePostRATDList() override;
/// startBlock - Initialize register live-range state for scheduling in
/// this block.
///
void startBlock(MachineBasicBlock *BB) override;
// Set the index of RegionEnd within the current BB.
void setEndIndex(unsigned EndIdx) { EndIndex = EndIdx; }
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
/// Initialize the scheduler state for the next scheduling region.
void enterRegion(MachineBasicBlock *bb,
MachineBasicBlock::iterator begin,
MachineBasicBlock::iterator end,
unsigned regioninstrs) override;
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
/// Notify that the scheduler has finished scheduling the current region.
void exitRegion() override;
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
/// Schedule - Schedule the instruction range using list scheduling.
///
void schedule() override;
void EmitSchedule();
/// Observe - Update liveness information to account for the current
/// instruction, which will not be scheduled.
///
void Observe(MachineInstr &MI, unsigned Count);
/// finishBlock - Clean up register live-range state.
///
void finishBlock() override;
private:
/// Apply each ScheduleDAGMutation step in order.
void postprocessDAG();
void ReleaseSucc(SUnit *SU, SDep *SuccEdge);
void ReleaseSuccessors(SUnit *SU);
void ScheduleNodeTopDown(SUnit *SU, unsigned CurCycle);
void ListScheduleTopDown();
void dumpSchedule() const;
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
void emitNoop(unsigned CurCycle);
};
}
char &llvm::PostRASchedulerID = PostRAScheduler::ID;
INITIALIZE_PASS(PostRAScheduler, DEBUG_TYPE,
"Post RA top-down list latency scheduler", false, false)
SchedulePostRATDList::SchedulePostRATDList(
MachineFunction &MF, MachineLoopInfo &MLI, AliasAnalysis *AA,
const RegisterClassInfo &RCI,
TargetSubtargetInfo::AntiDepBreakMode AntiDepMode,
SmallVectorImpl<const TargetRegisterClass *> &CriticalPathRCs)
: ScheduleDAGInstrs(MF, &MLI), AA(AA), EndIndex(0) {
const InstrItineraryData *InstrItins =
MF.getSubtarget().getInstrItineraryData();
HazardRec =
MF.getSubtarget().getInstrInfo()->CreateTargetPostRAHazardRecognizer(
InstrItins, this);
MF.getSubtarget().getPostRAMutations(Mutations);
This patch fixes a problem which arose when using the Post-RA scheduler on X86 Atom. Some of our tests failed because the tail merging part of the BranchFolding pass was creating new basic blocks which did not contain live-in information. When the anti-dependency code in the Post-RA scheduler ran, it would sometimes rename the register containing the function return value because the fact that the return value was live-in to the subsequent block had been lost. To fix this, it is necessary to run the RegisterScavenging code in the BranchFolding pass. This patch makes sure that the register scavenging code is invoked in the X86 subtarget only when post-RA scheduling is being done. Post RA scheduling in the X86 subtarget is only done for Atom. This patch adds a new function to the TargetRegisterClass to control whether or not live-ins should be preserved during branch folding. This is necessary in order for the anti-dependency optimizations done during the PostRASchedulerList pass to work properly when doing Post-RA scheduling for the X86 in general and for the Intel Atom in particular. The patch adds and invokes the new function trackLivenessAfterRegAlloc() instead of using the existing requiresRegisterScavenging(). It changes BranchFolding.cpp to call trackLivenessAfterRegAlloc() instead of requiresRegisterScavenging(). It changes the all the targets that implemented requiresRegisterScavenging() to also implement trackLivenessAfterRegAlloc(). It adds an assertion in the Post RA scheduler to make sure that post RA liveness information is available when it is needed. It changes the X86 break-anti-dependencies test to use –mcpu=atom, in order to avoid running into the added assertion. Finally, this patch restores the use of anti-dependency checking (which was turned off temporarily for the 3.1 release) for Intel Atom in the Post RA scheduler. Patch by Andy Zhang! Thanks to Jakob and Anton for their reviews. llvm-svn: 155395
2012-04-23 23:39:35 +02:00
assert((AntiDepMode == TargetSubtargetInfo::ANTIDEP_NONE ||
MRI.tracksLiveness()) &&
"Live-ins must be accurate for anti-dependency breaking");
AntiDepBreak = ((AntiDepMode == TargetSubtargetInfo::ANTIDEP_ALL)
? createAggressiveAntiDepBreaker(MF, RCI, CriticalPathRCs)
: ((AntiDepMode == TargetSubtargetInfo::ANTIDEP_CRITICAL)
? createCriticalAntiDepBreaker(MF, RCI)
: nullptr));
}
SchedulePostRATDList::~SchedulePostRATDList() {
delete HazardRec;
delete AntiDepBreak;
}
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
/// Initialize state associated with the next scheduling region.
