aboutsummaryrefslogtreecommitdiffstats
path: root/include
diff options
context:
space:
mode:
authorChandler Carruth <chandlerc@gmail.com>2013-01-07 01:37:14 +0000
committerChandler Carruth <chandlerc@gmail.com>2013-01-07 01:37:14 +0000
commitaeef83c6afa1e18d1cf9d359cc678ca0ad556175 (patch)
treed79f0bdd4339c6518779ad9a1db4e7d220606a14 /include
parent916d52a03ebd45f4b6d9dea185ee616623feeaf0 (diff)
downloadexternal_llvm-aeef83c6afa1e18d1cf9d359cc678ca0ad556175.zip
external_llvm-aeef83c6afa1e18d1cf9d359cc678ca0ad556175.tar.gz
external_llvm-aeef83c6afa1e18d1cf9d359cc678ca0ad556175.tar.bz2
Switch TargetTransformInfo from an immutable analysis pass that requires
a TargetMachine to construct (and thus isn't always available), to an analysis group that supports layered implementations much like AliasAnalysis does. This is a pretty massive change, with a few parts that I was unable to easily separate (sorry), so I'll walk through it. The first step of this conversion was to make TargetTransformInfo an analysis group, and to sink the nonce implementations in ScalarTargetTransformInfo and VectorTargetTranformInfo into a NoTargetTransformInfo pass. This allows other passes to add a hard requirement on TTI, and assume they will always get at least on implementation. The TargetTransformInfo analysis group leverages the delegation chaining trick that AliasAnalysis uses, where the base class for the analysis group delegates to the previous analysis *pass*, allowing all but tho NoFoo analysis passes to only implement the parts of the interfaces they support. It also introduces a new trick where each pass in the group retains a pointer to the top-most pass that has been initialized. This allows passes to implement one API in terms of another API and benefit when some other pass above them in the stack has more precise results for the second API. The second step of this conversion is to create a pass that implements the TargetTransformInfo analysis using the target-independent abstractions in the code generator. This replaces the ScalarTargetTransformImpl and VectorTargetTransformImpl classes in lib/Target with a single pass in lib/CodeGen called BasicTargetTransformInfo. This class actually provides most of the TTI functionality, basing it upon the TargetLowering abstraction and other information in the target independent code generator. The third step of the conversion adds support to all TargetMachines to register custom analysis passes. This allows building those passes with access to TargetLowering or other target-specific classes, and it also allows each target to customize the set of analysis passes desired in the pass manager. The baseline LLVMTargetMachine implements this interface to add the BasicTTI pass to the pass manager, and all of the tools that want to support target-aware TTI passes call this routine on whatever target machine they end up with to add the appropriate passes. The fourth step of the conversion created target-specific TTI analysis passes for the X86 and ARM backends. These passes contain the custom logic that was previously in their extensions of the ScalarTargetTransformInfo and VectorTargetTransformInfo interfaces. I separated them into their own file, as now all of the interface bits are private and they just expose a function to create the pass itself. Then I extended these target machines to set up a custom set of analysis passes, first adding BasicTTI as a fallback, and then adding their customized TTI implementations. The fourth step required logic that was shared between the target independent layer and the specific targets to move to a different interface, as they no longer derive from each other. As a consequence, a helper functions were added to TargetLowering representing the common logic needed both in the target implementation and the codegen implementation of the TTI pass. While technically this is the only change that could have been committed separately, it would have been a nightmare to extract. The final step of the conversion was just to delete all the old boilerplate. This got rid of the ScalarTargetTransformInfo and VectorTargetTransformInfo classes, all of the support in all of the targets for producing instances of them, and all of the support in the tools for manually constructing a pass based around them. Now that TTI is a relatively normal analysis group, two things become straightforward. First, we can sink it into lib/Analysis which is a more natural layer for it to live. Second, clients of this interface can depend on it *always* being available which will simplify their code and behavior. These (and other) simplifications will follow in subsequent commits, this one is clearly big enough. Finally, I'm very aware that much of the comments and documentation needs to be updated. As soon as I had this working, and plausibly well commented, I wanted to get it committed and in front of the build bots. I'll be doing a few passes over documentation later if it sticks. Commits to update DragonEgg and Clang will be made presently. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171681 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'include')
-rw-r--r--include/llvm/CodeGen/Passes.h6
-rw-r--r--include/llvm/InitializePasses.h1
-rw-r--r--include/llvm/Target/TargetLowering.h12
-rw-r--r--include/llvm/Target/TargetMachine.h12
-rw-r--r--include/llvm/Target/TargetTransformImpl.h104
-rw-r--r--include/llvm/TargetTransformInfo.h198
6 files changed, 47 insertions, 286 deletions
diff --git a/include/llvm/CodeGen/Passes.h b/include/llvm/CodeGen/Passes.h
index 7831d12..0d02700 100644
--- a/include/llvm/CodeGen/Passes.h
+++ b/include/llvm/CodeGen/Passes.h
@@ -280,6 +280,12 @@ protected:
/// List of target independent CodeGen pass IDs.
