mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2024-08-26 23:29:22 +00:00
The patch disabled unrolling in loop vectorization pass when VF==1 on x86 architecture, by setting MaxInterleaveFactor to 1. Unrolling in loop vectorization pass may introduce the cost of overflow check, memory boundary check and extra prologue/epilogue code when regular unroller will unroll the loop another time. Disable it when VF==1 remove the unnecessary cost on x86. The same can be done for other platforms after verifying interleaving/memory bound checking to be not perf critical on those platforms. Differential Revision: http://reviews.llvm.org/D9515 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@236613 91177308-0d34-0410-b5e6-96231b3b80d8
79 lines
2.7 KiB
C++
79 lines
2.7 KiB
C++
//===-- AMDGPUTargetTransformInfo.h - AMDGPU specific TTI -------*- C++ -*-===//
|
|
//
|
|
// The LLVM Compiler Infrastructure
|
|
//
|
|
// This file is distributed under the University of Illinois Open Source
|
|
// License. See LICENSE.TXT for details.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
/// \file
|
|
/// This file a TargetTransformInfo::Concept conforming object specific to the
|
|
/// AMDGPU target machine. It uses the target's detailed information to
|
|
/// provide more precise answers to certain TTI queries, while letting the
|
|
/// target independent and default TTI implementations handle the rest.
|
|
///
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#ifndef LLVM_LIB_TARGET_R600_AMDGPUTARGETTRANSFORMINFO_H
|
|
#define LLVM_LIB_TARGET_R600_AMDGPUTARGETTRANSFORMINFO_H
|
|
|
|
#include "AMDGPU.h"
|
|
#include "AMDGPUTargetMachine.h"
|
|
#include "llvm/Analysis/TargetTransformInfo.h"
|
|
#include "llvm/CodeGen/BasicTTIImpl.h"
|
|
#include "llvm/Target/TargetLowering.h"
|
|
|
|
namespace llvm {
|
|
|
|
class AMDGPUTTIImpl : public BasicTTIImplBase<AMDGPUTTIImpl> {
|
|
typedef BasicTTIImplBase<AMDGPUTTIImpl> BaseT;
|
|
typedef TargetTransformInfo TTI;
|
|
friend BaseT;
|
|
|
|
const AMDGPUSubtarget *ST;
|
|
const AMDGPUTargetLowering *TLI;
|
|
|
|
const AMDGPUSubtarget *getST() const { return ST; }
|
|
const AMDGPUTargetLowering *getTLI() const { return TLI; }
|
|
|
|
public:
|
|
explicit AMDGPUTTIImpl(const AMDGPUTargetMachine *TM)
|
|
: BaseT(TM), ST(TM->getSubtargetImpl()), TLI(ST->getTargetLowering()) {}
|
|
|
|
// Provide value semantics. MSVC requires that we spell all of these out.
|
|
AMDGPUTTIImpl(const AMDGPUTTIImpl &Arg)
|
|
: BaseT(static_cast<const BaseT &>(Arg)), ST(Arg.ST), TLI(Arg.TLI) {}
|
|
AMDGPUTTIImpl(AMDGPUTTIImpl &&Arg)
|
|
: BaseT(std::move(static_cast<BaseT &>(Arg))), ST(std::move(Arg.ST)),
|
|
TLI(std::move(Arg.TLI)) {}
|
|
AMDGPUTTIImpl &operator=(const AMDGPUTTIImpl &RHS) {
|
|
BaseT::operator=(static_cast<const BaseT &>(RHS));
|
|
ST = RHS.ST;
|
|
TLI = RHS.TLI;
|
|
return *this;
|
|
}
|
|
AMDGPUTTIImpl &operator=(AMDGPUTTIImpl &&RHS) {
|
|
BaseT::operator=(std::move(static_cast<BaseT &>(RHS)));
|
|
ST = std::move(RHS.ST);
|
|
TLI = std::move(RHS.TLI);
|
|
return *this;
|
|
}
|
|
|
|
bool hasBranchDivergence() { return true; }
|
|
|
|
void getUnrollingPreferences(Loop *L, TTI::UnrollingPreferences &UP);
|
|
|
|
TTI::PopcntSupportKind getPopcntSupport(unsigned TyWidth) {
|
|
assert(isPowerOf2_32(TyWidth) && "Ty width must be power of 2");
|
|
return ST->hasBCNT(TyWidth) ? TTI::PSK_FastHardware : TTI::PSK_Software;
|
|
}
|
|
|
|
unsigned getNumberOfRegisters(bool Vector);
|
|
unsigned getRegisterBitWidth(bool Vector);
|
|
unsigned getMaxInterleaveFactor(unsigned VF);
|
|
};
|
|
|
|
} // end namespace llvm
|
|
|
|
#endif
|