[llvm] 0b39d2d - Revert "[NFC] Separate Peeling Properties into its own struct"
Nikita Popov via llvm-commits
llvm-commits at lists.llvm.org
Wed Jul 8 12:46:04 PDT 2020
Author: Nikita Popov
Date: 2020-07-08T21:43:32+02:00
New Revision: 0b39d2d75275b80994dac06b7ad05031cbd09393
URL: https://github.com/llvm/llvm-project/commit/0b39d2d75275b80994dac06b7ad05031cbd09393
DIFF: https://github.com/llvm/llvm-project/commit/0b39d2d75275b80994dac06b7ad05031cbd09393.diff
LOG: Revert "[NFC] Separate Peeling Properties into its own struct"
This reverts commit 0369dc98f958a1ca2ec05f1897f091129bb16e8a.
Many failing tests.
Added:
Modified:
llvm/include/llvm/Analysis/TargetTransformInfo.h
llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
llvm/include/llvm/CodeGen/BasicTTIImpl.h
llvm/include/llvm/Transforms/Utils/UnrollLoop.h
llvm/lib/Analysis/TargetTransformInfo.cpp
llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
llvm/lib/Target/ARM/ARMTargetTransformInfo.h
llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.cpp
llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.h
llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp
llvm/lib/Target/PowerPC/PPCTargetTransformInfo.h
llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.cpp
llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.h
llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp
llvm/lib/Transforms/Utils/LoopUnrollPeel.cpp
Removed:
################################################################################
diff --git a/llvm/include/llvm/Analysis/TargetTransformInfo.h b/llvm/include/llvm/Analysis/TargetTransformInfo.h
index b6698eefdb01..695b7d6061c0 100644
--- a/llvm/include/llvm/Analysis/TargetTransformInfo.h
+++ b/llvm/include/llvm/Analysis/TargetTransformInfo.h
@@ -450,6 +450,11 @@ class TargetTransformInfo {
/// transformation will select an unrolling factor based on the current cost
/// threshold and other factors.
unsigned Count;
+ /// A forced peeling factor (the number of bodied of the original loop
+ /// that should be peeled off before the loop body). When set to 0, the
+ /// unrolling transformation will select a peeling factor based on profile
+ /// information and other factors.
+ unsigned PeelCount;
/// Default unroll count for loops with run-time trip count.
unsigned DefaultUnrollRuntimeCount;
// Set the maximum unrolling factor. The unrolling factor may be selected
@@ -483,10 +488,19 @@ class TargetTransformInfo {
bool Force;
/// Allow using trip count upper bound to unroll loops.
bool UpperBound;
+ /// Allow peeling off loop iterations.
+ bool AllowPeeling;
+ /// Allow peeling off loop iterations for loop nests.
+ bool AllowLoopNestsPeeling;
/// Allow unrolling of all the iterations of the runtime loop remainder.
bool UnrollRemainder;
/// Allow unroll and jam. Used to enable unroll and jam for the target.
bool UnrollAndJam;
+ /// Allow peeling basing on profile. Uses to enable peeling off all
+ /// iterations basing on provided profile.
+ /// If the value is true the peeling cost model can decide to peel only
+ /// some iterations and in this case it will set this to false.
+ bool PeelProfiledIterations;
/// Threshold for unroll and jam, for inner loop size. The 'Threshold'
/// value above is used during unroll and jam for the outer loop size.
/// This value is used in the same manner to limit the size of the inner
@@ -520,28 +534,6 @@ class TargetTransformInfo {
/// intrinsic is supported.
bool emitGetActiveLaneMask() const;
- // Parameters that control the loop peeling transformation
- struct PeelingPreferences {
- /// A forced peeling factor (the number of bodied of the original loop
- /// that should be peeled off before the loop body). When set to 0, the
- /// a peeling factor based on profile information and other factors.
- unsigned PeelCount;
- /// Allow peeling off loop iterations.
- bool AllowPeeling;
- /// Allow peeling off loop iterations for loop nests.
- bool AllowLoopNestsPeeling;
- /// Allow peeling basing on profile. Uses to enable peeling off all
- /// iterations basing on provided profile.
- /// If the value is true the peeling cost model can decide to peel only
- /// some iterations and in this case it will set this to false.