namespace llvm {
+ /// \brief Create a basic TargetTransformInfo analysis pass.
+ ///
+ /// This pass implements the target transform info analysis using the target
+ /// independent information available to the LLVM code generator.
+ ImmutablePass *createBasicTargetTransformInfoPass(const TargetLowering *TLI);
+
/// createUnreachableBlockEliminationPass - The LLVM code generator does not
/// work well with unreachable basic blocks (what live ranges make sense for a
/// block that cannot be reached?). As such, a code generator should either
diff --git a/include/llvm/InitializePasses.h b/include/llvm/InitializePasses.h
index 9ee74dd..79037f7 100644
--- a/include/llvm/InitializePasses.h
+++ b/include/llvm/InitializePasses.h
@@ -69,6 +69,7 @@ void initializeArgPromotionPass(PassRegistry&);
void initializeBarrierNoopPass(PassRegistry&);
void initializeBasicAliasAnalysisPass(PassRegistry&);
void initializeBasicCallGraphPass(PassRegistry&);
+void initializeBasicTTIPass(PassRegistry&);
void initializeBlockExtractorPassPass(PassRegistry&);
void initializeBlockFrequencyInfoPass(PassRegistry&);
void initializeBlockPlacementPass(PassRegistry&);
diff --git a/include/llvm/Target/TargetLowering.h b/include/llvm/Target/TargetLowering.h
index ce6abc5..8359a5c 100644
--- a/include/llvm/Target/TargetLowering.h
+++ b/include/llvm/Target/TargetLowering.h
@@ -895,6 +895,18 @@ public:
}
//===--------------------------------------------------------------------===//
+ /// \name Helpers for TargetTransformInfo implementations
+ /// @{
+
+ /// Get the ISD node that corresponds to the Instruction class opcode.
+ int InstructionOpcodeToISD(unsigned Opcode) const;
+
+ /// Estimate the cost of type-legalization and the legalized type.
+ std::pair<unsigned, MVT> getTypeLegalizationCost(Type *Ty) const;
+
+ /// @}
+
+ //===--------------------------------------------------------------------===//
// TargetLowering Optimization Methods
//
diff --git a/include/llvm/Target/TargetMachine.h b/include/llvm/Target/TargetMachine.h
index 5756f2c..aa049f0 100644
--- a/include/llvm/Target/TargetMachine.h
+++ b/include/llvm/Target/TargetMachine.h
@@ -108,10 +108,6 @@ public:
virtual const TargetLowering *getTargetLowering() const { return 0; }
virtual const TargetSelectionDAGInfo *getSelectionDAGInfo() const{ return 0; }
virtual const DataLayout *getDataLayout() const { return 0; }
- virtual const ScalarTargetTransformInfo*
- getScalarTargetTransformInfo() const { return 0; }
- virtual const VectorTargetTransformInfo*
- getVectorTargetTransformInfo() const { return 0; }
/// getMCAsmInfo - Return target specific asm information.
///
@@ -232,6 +228,9 @@ public:
/// sections.
static void setFunctionSections(bool);
+ /// \brief Register analysis passes for this target with a pass manager.
+ virtual void addAnalysisPasses(PassManagerBase &) {}
+
/// CodeGenFileType - These enums are meant to be passed into
/// addPassesToEmitFile to indicate what type of file to emit, and returned by
/// it to indicate what type of file could actually be made.
@@ -290,6 +289,11 @@ protected: // Can only create subclasses.
CodeGenOpt::Level OL);
public:
+ /// \brief Register analysis passes for this target with a pass manager.
+ ///
+ /// This registers target independent analysis passes.