void SchedulePostRATDList::enterRegion(MachineBasicBlock *bb,
MachineBasicBlock::iterator begin,
MachineBasicBlock::iterator end,
unsigned regioninstrs) {
ScheduleDAGInstrs::enterRegion(bb, begin, end, regioninstrs);
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
Sequence.clear();
}
/// Print the schedule before exiting the region.
void SchedulePostRATDList::exitRegion() {
LLVM_DEBUG({
dbgs() << "*** Final schedule ***\n";
dumpSchedule();
dbgs() << '\n';
});
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
ScheduleDAGInstrs::exitRegion();
}
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
/// dumpSchedule - dump the scheduled Sequence.
LLVM_DUMP_METHOD void SchedulePostRATDList::dumpSchedule() const {
for (unsigned i = 0, e = Sequence.size(); i != e; i++) {
if (SUnit *SU = Sequence[i])
dumpNode(*SU);
else
dbgs() << "**** NOOP ****\n";
}
}
#endif
bool PostRAScheduler::enablePostRAScheduler(
const TargetSubtargetInfo &ST,
CodeGenOpt::Level OptLevel,
TargetSubtargetInfo::AntiDepBreakMode &Mode,
TargetSubtargetInfo::RegClassVector &CriticalPathRCs) const {
Mode = ST.getAntiDepBreakMode();
ST.getCriticalPathRCs(CriticalPathRCs);
// Check for explicit enable/disable of post-ra scheduling.
if (EnablePostRAScheduler.getPosition() > 0)
return EnablePostRAScheduler;
return ST.enablePostRAScheduler() &&
OptLevel >= ST.getOptLevelToEnablePostRAScheduler();
}
bool PostRAScheduler::runOnMachineFunction(MachineFunction &Fn) {
if (skipFunction(Fn.getFunction()))
return false;
TII = Fn.getSubtarget().getInstrInfo();
MachineLoopInfo &MLI = getAnalysis<MachineLoopInfo>();
[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
2015-09-09 19:55:00 +02:00
AliasAnalysis *AA = &getAnalysis<AAResultsWrapperPass>().getAAResults();
TargetPassConfig *PassConfig = &getAnalysis<TargetPassConfig>();
RegClassInfo.runOnMachineFunction(Fn);
TargetSubtargetInfo::AntiDepBreakMode AntiDepMode =
TargetSubtargetInfo::ANTIDEP_NONE;
SmallVector<const TargetRegisterClass*, 4> CriticalPathRCs;
// Check that post-RA scheduling is enabled for this target.
// This may upgrade the AntiDepMode.
if (!enablePostRAScheduler(Fn.getSubtarget(), PassConfig->getOptLevel(),
AntiDepMode, CriticalPathRCs))
return false;
// Check for antidep breaking override...