- bool PeelProfiledIterations;
- };
-
- /// Get target-customized preferences for the generic loop peeling
- /// transformation. The caller will initialize \p PP with the current
- /// target-independent defaults with information from \p L and \p SE.
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- PeelingPreferences &PP) const;
/// @}
/// \name Scalar Target Information
@@ -1290,8 +1282,6 @@ class TargetTransformInfo::Concept {
virtual bool isLoweredToCall(const Function *F) = 0;
virtual void getUnrollingPreferences(Loop *L, ScalarEvolution &,
UnrollingPreferences &UP) = 0;
- virtual void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- PeelingPreferences &PP) = 0;
virtual bool isHardwareLoopProfitable(Loop *L, ScalarEvolution &SE,
AssumptionCache &AC,
TargetLibraryInfo *LibInfo,
@@ -1570,10 +1560,6 @@ class TargetTransformInfo::Model final : public TargetTransformInfo::Concept {
UnrollingPreferences &UP) override {
return Impl.getUnrollingPreferences(L, SE, UP);
}
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- PeelingPreferences &PP) override {
- return Impl.getPeelingPreferences(L, SE, PP);
- }
bool isHardwareLoopProfitable(Loop *L, ScalarEvolution &SE,
AssumptionCache &AC, TargetLibraryInfo *LibInfo,
HardwareLoopInfo &HWLoopInfo) override {
diff --git a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
index 0ce975d6d4b5..ca7106ab98aa 100644
--- a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
+++ b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
@@ -150,9 +150,6 @@ class TargetTransformInfoImplBase {
void getUnrollingPreferences(Loop *, ScalarEvolution &,
TTI::UnrollingPreferences &) {}
- void getPeelingPreferences(Loop *, ScalarEvolution &,
- TTI::PeelingPreferences &) {}
-
bool isLegalAddImmediate(int64_t Imm) { return false; }
bool isLegalICmpImmediate(int64_t Imm) { return false; }
diff --git a/llvm/include/llvm/CodeGen/BasicTTIImpl.h b/llvm/include/llvm/CodeGen/BasicTTIImpl.h
index f9d32eadd23e..c6a9a65ae6c1 100644
--- a/llvm/include/llvm/CodeGen/BasicTTIImpl.h
+++ b/llvm/include/llvm/CodeGen/BasicTTIImpl.h
@@ -451,14 +451,6 @@ class BasicTTIImplBase : public TargetTransformInfoImplCRTPBase<T> {
UP.BEInsns = 2;
}
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- PP.PeelCount = 0;
- PP.AllowPeeling = true;
- PP.AllowLoopNestsPeeling = false;
- PP.PeelProfiledIterations = true;
- }
-
bool isHardwareLoopProfitable(Loop *L, ScalarEvolution &SE,
AssumptionCache &AC,
TargetLibraryInfo *LibInfo,
diff --git a/llvm/include/llvm/Transforms/Utils/UnrollLoop.h b/llvm/include/llvm/Transforms/Utils/UnrollLoop.h
index bb3d02b95956..1970cefcefba 100644
--- a/llvm/include/llvm/Transforms/Utils/UnrollLoop.h
+++ b/llvm/include/llvm/Transforms/Utils/UnrollLoop.h
@@ -94,7 +94,6 @@ bool UnrollRuntimeLoopRemainder(
void computePeelCount(Loop *L, unsigned LoopSize,
TargetTransformInfo::UnrollingPreferences &UP,
- TargetTransformInfo::PeelingPreferences &PP,
unsigned &TripCount, ScalarEvolution &SE);
bool canPeel(Loop *L);
@@ -120,8 +119,6 @@ bool computeUnrollCount(Loop *L, const TargetTransformInfo &TTI,
unsigned MaxTripCount, bool MaxOrZero,
unsigned &TripMultiple, unsigned LoopSize,
TargetTransformInfo::UnrollingPreferences &UP,
- TargetTransformInfo::PeelingPreferences &PP,
-
bool &UseUpperBound);
void simplifyLoopAfterUnroll(Loop *L, bool SimplifyIVs, LoopInfo *LI,