+ virtual void addAnalysisPasses(PassManagerBase &PM);
+
/// createPassConfig - Create a pass configuration object to be used by
/// addPassToEmitX methods for generating a pipeline of CodeGen passes.
virtual TargetPassConfig *createPassConfig(PassManagerBase &PM);
diff --git a/include/llvm/Target/TargetTransformImpl.h b/include/llvm/Target/TargetTransformImpl.h
deleted file mode 100644
index 2069927..0000000
--- a/include/llvm/Target/TargetTransformImpl.h
+++ /dev/null
@@ -1,104 +0,0 @@
-//=- llvm/Target/TargetTransformImpl.h - Target Loop Trans Info----*- C++ -*-=//
-//
-// The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-//
-// This file contains the target-specific implementations of the
-// TargetTransform interfaces.
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef LLVM_TARGET_TARGET_TRANSFORMATION_IMPL_H
-#define LLVM_TARGET_TARGET_TRANSFORMATION_IMPL_H
-
-#include "llvm/CodeGen/ValueTypes.h"
-#include "llvm/TargetTransformInfo.h"
-
-namespace llvm {
-
-class TargetLowering;
-
-/// ScalarTargetTransformInfo - This is a default implementation for the
-/// ScalarTargetTransformInfo interface. Different targets can implement
-/// this interface differently.
-class ScalarTargetTransformImpl : public ScalarTargetTransformInfo {
-protected:
- const TargetLowering *TLI;
-
-public:
- /// Ctor
- explicit ScalarTargetTransformImpl(const TargetLowering *TL) : TLI(TL) {}
-
- virtual bool isLegalAddImmediate(int64_t imm) const;
-
- virtual bool isLegalICmpImmediate(int64_t imm) const;
-
- virtual bool isLegalAddressingMode(Type *Ty, GlobalValue *BaseGV,
- int64_t BaseOffset, bool HasBaseReg,
- int64_t Scale) const;
-
- virtual bool isTruncateFree(Type *Ty1, Type *Ty2) const;
-
- virtual bool isTypeLegal(Type *Ty) const;
-
- virtual unsigned getJumpBufAlignment() const;
-
- virtual unsigned getJumpBufSize() const;
-
- virtual bool shouldBuildLookupTables() const;
-};
-
-class VectorTargetTransformImpl : public VectorTargetTransformInfo {
-protected:
- const TargetLowering *TLI;
-
- /// Estimate the cost of type-legalization and the legalized type.
- std::pair<unsigned, MVT> getTypeLegalizationCost(Type *Ty) const;
-
- /// Estimate the overhead of scalarizing an instruction. Insert and Extract
- /// are set if the result needs to be inserted and/or extracted from vectors.
- unsigned getScalarizationOverhead(Type *Ty, bool Insert, bool Extract) const;
-
- // Get the ISD node that corresponds to the Instruction class opcode.
- int InstructionOpcodeToISD(unsigned Opcode) const;
-
-public:
- explicit VectorTargetTransformImpl(const TargetLowering *TL) : TLI(TL) {}
-
- virtual ~VectorTargetTransformImpl() {}
-
- virtual unsigned getNumberOfRegisters(bool Vector) const;
-
- virtual unsigned getArithmeticInstrCost(unsigned Opcode, Type *Ty) const;
-
- virtual unsigned getShuffleCost(ShuffleKind Kind, Type *Tp,
- int Index, Type *SubTp) const;
-
- virtual unsigned getCastInstrCost(unsigned Opcode, Type *Dst,
- Type *Src) const;
-
- virtual unsigned getCFInstrCost(unsigned Opcode) const;
-
- virtual unsigned getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
- Type *CondTy) const;
-
- virtual unsigned getVectorInstrCost(unsigned Opcode, Type *Val,
- unsigned Index) const;
-
- virtual unsigned getMemoryOpCost(unsigned Opcode, Type *Src,
- unsigned Alignment,
- unsigned AddressSpace) const;
-
- virtual unsigned getIntrinsicInstrCost(Intrinsic::ID, Type *RetTy,
- ArrayRef<Type*> Tys) const;
-
- virtual unsigned getNumberOfParts(Type *Tp) const;
-};
-
-} // end llvm namespace
-
-#endif
diff --git a/include/llvm/TargetTransformInfo.h b/include/llvm/TargetTransformInfo.h
index 97d0304..253960d 100644
--- a/include/llvm/TargetTransformInfo.h
+++ b/include/llvm/TargetTransformInfo.h
@@ -30,9 +30,6 @@
namespace llvm {
-class ScalarTargetTransformInfo;
-class VectorTargetTransformInfo;
-
/// TargetTransformInfo - This pass provides access to the codegen
/// interfaces that are needed for IR-level transformations.
class TargetTransformInfo {
@@ -42,11 +39,26 @@ protected:
/// This is used to implement the default behavior all of the methods which
/// is to delegate up through the stack of TTIs until one can answer the
/// query.