if (EnableAntiDepBreaking.getPosition() > 0) {
AntiDepMode = (EnableAntiDepBreaking == "all")
? TargetSubtargetInfo::ANTIDEP_ALL
: ((EnableAntiDepBreaking == "critical")
? TargetSubtargetInfo::ANTIDEP_CRITICAL
: TargetSubtargetInfo::ANTIDEP_NONE);
}
LLVM_DEBUG(dbgs() << "PostRAScheduler\n");
SchedulePostRATDList Scheduler(Fn, MLI, AA, RegClassInfo, AntiDepMode,
CriticalPathRCs);
// Loop over all of the basic blocks
for (auto &MBB : Fn) {
#ifndef NDEBUG
// If DebugDiv > 0 then only schedule MBB with (ID % DebugDiv) == DebugMod
if (DebugDiv > 0) {
static int bbcnt = 0;
if (bbcnt++ % DebugDiv != DebugMod)
continue;
dbgs() << "*** DEBUG scheduling " << Fn.getName() << ":"
<< printMBBReference(MBB) << " ***\n";
}
#endif
// Initialize register live-range state for scheduling in this block.
Scheduler.startBlock(&MBB);
// Schedule each sequence of instructions not interrupted by a label
// or anything else that effectively needs to shut down scheduling.
MachineBasicBlock::iterator Current = MBB.end();
unsigned Count = MBB.size(), CurrentCount = Count;
for (MachineBasicBlock::iterator I = Current; I != MBB.begin();) {
MachineInstr &MI = *std::prev(I);
--Count;
// Calls are not scheduling boundaries before register allocation, but
// post-ra we don't gain anything by scheduling across calls since we
// don't need to worry about register pressure.
if (MI.isCall() || TII->isSchedulingBoundary(MI, &MBB, Fn)) {
Scheduler.enterRegion(&MBB, I, Current, CurrentCount - Count);
Scheduler.setEndIndex(CurrentCount);
Scheduler.schedule();
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
Scheduler.exitRegion();
Scheduler.EmitSchedule();
Current = &MI;
CurrentCount = Count;
Scheduler.Observe(MI, CurrentCount);
}
I = MI;
if (MI.isBundle())
Count -= MI.getBundleSize();
}
assert(Count == 0 && "Instruction count mismatch!");
assert((MBB.begin() == Current || CurrentCount != 0) &&
"Instruction count mismatch!");
Scheduler.enterRegion(&MBB, MBB.begin(), Current, CurrentCount);
Scheduler.setEndIndex(CurrentCount);
Scheduler.schedule();
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
Scheduler.exitRegion();
Scheduler.EmitSchedule();
// Clean up register live-range state.
Scheduler.finishBlock();
// Update register kills
Scheduler.fixupKills(MBB);
}
return true;
}
/// StartBlock - Initialize register live-range state for scheduling in
/// this block.
///
void SchedulePostRATDList::startBlock(MachineBasicBlock *BB) {
// Call the superclass.
ScheduleDAGInstrs::startBlock(BB);
// Reset the hazard recognizer and anti-dep breaker.
HazardRec->Reset();
if (AntiDepBreak)
AntiDepBreak->StartBlock(BB);
}
/// Schedule - Schedule the instruction range using list scheduling.
///
void SchedulePostRATDList::schedule() {
// Build the scheduling graph.
buildSchedGraph(AA);
if (AntiDepBreak) {
unsigned Broken =
AntiDepBreak->BreakAntiDependencies(SUnits, RegionBegin, RegionEnd,
EndIndex, DbgValues);
if (Broken != 0) {
// We made changes. Update the dependency graph.
// Theoretically we could update the graph in place:
// When a live range is changed to use a different register, remove
// the def's anti-dependence *and* output-dependence edges due to
// that register, and add new anti-dependence and output-dependence
// edges based on the next live range of the register.
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles. ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation. ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class. ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target. Specific changes: - Removed driver code from ScheduleDAG. clearDAG is the only interface needed. - Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls. - Added ScheduleDAGInstrs::begin()/end() public API. - Moved Sequence into the driver layer, which is specific to the scheduling algorithm. llvm-svn: 152208
2012-03-07 06:21:52 +01:00
ScheduleDAG::clearDAG();
buildSchedGraph(AA);
NumFixedAnti += Broken;
}
}
postprocessDAG();
LLVM_DEBUG(dbgs() << "********** List Scheduling **********\n");
LLVM_DEBUG(dump());
AvailableQueue.initNodes(SUnits);
ListScheduleTopDown();
AvailableQueue.releaseState();
}
/// Observe - Update liveness information to account for the current
/// instruction, which will not be scheduled.