@@ -136,13 +133,9 @@ TargetTransformInfo::UnrollingPreferences gatherUnrollingPreferences(
BlockFrequencyInfo *BFI, ProfileSummaryInfo *PSI, int OptLevel,
Optional<unsigned> UserThreshold, Optional<unsigned> UserCount,
Optional<bool> UserAllowPartial, Optional<bool> UserRuntime,
- Optional<bool> UserUpperBound, Optional<unsigned> UserFullUnrollMaxCount);
-
-TargetTransformInfo::PeelingPreferences
-gatherPeelingPreferences(Loop *L, ScalarEvolution &SE,
- const TargetTransformInfo &TTI,
- Optional<bool> UserAllowPeeling,
- Optional<bool> UserAllowProfileBasedPeeling);
+ Optional<bool> UserUpperBound, Optional<bool> UserAllowPeeling,
+ Optional<bool> UserAllowProfileBasedPeeling,
+ Optional<unsigned> UserFullUnrollMaxCount);
unsigned ApproximateLoopSize(const Loop *L, unsigned &NumCalls,
bool &NotDuplicatable, bool &Convergent,
diff --git a/llvm/lib/Analysis/TargetTransformInfo.cpp b/llvm/lib/Analysis/TargetTransformInfo.cpp
index 2f051e53790b..87c6f83938ed 100644
--- a/llvm/lib/Analysis/TargetTransformInfo.cpp
+++ b/llvm/lib/Analysis/TargetTransformInfo.cpp
@@ -327,11 +327,6 @@ void TargetTransformInfo::getUnrollingPreferences(
return TTIImpl->getUnrollingPreferences(L, SE, UP);
}
-void TargetTransformInfo::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- PeelingPreferences &PP) const {
- return TTIImpl->getPeelingPreferences(L, SE, PP);
-}
-
bool TargetTransformInfo::isLegalAddImmediate(int64_t Imm) const {
return TTIImpl->isLegalAddImmediate(Imm);
}
diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
index cf6de797727b..be0c51b83a25 100644
--- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
@@ -859,11 +859,6 @@ void AArch64TTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
getFalkorUnrollingPreferences(L, SE, UP);
}
-void AArch64TTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- BaseT::getPeelingPreferences(L, SE, PP);
-}
-
Value *AArch64TTIImpl::getOrCreateResultFromMemIntrinsic(IntrinsicInst *Inst,
Type *ExpectedType) {
switch (Inst->getIntrinsicID()) {
diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
index 094b04c95db4..27afb2e5a7d6 100644
--- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
+++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
@@ -153,9 +153,6 @@ class AArch64TTIImpl : public BasicTTIImplBase<AArch64TTIImpl> {
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP);
-
Value *getOrCreateResultFromMemIntrinsic(IntrinsicInst *Inst,
Type *ExpectedType);
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
index 46051ac14b59..24f079ffe929 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
@@ -236,10 +236,6 @@ void AMDGPUTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
}
}
-void AMDGPUTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- BaseT::getPeelingPreferences(L, SE, PP);
-}
unsigned GCNTTIImpl::getHardwareNumberOfRegisters(bool Vec) const {
// The concept of vector registers doesn't really exist. Some packed vector
// operations operate on the normal 32-bit registers.
@@ -994,11 +990,6 @@ void GCNTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
CommonTTI.getUnrollingPreferences(L, SE, UP);
}
-void GCNTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- CommonTTI.getPeelingPreferences(L, SE, PP);
-}
-
unsigned R600TTIImpl::getHardwareNumberOfRegisters(bool Vec) const {
return 4 * 128; // XXX - 4 channels. Should these count as vector instead?
