Inlining and unrolling heuristics should be aware of free truncs.

We want heuristics to be based on accurate data, but more importantly
we don't want llvm to behave randomly. A benign trunc inserted by an
upstream pass should not cause a wild swings in optimization
level. See PR11034. It's a general problem with threshold-based
heuristics, but we can make it less bad.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140919 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Andrew Trick 2011-10-01 01:39:05 +00:00
parent 5c655413cf
commit b2ab2fa524
6 changed files with 44 additions and 18 deletions

View File

@ -18,6 +18,9 @@
#include "llvm/ADT/DenseMap.h"
namespace llvm {
class TargetData;
// CodeMetrics - Calculate size and a few similar metrics for a set of
// basic blocks.
struct CodeMetrics {
@ -68,11 +71,11 @@ namespace llvm {
/// analyzeBasicBlock - Add information about the specified basic block
/// to the current structure.
void analyzeBasicBlock(const BasicBlock *BB);
void analyzeBasicBlock(const BasicBlock *BB, const TargetData *TD = 0);
/// analyzeFunction - Add information about the specified function
/// to the current structure.
void analyzeFunction(Function *F);
void analyzeFunction(Function *F, const TargetData *TD = 0);
/// CountCodeReductionForConstant - Figure out an approximation for how
/// many instructions will be constant folded if the specified value is

View File

@ -29,6 +29,7 @@ namespace llvm {
class CallSite;
template<class PtrType, unsigned SmallSize>
class SmallPtrSet;
class TargetData;
namespace InlineConstants {
// Various magic constants used to adjust heuristics.
@ -113,7 +114,7 @@ namespace llvm {
/// analyzeFunction - Add information about the specified function
/// to the current structure.
void analyzeFunction(Function *F);
void analyzeFunction(Function *F, const TargetData *TD);
/// NeverInline - Returns true if the function should never be
/// inlined into any caller.
@ -124,11 +125,17 @@ namespace llvm {
// the ValueMap will update itself when this happens.
ValueMap<const Function *, FunctionInfo> CachedFunctionInfo;
// TargetData if available, or null.
const TargetData *TD;
int CountBonusForConstant(Value *V, Constant *C = NULL);
int ConstantFunctionBonus(CallSite CS, Constant *C);
int getInlineSize(CallSite CS, Function *Callee);
int getInlineBonuses(CallSite CS, Function *Callee);
public:
InlineCostAnalyzer(): TD(0) {}
void setTargetData(const TargetData *TData) { TD = TData; }
/// getInlineCost - The heuristic used to determine if we should inline the
/// function call or not.

View File

@ -15,6 +15,7 @@
#include "llvm/Support/CallSite.h"
#include "llvm/CallingConv.h"
#include "llvm/IntrinsicInst.h"
#include "llvm/Target/TargetData.h"
#include "llvm/ADT/SmallPtrSet.h"
using namespace llvm;
@ -52,7 +53,8 @@ bool llvm::callIsSmall(const Function *F) {
/// analyzeBasicBlock - Fill in the current structure with information gleaned
/// from the specified block.
void CodeMetrics::analyzeBasicBlock(const BasicBlock *BB) {
void CodeMetrics::analyzeBasicBlock(const BasicBlock *BB,
const TargetData *TD) {
++NumBlocks;
unsigned NumInstsBeforeThisBB = NumInsts;
for (BasicBlock::const_iterator II = BB->begin(), E = BB->end();
@ -105,6 +107,11 @@ void CodeMetrics::analyzeBasicBlock(const BasicBlock *BB) {
if (CI->isLosslessCast() || isa<IntToPtrInst>(CI) ||
isa<PtrToIntInst>(CI))
continue;
// trunc to a native type is free (assuming the target has compare and
// shift-right of the same width).
if (isa<TruncInst>(CI) && TD &&
TD->isLegalInteger(TD->getTypeSizeInBits(CI->getType())))
continue;
// Result of a cmp instruction is often extended (to be used by other
// cmp instructions, logical or return instructions). These are usually
// nop on most sane targets.
@ -217,7 +224,7 @@ unsigned CodeMetrics::CountCodeReductionForAlloca(Value *V) {
/// analyzeFunction - Fill in the current structure with information gleaned
/// from the specified function.
void CodeMetrics::analyzeFunction(Function *F) {
void CodeMetrics::analyzeFunction(Function *F, const TargetData *TD) {
// If this function contains a call to setjmp or _setjmp, never inline
// it. This is a hack because we depend on the user marking their local
// variables as volatile if they are live across a setjmp call, and they
@ -227,13 +234,14 @@ void CodeMetrics::analyzeFunction(Function *F) {
// Look at the size of the callee.
for (Function::const_iterator BB = F->begin(), E = F->end(); BB != E; ++BB)
analyzeBasicBlock(&*BB);
analyzeBasicBlock(&*BB, TD);
}
/// analyzeFunction - Fill in the current structure with information gleaned
/// from the specified function.
void InlineCostAnalyzer::FunctionInfo::analyzeFunction(Function *F) {
Metrics.analyzeFunction(F);
void InlineCostAnalyzer::FunctionInfo::analyzeFunction(Function *F,
const TargetData *TD) {
Metrics.analyzeFunction(F, TD);
// A function with exactly one return has it removed during the inlining
// process (see InlineFunction), so don't count it.
@ -275,7 +283,7 @@ int InlineCostAnalyzer::getSpecializationBonus(Function *Callee,
// If we haven't calculated this information yet, do so now.
if (CalleeFI->Metrics.NumBlocks == 0)
CalleeFI->analyzeFunction(Callee);
CalleeFI->analyzeFunction(Callee, TD);
unsigned ArgNo = 0;
unsigned i = 0;
@ -365,7 +373,7 @@ int InlineCostAnalyzer::getInlineSize(CallSite CS, Function *Callee) {
// If we haven't calculated this information yet, do so now.
if (CalleeFI->Metrics.NumBlocks == 0)
CalleeFI->analyzeFunction(Callee);
CalleeFI->analyzeFunction(Callee, TD);
// InlineCost - This value measures how good of an inline candidate this call
// site is to inline. A lower inline cost make is more likely for the call to
@ -418,7 +426,7 @@ int InlineCostAnalyzer::getInlineBonuses(CallSite CS, Function *Callee) {
// If we haven't calculated this information yet, do so now.
if (CalleeFI->Metrics.NumBlocks == 0)
CalleeFI->analyzeFunction(Callee);
CalleeFI->analyzeFunction(Callee, TD);
bool isDirectCall = CS.getCalledFunction() == Callee;
Instruction *TheCall = CS.getInstruction();
@ -486,7 +494,7 @@ InlineCost InlineCostAnalyzer::getInlineCost(CallSite CS,
// If we haven't calculated this information yet, do so now.
if (CalleeFI->Metrics.NumBlocks == 0)
CalleeFI->analyzeFunction(Callee);
CalleeFI->analyzeFunction(Callee, TD);
// If we should never inline this, return a huge cost.
if (CalleeFI->NeverInline())
@ -505,7 +513,7 @@ InlineCost InlineCostAnalyzer::getInlineCost(CallSite CS,
// If we haven't calculated this information yet, do so now.
if (CallerFI.Metrics.NumBlocks == 0) {
CallerFI.analyzeFunction(Caller);
CallerFI.analyzeFunction(Caller, TD);
// Recompute the CalleeFI pointer, getting Caller could have invalidated
// it.
@ -544,7 +552,7 @@ InlineCost InlineCostAnalyzer::getSpecializationCost(Function *Callee,
// If we haven't calculated this information yet, do so now.
if (CalleeFI->Metrics.NumBlocks == 0)
CalleeFI->analyzeFunction(Callee);
CalleeFI->analyzeFunction(Callee, TD);
int Cost = 0;
@ -570,7 +578,7 @@ float InlineCostAnalyzer::getInlineFudgeFactor(CallSite CS) {
// If we haven't calculated this information yet, do so now.
if (CalleeFI.Metrics.NumBlocks == 0)
CalleeFI.analyzeFunction(Callee);
CalleeFI.analyzeFunction(Callee, TD);
float Factor = 1.0f;
// Single BB functions are often written to be inlined.