///
void SchedulePostRATDList::Observe(MachineInstr &MI, unsigned Count) {
if (AntiDepBreak)
AntiDepBreak->Observe(MI, Count, EndIndex);
}
/// FinishBlock - Clean up register live-range state.
///
void SchedulePostRATDList::finishBlock() {
if (AntiDepBreak)
AntiDepBreak->FinishBlock();
// Call the superclass.
ScheduleDAGInstrs::finishBlock();
}
/// Apply each ScheduleDAGMutation step in order.
void SchedulePostRATDList::postprocessDAG() {
for (auto &M : Mutations)
M->apply(this);
}
//===----------------------------------------------------------------------===//
// Top-Down Scheduling
//===----------------------------------------------------------------------===//
/// ReleaseSucc - Decrement the NumPredsLeft count of a successor. Add it to
/// the PendingQueue if the count reaches zero.
void SchedulePostRATDList::ReleaseSucc(SUnit *SU, SDep *SuccEdge) {
SUnit *SuccSU = SuccEdge->getSUnit();
if (SuccEdge->isWeak()) {
--SuccSU->WeakPredsLeft;
return;
}
#ifndef NDEBUG
if (SuccSU->NumPredsLeft == 0) {
dbgs() << "*** Scheduling failed! ***\n";
dumpNode(*SuccSU);
dbgs() << " has been released too many times!\n";
llvm_unreachable(nullptr);
}
#endif
--SuccSU->NumPredsLeft;
// Standard scheduler algorithms will recompute the depth of the successor
// here as such:
// SuccSU->setDepthToAtLeast(SU->getDepth() + SuccEdge->getLatency());
//
// However, we lazily compute node depth instead. Note that
// ScheduleNodeTopDown has already updated the depth of this node which causes
// all descendents to be marked dirty. Setting the successor depth explicitly
// here would cause depth to be recomputed for all its ancestors. If the
// successor is not yet ready (because of a transitively redundant edge) then
// this causes depth computation to be quadratic in the size of the DAG.
// If all the node's predecessors are scheduled, this node is ready
// to be scheduled. Ignore the special ExitSU node.
if (SuccSU->NumPredsLeft == 0 && SuccSU != &ExitSU)
PendingQueue.push_back(SuccSU);
}
/// ReleaseSuccessors - Call ReleaseSucc on each of SU's successors.
void SchedulePostRATDList::ReleaseSuccessors(SUnit *SU) {
for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end();
I != E; ++I) {
ReleaseSucc(SU, &*I);
}
}
/// ScheduleNodeTopDown - Add the node to the schedule. Decrement the pending
/// count of its successors. If a successor pending count is zero, add it to
/// the Available queue.
void SchedulePostRATDList::ScheduleNodeTopDown(SUnit *SU, unsigned CurCycle) {
LLVM_DEBUG(dbgs() << "*** Scheduling [" << CurCycle << "]: ");
LLVM_DEBUG(dumpNode(*SU));
Sequence.push_back(SU);
assert(CurCycle >= SU->getDepth() &&
"Node scheduled above its depth!");
SU->setDepthToAtLeast(CurCycle);
ReleaseSuccessors(SU);
SU->isScheduled = true;
AvailableQueue.scheduledNode(SU);
}
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
/// emitNoop - Add a noop to the current instruction sequence.
void SchedulePostRATDList::emitNoop(unsigned CurCycle) {
LLVM_DEBUG(dbgs() << "*** Emitting noop in cycle " << CurCycle << '\n');
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
HazardRec->EmitNoop();
Sequence.push_back(nullptr); // NULL here means noop
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
++NumNoops;
}
/// ListScheduleTopDown - The main loop of list scheduling for top-down
/// schedulers.
void SchedulePostRATDList::ListScheduleTopDown() {
unsigned CurCycle = 0;
// We're scheduling top-down but we're visiting the regions in
// bottom-up order, so we don't know the hazards at the start of a
// region. So assume no hazards (this should usually be ok as most
// blocks are a single region).