}
@@ -1105,8 +1096,3 @@ void R600TTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP) {
CommonTTI.getUnrollingPreferences(L, SE, UP);
}
-
-void R600TTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- CommonTTI.getPeelingPreferences(L, SE, PP);
-}
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
index b913f5194e40..508ed061e935 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
+++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
@@ -61,9 +61,6 @@ class AMDGPUTTIImpl final : public BasicTTIImplBase<AMDGPUTTIImpl> {
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
-
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP);
};
class GCNTTIImpl final : public BasicTTIImplBase<GCNTTIImpl> {
@@ -144,9 +141,6 @@ class GCNTTIImpl final : public BasicTTIImplBase<GCNTTIImpl> {
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP);
-
TTI::PopcntSupportKind getPopcntSupport(unsigned TyWidth) {
assert(isPowerOf2_32(TyWidth) && "Ty width must be power of 2");
return TTI::PSK_FastHardware;
@@ -264,8 +258,6 @@ class R600TTIImpl final : public BasicTTIImplBase<R600TTIImpl> {
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP);
unsigned getHardwareNumberOfRegisters(bool Vec) const;
unsigned getNumberOfRegisters(bool Vec) const;
unsigned getRegisterBitWidth(bool Vector) const;
diff --git a/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp b/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
index 74b1331216a0..44dfb9e8c129 100644
--- a/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
+++ b/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
@@ -1582,11 +1582,6 @@ void ARMTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
UP.Force = true;
}
-void ARMTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- BaseT::getPeelingPreferences(L, SE, PP);
-}
-
bool ARMTTIImpl::useReductionIntrinsic(unsigned Opcode, Type *Ty,
TTI::ReductionFlags Flags) const {
return ST->hasMVEIntegerOps();
diff --git a/llvm/lib/Target/ARM/ARMTargetTransformInfo.h b/llvm/lib/Target/ARM/ARMTargetTransformInfo.h
index 537a546361ee..5d914227c968 100644
--- a/llvm/lib/Target/ARM/ARMTargetTransformInfo.h
+++ b/llvm/lib/Target/ARM/ARMTargetTransformInfo.h
@@ -251,8 +251,6 @@ class ARMTTIImpl : public BasicTTIImplBase<ARMTTIImpl> {
bool emitGetActiveLaneMask() const;
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP);
bool shouldBuildLookupTablesForConstant(Constant *C) const {
// In the ROPI and RWPI relocation models we can't have pointers to global
// variables or functions in constant data, so don't convert switches to
diff --git a/llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.cpp b/llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.cpp
index 80c8736cb74a..76df4e8e1931 100644
--- a/llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.cpp
+++ b/llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.cpp
@@ -78,17 +78,12 @@ HexagonTTIImpl::getPopcntSupport(unsigned IntTyWidthInBit) const {
void HexagonTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP) {
UP.Runtime = UP.Partial = true;
-}
-
-void HexagonTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- BaseT::getPeelingPreferences(L, SE, PP);
// Only try to peel innermost loops with small runtime trip counts.
if (L && L->empty() && canPeel(L) &&
SE.getSmallConstantTripCount(L) == 0 &&
SE.getSmallConstantMaxTripCount(L) > 0 &&
SE.getSmallConstantMaxTripCount(L) <= 5) {
- PP.PeelCount = 2;
+ UP.PeelCount = 2;
}
}
diff --git a/llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.h b/llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.h
index 5fe397486402..3365c5bf1cb1 100644
--- a/llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.h
+++ b/llvm/lib/Target/Hexagon/HexagonTargetTransformInfo.h
@@ -64,9 +64,6 @@ class HexagonTTIImpl : public BasicTTIImplBase<HexagonTTIImpl> {
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP);
-
/// Bias LSR towards creating post-increment opportunities.
bool shouldFavorPostInc() const;
diff --git a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
index 3873c73fb2e0..5c14d0f1a24d 100644
--- a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
+++ b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
@@ -155,8 +155,3 @@ void NVPTXTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
UP.Partial = UP.Runtime = true;
UP.PartialThreshold = UP.Threshold / 4;
}
-
-void NVPTXTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- BaseT::getPeelingPreferences(L, SE, PP);
-}
diff --git a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
index cb832031f1ad..88156f687284 100644
--- a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
+++ b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
@@ -95,10 +95,6 @@ class NVPTXTTIImpl : public BasicTTIImplBase<NVPTXTTIImpl> {
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
-
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP);
-
bool hasVolatileVariant(Instruction *I, unsigned AddrSpace) {
// Volatile loads/stores are only supported for shared and global address
// spaces, or for generic AS that maps to them.
diff --git a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp
index 53556ffc267d..f2c746a14299 100644
--- a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp
+++ b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp
@@ -568,10 +568,6 @@ void PPCTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
BaseT::getUnrollingPreferences(L, SE, UP);
}
-void PPCTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- BaseT::getPeelingPreferences(L, SE, PP);
-}
// This function returns true to allow using coldcc calling convention.