- const TargetTransformInfo *PrevTTI;
+ TargetTransformInfo *PrevTTI;
- /// Every subclass must initialize the base with the previous TTI in the
- /// stack, or 0 if there is no previous TTI in the stack.
- TargetTransformInfo(const TargetTransformInfo *PrevTTI) : PrevTTI(PrevTTI) {}
+ /// \brief The top of the stack of TTI analyses available.
+ ///
+ /// This is a convenience routine maintained as TTI analyses become available
+ /// that complements the PrevTTI delegation chain. When one part of an
+ /// analysis pass wants to query another part of the analysis pass it can use
+ /// this to start back at the top of the stack.
+ TargetTransformInfo *TopTTI;
+
+ /// All pass subclasses must in their initializePass routine call
+ /// pushTTIStack with themselves to update the pointers tracking the previous
+ /// TTI instance in the analysis group's stack, and the top of the analysis
+ /// group's stack.
+ void pushTTIStack(Pass *P);
+
+ /// All pass subclasses must in their finalizePass routine call popTTIStack
+ /// to update the pointers tracking the previous TTI instance in the analysis
+ /// group's stack, and the top of the analysis group's stack.
+ void popTTIStack();
/// All pass subclasses must call TargetTransformInfo::getAnalysisUsage.
virtual void getAnalysisUsage(AnalysisUsage &AU) const;
@@ -184,177 +196,7 @@ public:
/// This class provides the base case for the stack of TTI analyses. It doesn't
/// delegate to anything and uses the STTI and VTTI objects passed in to
/// satisfy the queries.
-ImmutablePass *createNoTTIPass(const ScalarTargetTransformInfo *S,
- const VectorTargetTransformInfo *V);
-
-
-// ---------------------------------------------------------------------------//
-// The classes below are inherited and implemented by target-specific classes
-// in the codegen.
-// ---------------------------------------------------------------------------//
-
-/// ScalarTargetTransformInfo - This interface is used by IR-level passes
-/// that need target-dependent information for generic scalar transformations.
-/// LSR, and LowerInvoke use this interface.
-class ScalarTargetTransformInfo {
-public:
- /// PopcntHwSupport - Hardware support for population count. Compared to the
- /// SW implementation, HW support is supposed to significantly boost the
- /// performance when the population is dense, and it may or may not degrade
- /// performance if the population is sparse. A HW support is considered as
- /// "Fast" if it can outperform, or is on a par with, SW implementaion when
- /// the population is sparse; otherwise, it is considered as "Slow".
- enum PopcntHwSupport {
- None,
- Fast,
- Slow
- };
-
- virtual ~ScalarTargetTransformInfo() {}
-
- /// isLegalAddImmediate - Return true if the specified immediate is legal
- /// add immediate, that is the target has add instructions which can add
- /// a register with the immediate without having to materialize the
- /// immediate into a register.
- virtual bool isLegalAddImmediate(int64_t) const {
- return false;
- }
- /// isLegalICmpImmediate - Return true if the specified immediate is legal
- /// icmp immediate, that is the target has icmp instructions which can compare
- /// a register against the immediate without having to materialize the
- /// immediate into a register.
- virtual bool isLegalICmpImmediate(int64_t) const {
- return false;
- }
- /// isLegalAddressingMode - Return true if the addressing mode represented by
- /// AM is legal for this target, for a load/store of the specified type.
- /// The type may be VoidTy, in which case only return true if the addressing
- /// mode is legal for a load/store of any legal type.
- /// TODO: Handle pre/postinc as well.
- virtual bool isLegalAddressingMode(Type *Ty, GlobalValue *BaseGV,
- int64_t BaseOffset, bool HasBaseReg,
- int64_t Scale) const {
- return false;
- }
- /// isTruncateFree - Return true if it's free to truncate a value of
- /// type Ty1 to type Ty2. e.g. On x86 it's free to truncate a i32 value in
- /// register EAX to i16 by referencing its sub-register AX.