HazardRec->Reset();
// Release any successors of the special Entry node.
ReleaseSuccessors(&EntrySU);
// Add all leaves to Available queue.
for (unsigned i = 0, e = SUnits.size(); i != e; ++i) {
// It is available if it has no predecessors.
if (!SUnits[i].NumPredsLeft && !SUnits[i].isAvailable) {
AvailableQueue.push(&SUnits[i]);
SUnits[i].isAvailable = true;
}
}
// In any cycle where we can't schedule any instructions, we must
// stall or emit a noop, depending on the target.
bool CycleHasInsts = false;
// While Available queue is not empty, grab the node with the highest
// priority. If it is not ready put it back. Schedule the node.
std::vector<SUnit*> NotReady;
Sequence.reserve(SUnits.size());
while (!AvailableQueue.empty() || !PendingQueue.empty()) {
// Check to see if any of the pending instructions are ready to issue. If
// so, add them to the available queue.
unsigned MinDepth = ~0u;
for (unsigned i = 0, e = PendingQueue.size(); i != e; ++i) {
if (PendingQueue[i]->getDepth() <= CurCycle) {
AvailableQueue.push(PendingQueue[i]);
PendingQueue[i]->isAvailable = true;
PendingQueue[i] = PendingQueue.back();
PendingQueue.pop_back();
--i; --e;
} else if (PendingQueue[i]->getDepth() < MinDepth)
MinDepth = PendingQueue[i]->getDepth();
}
2009-08-11 19:35:23 +02:00
LLVM_DEBUG(dbgs() << "\n*** Examining Available\n";
AvailableQueue.dump(this));
2009-08-11 19:35:23 +02:00
SUnit *FoundSUnit = nullptr, *NotPreferredSUnit = nullptr;
bool HasNoopHazards = false;
while (!AvailableQueue.empty()) {
SUnit *CurSUnit = AvailableQueue.pop();
ScheduleHazardRecognizer::HazardType HT =
HazardRec->getHazardType(CurSUnit, 0/*no stalls*/);
if (HT == ScheduleHazardRecognizer::NoHazard) {
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
if (HazardRec->ShouldPreferAnother(CurSUnit)) {
if (!NotPreferredSUnit) {
2014-10-29 16:23:11 +01:00
// If this is the first non-preferred node for this cycle, then
// record it and continue searching for a preferred node. If this
// is not the first non-preferred node, then treat it as though
// there had been a hazard.
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
NotPreferredSUnit = CurSUnit;
continue;
}
} else {
FoundSUnit = CurSUnit;
break;
}
}
// Remember if this is a noop hazard.
HasNoopHazards |= HT == ScheduleHazardRecognizer::NoopHazard;
NotReady.push_back(CurSUnit);
}
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
// If we have a non-preferred node, push it back onto the available list.
// If we did not find a preferred node, then schedule this first
// non-preferred node.
if (NotPreferredSUnit) {
if (!FoundSUnit) {
LLVM_DEBUG(
dbgs() << "*** Will schedule a non-preferred instruction...\n");
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
FoundSUnit = NotPreferredSUnit;
} else {
AvailableQueue.push(NotPreferredSUnit);
}
NotPreferredSUnit = nullptr;
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
}
// Add the nodes that aren't ready back onto the available list.
if (!NotReady.empty()) {
AvailableQueue.push_all(NotReady);
NotReady.clear();
}
// If we found a node to schedule...
if (FoundSUnit) {
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
// If we need to emit noops prior to this instruction, then do so.
unsigned NumPreNoops = HazardRec->PreEmitNoops(FoundSUnit);
for (unsigned i = 0; i != NumPreNoops; ++i)
emitNoop(CurCycle);
// ... schedule the node...