// Returning true results in coldcc being used for functions which are cold at
// all call sites when the callers of the functions are not calling any other
diff --git a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.h b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.h
index d998521084e1..b831789d3e6e 100644
--- a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.h
+++ b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.h
@@ -66,8 +66,6 @@ class PPCTTIImpl : public BasicTTIImplBase<PPCTTIImpl> {
TargetLibraryInfo *LibInfo);
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP);
bool isLSRCostLess(TargetTransformInfo::LSRCost &C1,
TargetTransformInfo::LSRCost &C2);
diff --git a/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.cpp b/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.cpp
index 864200e5f71c..36141426e27d 100644
--- a/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.cpp
+++ b/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.cpp
@@ -294,10 +294,6 @@ void SystemZTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
UP.Force = true;
}
-void SystemZTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP) {
- BaseT::getPeelingPreferences(L, SE, PP);
-}
bool SystemZTTIImpl::isLSRCostLess(TargetTransformInfo::LSRCost &C1,
TargetTransformInfo::LSRCost &C2) {
diff --git a/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.h b/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.h
index 7f8f7f6f923f..d20541774da1 100644
--- a/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.h
+++ b/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.h
@@ -50,9 +50,6 @@ class SystemZTTIImpl : public BasicTTIImplBase<SystemZTTIImpl> {
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
- void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
- TTI::PeelingPreferences &PP);
-
bool isLSRCostLess(TargetTransformInfo::LSRCost &C1,
TargetTransformInfo::LSRCost &C2);
/// @}
diff --git a/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp b/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
index 285cba6ee205..f0ece1faa5fd 100644
--- a/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
+++ b/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
@@ -158,8 +158,7 @@ static bool computeUnrollAndJamCount(
const SmallPtrSetImpl<const Value *> &EphValues,
OptimizationRemarkEmitter *ORE, unsigned OuterTripCount,
unsigned OuterTripMultiple, unsigned OuterLoopSize, unsigned InnerTripCount,
- unsigned InnerLoopSize, TargetTransformInfo::UnrollingPreferences &UP,
- TargetTransformInfo::PeelingPreferences &PP) {
+ unsigned InnerLoopSize, TargetTransformInfo::UnrollingPreferences &UP) {
// First up use computeUnrollCount from the loop unroller to get a count
// for unrolling the outer loop, plus any loops requiring explicit
// unrolling we leave to the unroller. This uses UP.Threshold /
@@ -169,8 +168,7 @@ static bool computeUnrollAndJamCount(
bool UseUpperBound = false;
bool ExplicitUnroll = computeUnrollCount(
L, TTI, DT, LI, SE, EphValues, ORE, OuterTripCount, MaxTripCount,
- /*MaxOrZero*/ false, OuterTripMultiple, OuterLoopSize, UP, PP,
- UseUpperBound);
+ /*MaxOrZero*/ false, OuterTripMultiple, OuterLoopSize, UP, UseUpperBound);
if (ExplicitUnroll || UseUpperBound) {
// If the user explicitly set the loop as unrolled, dont UnJ it. Leave it
// for the unroller instead.
@@ -284,9 +282,7 @@ tryToUnrollAndJamLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
OptimizationRemarkEmitter &ORE, int OptLevel) {
TargetTransformInfo::UnrollingPreferences UP =
gatherUnrollingPreferences(L, SE, TTI, nullptr, nullptr, OptLevel, None,
- None, None, None, None, None);
- TargetTransformInfo::PeelingPreferences PP =
- gatherPeelingPreferences(L, SE, TTI, None, None);
+ None, None, None, None, None, None, None);
if (AllowUnrollAndJam.getNumOccurrences() > 0)
UP.UnrollAndJam = AllowUnrollAndJam;
if (UnrollAndJamThreshold.getNumOccurrences() > 0)
@@ -371,7 +367,7 @@ tryToUnrollAndJamLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
// Decide if, and by how much, to unroll
bool IsCountSetExplicitly = computeUnrollAndJamCount(
L, SubLoop, TTI, DT, LI, SE, EphValues, &ORE, OuterTripCount,
- OuterTripMultiple, OuterLoopSize, InnerTripCount, InnerLoopSize, UP, PP);
+ OuterTripMultiple, OuterLoopSize, InnerTripCount, InnerLoopSize, UP);
if (UP.Count <= 1)
return LoopUnrollResult::Unmodified;
// Unroll factor (Count) must be less or equal to TripCount.