View File

@ -23,6 +23,7 @@
#include "llvm/Support/CallSite.h"
#include "llvm/Transforms/IPO.h"
#include "llvm/Transforms/IPO/InlinerPass.h"
#include "llvm/Target/TargetData.h"
#include "llvm/ADT/SmallPtrSet.h"
using namespace llvm;
@ -74,6 +75,8 @@ Pass *llvm::createAlwaysInlinerPass() { return new AlwaysInliner(); }
// doInitialization - Initializes the vector of functions that have not
// been annotated with the "always inline" attribute.
bool AlwaysInliner::doInitialization(CallGraph &CG) {
CA.setTargetData(getAnalysisIfAvailable<TargetData>());
Module &M = CG.getModule();
for (Module::iterator I = M.begin(), E = M.end();

View File

@ -22,6 +22,7 @@
#include "llvm/Support/CallSite.h"
#include "llvm/Transforms/IPO.h"
#include "llvm/Transforms/IPO/InlinerPass.h"
#include "llvm/Target/TargetData.h"
#include "llvm/ADT/SmallPtrSet.h"
using namespace llvm;
@ -75,6 +76,7 @@ Pass *llvm::createFunctionInliningPass(int Threshold) {
// doInitialization - Initializes the vector of functions that have been
// annotated with the noinline attribute.
bool SimpleInliner::doInitialization(CallGraph &CG) {
CA.setTargetData(getAnalysisIfAvailable<TargetData>());
Module &M = CG.getModule();

View File

@ -22,6 +22,7 @@
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/Utils/UnrollLoop.h"
#include "llvm/Target/TargetData.h"
#include <climits>
using namespace llvm;
@ -107,11 +108,12 @@ Pass *llvm::createLoopUnrollPass(int Threshold, int Count, int AllowPartial) {
}
/// ApproximateLoopSize - Approximate the size of the loop.
static unsigned ApproximateLoopSize(const Loop *L, unsigned &NumCalls) {
static unsigned ApproximateLoopSize(const Loop *L, unsigned &NumCalls,
const TargetData *TD) {
CodeMetrics Metrics;
for (Loop::block_iterator I = L->block_begin(), E = L->block_end();
I != E; ++I)
Metrics.analyzeBasicBlock(*I);
Metrics.analyzeBasicBlock(*I, TD);
NumCalls = Metrics.NumInlineCandidates;
unsigned LoopSize = Metrics.NumInsts;
@ -174,8 +176,9 @@ bool LoopUnroll::runOnLoop(Loop *L, LPPassManager &LPM) {
// Enforce the threshold.
if (Threshold != NoThreshold) {
const TargetData *TD = getAnalysisIfAvailable<TargetData>();
unsigned NumInlineCandidates;
unsigned LoopSize = ApproximateLoopSize(L, NumInlineCandidates);
unsigned LoopSize = ApproximateLoopSize(L, NumInlineCandidates, TD);
DEBUG(dbgs() << " Loop Size = " << LoopSize << "\n");
if (NumInlineCandidates != 0) {
DEBUG(dbgs() << " Not unrolling loop with inlinable calls.\n");