- virtual bool isTruncateFree(Type *Ty1, Type *Ty2) const {
- return false;
- }
- /// Is this type legal.
- virtual bool isTypeLegal(Type *Ty) const {
- return false;
- }
- /// getJumpBufAlignment - returns the target's jmp_buf alignment in bytes
- virtual unsigned getJumpBufAlignment() const {
- return 0;
- }
- /// getJumpBufSize - returns the target's jmp_buf size in bytes.
- virtual unsigned getJumpBufSize() const {
- return 0;
- }
- /// shouldBuildLookupTables - Return true if switches should be turned into
- /// lookup tables for the target.
- virtual bool shouldBuildLookupTables() const {
- return true;
- }
- /// getPopcntHwSupport - Return hardware support for population count.
- virtual PopcntHwSupport getPopcntHwSupport(unsigned IntTyWidthInBit) const {
- return None;
- }
- /// getIntImmCost - Return the expected cost of materializing the given
- /// integer immediate of the specified type.
- virtual unsigned getIntImmCost(const APInt&, Type*) const {
- // The default assumption is that the immediate is cheap.
- return 1;
- }
-};
-
-/// VectorTargetTransformInfo - This interface is used by the vectorizers
-/// to estimate the profitability of vectorization for different instructions.
-/// This interface provides the cost of different IR instructions. The cost
-/// is unit-less and represents the estimated throughput of the instruction
-/// (not the latency!) assuming that all branches are predicted, cache is hit,
-/// etc.
-class VectorTargetTransformInfo {
-public:
- virtual ~VectorTargetTransformInfo() {}
-
- enum ShuffleKind {
- Broadcast, // Broadcast element 0 to all other elements.
- Reverse, // Reverse the order of the vector.
- InsertSubvector, // InsertSubvector. Index indicates start offset.
- ExtractSubvector // ExtractSubvector Index indicates start offset.
- };
-
- /// \return The number of scalar or vector registers that the target has.
- /// If 'Vectors' is true, it returns the number of vector registers. If it is
- /// set to false, it returns the number of scalar registers.
- virtual unsigned getNumberOfRegisters(bool Vector) const {
- return 8;
- }
-
- /// \return The expected cost of arithmetic ops, such as mul, xor, fsub, etc.
- virtual unsigned getArithmeticInstrCost(unsigned Opcode, Type *Ty) const {
- return 1;
- }
-
- /// \return The cost of a shuffle instruction of kind Kind and of type Tp.
- /// The index and subtype parameters are used by the subvector insertion and
- /// extraction shuffle kinds.
- virtual unsigned getShuffleCost(ShuffleKind Kind, Type *Tp,
- int Index = 0, Type *SubTp = 0) const {
- return 1;
- }
-
- /// \return The expected cost of cast instructions, such as bitcast, trunc,
- /// zext, etc.
- virtual unsigned getCastInstrCost(unsigned Opcode, Type *Dst,
- Type *Src) const {
- return 1;
- }
-
- /// \return The expected cost of control-flow related instrutctions such as
- /// Phi, Ret, Br.
- virtual unsigned getCFInstrCost(unsigned Opcode) const {
- return 1;
- }
-
- /// \returns The expected cost of compare and select instructions.
- virtual unsigned getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
- Type *CondTy = 0) const {
- return 1;
- }
-
- /// \return The expected cost of vector Insert and Extract.
- /// Use -1 to indicate that there is no information on the index value.
- virtual unsigned getVectorInstrCost(unsigned Opcode, Type *Val,
- unsigned Index = -1) const {
- return 1;
- }
-
- /// \return The cost of Load and Store instructions.
- virtual unsigned getMemoryOpCost(unsigned Opcode, Type *Src,
- unsigned Alignment,
- unsigned AddressSpace) const {
- return 1;
- }
-
- /// \returns The cost of Intrinsic instructions.
- virtual unsigned getIntrinsicInstrCost(Intrinsic::ID,
- Type *RetTy,
- ArrayRef<Type*> Tys) const {
- return 1;
- }
-
- /// \returns The number of pieces into which the provided type must be
- /// split during legalization. Zero is returned when the answer is unknown.
- virtual unsigned getNumberOfParts(Type *Tp) const {
- return 0;
- }
-};
-
+ImmutablePass *createNoTargetTransformInfoPass();
} // End llvm namespace