diff --git a/llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp b/llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp
index 88845cde8d4f..ec56610e41e5 100644
--- a/llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp
+++ b/llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp
@@ -193,7 +193,9 @@ TargetTransformInfo::UnrollingPreferences llvm::gatherUnrollingPreferences(
BlockFrequencyInfo *BFI, ProfileSummaryInfo *PSI, int OptLevel,
Optional<unsigned> UserThreshold, Optional<unsigned> UserCount,
Optional<bool> UserAllowPartial, Optional<bool> UserRuntime,
- Optional<bool> UserUpperBound, Optional<unsigned> UserFullUnrollMaxCount) {
+ Optional<bool> UserUpperBound, Optional<bool> UserAllowPeeling,
+ Optional<bool> UserAllowProfileBasedPeeling,
+ Optional<unsigned> UserFullUnrollMaxCount) {
TargetTransformInfo::UnrollingPreferences UP;
// Set up the defaults
@@ -204,6 +206,7 @@ TargetTransformInfo::UnrollingPreferences llvm::gatherUnrollingPreferences(
UP.PartialThreshold = 150;
UP.PartialOptSizeThreshold = 0;
UP.Count = 0;
+ UP.PeelCount = 0;
UP.DefaultUnrollRuntimeCount = 8;
UP.MaxCount = std::numeric_limits<unsigned>::max();
UP.FullUnrollMaxCount = std::numeric_limits<unsigned>::max();
@@ -215,7 +218,10 @@ TargetTransformInfo::UnrollingPreferences llvm::gatherUnrollingPreferences(
UP.AllowExpensiveTripCount = false;
UP.Force = false;
UP.UpperBound = false;
+ UP.AllowPeeling = true;
+ UP.AllowLoopNestsPeeling = false;
UP.UnrollAndJam = false;
+ UP.PeelProfiledIterations = true;
UP.UnrollAndJamInnerLoopThreshold = 60;
UP.MaxIterationsCountToAnalyze = UnrollMaxIterationsCountToAnalyze;
@@ -243,6 +249,8 @@ TargetTransformInfo::UnrollingPreferences llvm::gatherUnrollingPreferences(
UP.MaxCount = UnrollMaxCount;
if (UnrollFullMaxCount.getNumOccurrences() > 0)
UP.FullUnrollMaxCount = UnrollFullMaxCount;
+ if (UnrollPeelCount.getNumOccurrences() > 0)
+ UP.PeelCount = UnrollPeelCount;
if (UnrollAllowPartial.getNumOccurrences() > 0)
UP.Partial = UnrollAllowPartial;
if (UnrollAllowRemainder.getNumOccurrences() > 0)
@@ -251,6 +259,10 @@ TargetTransformInfo::UnrollingPreferences llvm::gatherUnrollingPreferences(
UP.Runtime = UnrollRuntime;
if (UnrollMaxUpperBound == 0)
UP.UpperBound = false;
+ if (UnrollAllowPeeling.getNumOccurrences() > 0)
+ UP.AllowPeeling = UnrollAllowPeeling;
+ if (UnrollAllowLoopNestsPeeling.getNumOccurrences() > 0)
+ UP.AllowLoopNestsPeeling = UnrollAllowLoopNestsPeeling;
if (UnrollUnrollRemainder.getNumOccurrences() > 0)
UP.UnrollRemainder = UnrollUnrollRemainder;
if (UnrollMaxIterationsCountToAnalyze.getNumOccurrences() > 0)
@@ -269,39 +281,16 @@ TargetTransformInfo::UnrollingPreferences llvm::gatherUnrollingPreferences(
UP.Runtime = *UserRuntime;
if (UserUpperBound.hasValue())
UP.UpperBound = *UserUpperBound;
+ if (UserAllowPeeling.hasValue())
+ UP.AllowPeeling = *UserAllowPeeling;
+ if (UserAllowProfileBasedPeeling.hasValue())
+ UP.PeelProfiledIterations = *UserAllowProfileBasedPeeling;
if (UserFullUnrollMaxCount.hasValue())
UP.FullUnrollMaxCount = *UserFullUnrollMaxCount;
return UP;
}
-TargetTransformInfo::PeelingPreferences
-llvm::gatherPeelingPreferences(Loop *L, ScalarEvolution &SE,
- const TargetTransformInfo &TTI,
- Optional<bool> UserAllowPeeling,
- Optional<bool> UserAllowProfileBasedPeeling) {