ScheduleNodeTopDown(FoundSUnit, CurCycle);
HazardRec->EmitInstruction(FoundSUnit);
CycleHasInsts = true;
if (HazardRec->atIssueLimit()) {
LLVM_DEBUG(dbgs() << "*** Max instructions per cycle " << CurCycle
<< '\n');
HazardRec->AdvanceCycle();
++CurCycle;
CycleHasInsts = false;
}
} else {
if (CycleHasInsts) {
LLVM_DEBUG(dbgs() << "*** Finished cycle " << CurCycle << '\n');
HazardRec->AdvanceCycle();
} else if (!HasNoopHazards) {
// Otherwise, we have a pipeline stall, but no other problem,
// just advance the current cycle and try again.
LLVM_DEBUG(dbgs() << "*** Stall in cycle " << CurCycle << '\n');
HazardRec->AdvanceCycle();
++NumStalls;
} else {
// Otherwise, we have no instructions to issue and we have instructions
// that will fault if we don't do this right. This is the case for
// processors without pipeline interlocks and other cases.
Add two additional hazard recognizer functions This adds two additional functions to the hazard recognizer interface. These are optional (in the sense that the default implementations preserve the current behavior), and used by the post-RA scheduler. Upcoming commits will use this functionality in order to improve dispatch-group formation on the POWER7 and related cores. Dispatch groups are an odd construct: sometimes we need to insert nops to force a new one to start (for performance reasons), and some instructions need to appear in certain positions within a group, but the groups are not fundamentally cycle based (they can contain instructions with data dependencies with non-trivial latencies). Motivation: unsigned PreEmitNoops(SUnit *) - Used to force the post-RA scheduler to insert nops to force a new dispatch group to begin. We already have a NoopHazard, and this is also still needed. However, NoopHazard only causes a nop to be inserted if there are no other available instructions, and so is not always sufficient. The number of nops to insert depends on state that only the hazard recognizer has, so a general callback is necessary. bool ShouldPreferAnother(SUnit *) - Used to avoid scheduling instructions that would start a new dispatch group when others are available that could be part of the current dispatch group. In this case, we don't want to issue nops, because the non-preferred instruction will implicitly start a new dispatch group regardless. Although the motivation for these functions is driven by the PowerPC backend, they are completely general. llvm-svn: 197084
2013-12-11 23:33:43 +01:00
emitNoop(CurCycle);
}
++CurCycle;
CycleHasInsts = false;
}
}
#ifndef NDEBUG
unsigned ScheduledNodes = VerifyScheduledDAG(/*isBottomUp=*/false);
unsigned Noops = 0;
for (unsigned i = 0, e = Sequence.size(); i != e; ++i)
if (!Sequence[i])
++Noops;
assert(Sequence.size() - Noops == ScheduledNodes &&
"The number of nodes scheduled doesn't match the expected number!");
#endif // NDEBUG
}
// EmitSchedule - Emit the machine code in scheduled order.
void SchedulePostRATDList::EmitSchedule() {
RegionBegin = RegionEnd;
// If first instruction was a DBG_VALUE then put it back.
if (FirstDbgValue)
BB->splice(RegionEnd, BB, FirstDbgValue);
// Then re-insert them according to the given schedule.
for (unsigned i = 0, e = Sequence.size(); i != e; i++) {
if (SUnit *SU = Sequence[i])
BB->splice(RegionEnd, BB, SU->getInstr());
else
// Null SUnit* is a noop.
TII->insertNoop(*BB, RegionEnd);
// Update the Begin iterator, as the first instruction in the block
// may have been scheduled later.
if (i == 0)
RegionBegin = std::prev(RegionEnd);
}
// Reinsert any remaining debug_values.
for (std::vector<std::pair<MachineInstr *, MachineInstr *> >::iterator
DI = DbgValues.end(), DE = DbgValues.begin(); DI != DE; --DI) {
std::pair<MachineInstr *, MachineInstr *> P = *std::prev(DI);
MachineInstr *DbgValue = P.first;
MachineBasicBlock::iterator OrigPrivMI = P.second;
BB->splice(++OrigPrivMI, BB, DbgValue);
}
DbgValues.clear();
FirstDbgValue = nullptr;
}