- TargetTransformInfo::PeelingPreferences PP;
-
- // Get Target Specifc Values
- TTI.getPeelingPreferences(L, SE, PP);
-
- // User Specified Values using cl::opt
- if (UnrollPeelCount.getNumOccurrences() > 0)
- PP.PeelCount = UnrollPeelCount;
- if (UnrollAllowPeeling.getNumOccurrences() > 0)
- PP.AllowPeeling = UnrollAllowPeeling;
- if (UnrollAllowLoopNestsPeeling.getNumOccurrences() > 0)
- PP.AllowLoopNestsPeeling = UnrollAllowLoopNestsPeeling;
-
- // User Specifed values provided by argument
- if (UserAllowPeeling.hasValue())
- PP.AllowPeeling = *UserAllowPeeling;
- if (UserAllowProfileBasedPeeling.hasValue())
- PP.PeelProfiledIterations = *UserAllowProfileBasedPeeling;
-
- return PP;
-}
-
namespace {
/// A struct to densely store the state of an instruction after unrolling at
@@ -772,8 +761,7 @@ bool llvm::computeUnrollCount(
ScalarEvolution &SE, const SmallPtrSetImpl<const Value *> &EphValues,
OptimizationRemarkEmitter *ORE, unsigned &TripCount, unsigned MaxTripCount,
bool MaxOrZero, unsigned &TripMultiple, unsigned LoopSize,
- TargetTransformInfo::UnrollingPreferences &UP,
- TargetTransformInfo::PeelingPreferences &PP, bool &UseUpperBound) {
+ TargetTransformInfo::UnrollingPreferences &UP, bool &UseUpperBound) {
// Check for explicit Count.
// 1st priority is unroll count set by "unroll-count" option.
@@ -875,8 +863,8 @@ bool llvm::computeUnrollCount(
}
// 4th priority is loop peeling.
- computePeelCount(L, LoopSize, UP, PP, TripCount, SE);
- if (PP.PeelCount) {
+ computePeelCount(L, LoopSize, UP, TripCount, SE);
+ if (UP.PeelCount) {
UP.Runtime = false;
UP.Count = 1;
return ExplicitUnroll;
@@ -1079,9 +1067,8 @@ static LoopUnrollResult tryToUnrollLoop(
TargetTransformInfo::UnrollingPreferences UP = gatherUnrollingPreferences(
L, SE, TTI, BFI, PSI, OptLevel, ProvidedThreshold, ProvidedCount,
ProvidedAllowPartial, ProvidedRuntime, ProvidedUpperBound,
+ ProvidedAllowPeeling, ProvidedAllowProfileBasedPeeling,
ProvidedFullUnrollMaxCount);
- TargetTransformInfo::PeelingPreferences PP = gatherPeelingPreferences(
- L, SE, TTI, ProvidedAllowPeeling, ProvidedAllowProfileBasedPeeling);
// Exit early if unrolling is disabled. For OptForSize, we pick the loop size
// as threshold later on.
@@ -1155,7 +1142,7 @@ static LoopUnrollResult tryToUnrollLoop(
bool UseUpperBound = false;
bool IsCountSetExplicitly = computeUnrollCount(
L, TTI, DT, LI, SE, EphValues, &ORE, TripCount, MaxTripCount, MaxOrZero,
- TripMultiple, LoopSize, UP, PP, UseUpperBound);
+ TripMultiple, LoopSize, UP, UseUpperBound);
if (!UP.Count)
return LoopUnrollResult::Unmodified;
// Unroll factor (Count) must be less or equal to TripCount.
@@ -1170,7 +1157,7 @@ static LoopUnrollResult tryToUnrollLoop(
LoopUnrollResult UnrollResult = UnrollLoop(
L,
{UP.Count, TripCount, UP.Force, UP.Runtime, UP.AllowExpensiveTripCount,
- UseUpperBound, MaxOrZero, TripMultiple, PP.PeelCount, UP.UnrollRemainder,
+ UseUpperBound, MaxOrZero, TripMultiple, UP.PeelCount, UP.UnrollRemainder,
ForgetAllSCEV},
LI, &SE, &DT, &AC, &TTI, &ORE, PreserveLCSSA, &RemainderLoop);
if (UnrollResult == LoopUnrollResult::Unmodified)
@@ -1202,7 +1189,7 @@ static LoopUnrollResult tryToUnrollLoop(
// If the loop was peeled, we already "used up" the profile information
// we had, so we don't want to unroll or peel again.
if (UnrollResult != LoopUnrollResult::FullyUnrolled &&
- (IsCountSetExplicitly || (PP.PeelProfiledIterations && PP.PeelCount)))
+ (IsCountSetExplicitly || (UP.PeelProfiledIterations && UP.PeelCount)))
L->setLoopAlreadyUnrolled();
return UnrollResult;
diff --git a/llvm/lib/Transforms/Utils/LoopUnrollPeel.cpp b/llvm/lib/Transforms/Utils/LoopUnrollPeel.cpp
index c653aacbee6c..43dfaf3e50dc 100644
--- a/llvm/lib/Transforms/Utils/LoopUnrollPeel.cpp
+++ b/llvm/lib/Transforms/Utils/LoopUnrollPeel.cpp
@@ -279,20 +279,19 @@ static unsigned countToEliminateCompares(Loop &L, unsigned MaxPeelCount,
// Return the number of iterations we want to peel off.
void llvm::computePeelCount(Loop *L, unsigned LoopSize,
TargetTransformInfo::UnrollingPreferences &UP,
- TargetTransformInfo::PeelingPreferences &PP,
unsigned &TripCount, ScalarEvolution &SE) {
assert(LoopSize > 0 && "Zero loop size is not allowed!");
- // Save the PP.PeelCount value set by the target in
- // TTI.getPeelingPreferences or by the flag -unroll-peel-count.
- unsigned TargetPeelCount = PP.PeelCount;
- PP.PeelCount = 0;
+ // Save the UP.PeelCount value set by the target in
+ // TTI.getUnrollingPreferences or by the flag -unroll-peel-count.
+ unsigned TargetPeelCount = UP.PeelCount;
+ UP.PeelCount = 0;
if (!canPeel(L))
return;
// Only try to peel innermost loops by default.
// The constraint can be relaxed by the target in TTI.getUnrollingPreferences
// or by the flag -unroll-allow-loop-nests-peeling.
- if (!PP.AllowLoopNestsPeeling && !L->empty())
+ if (!UP.AllowLoopNestsPeeling && !L->empty())
return;
// If the user provided a peel count, use that.
@@ -300,13 +299,13 @@ void llvm::computePeelCount(Loop *L, unsigned LoopSize,
if (UserPeelCount) {
LLVM_DEBUG(dbgs() << "Force-peeling first " << UnrollForcePeelCount
<< " iterations.\n");
- PP.PeelCount = UnrollForcePeelCount;
- PP.PeelProfiledIterations = true;
+ UP.PeelCount = UnrollForcePeelCount;
+ UP.PeelProfiledIterations = true;
return;
}
// Skip peeling if it's disabled.
- if (!PP.AllowPeeling)
+ if (!UP.AllowPeeling)
return;
unsigned AlreadyPeeled = 0;
@@ -355,8 +354,8 @@ void llvm::computePeelCount(Loop *L, unsigned LoopSize,
LLVM_DEBUG(dbgs() << "Peel " << DesiredPeelCount
<< " iteration(s) to turn"
<< " some Phis into invariants.\n");
- PP.PeelCount = DesiredPeelCount;
- PP.PeelProfiledIterations = false;
+ UP.PeelCount = DesiredPeelCount;
+ UP.PeelProfiledIterations = false;
return;
}
}
@@ -368,7 +367,7 @@ void llvm::computePeelCount(Loop *L, unsigned LoopSize,
return;
// Do not apply profile base peeling if it is disabled.
- if (!PP.PeelProfiledIterations)
+ if (!UP.PeelProfiledIterations)
return;
// If we don't know the trip count, but have reason to believe the average
// trip count is low, peeling should be beneficial, since we will usually
@@ -388,7 +387,7 @@ void llvm::computePeelCount(Loop *L, unsigned LoopSize,
(LoopSize * (*PeelCount + 1) <= UP.Threshold)) {
LLVM_DEBUG(dbgs() << "Peeling first " << *PeelCount
<< " iterations.\n");
- PP.PeelCount = *PeelCount;
+ UP.PeelCount = *PeelCount;
return;
}
LLVM_DEBUG(dbgs() << "Requested peel count: " << *PeelCount << "\n");
More information about the llvm-commits
mailing list