llvm-project/llvm/lib/IR/PseudoProbe.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

100 lines
3.7 KiB
C++
Raw Normal View History

[CSSPGO] Consume pseudo-probe-based AutoFDO profile This change enables pseudo-probe-based sample counts to be consumed by the sample profile loader under the regular `-fprofile-sample-use` switch with minimal adjustments to the existing sample file formats. After the counts are imported, a probe helper, aka, a `PseudoProbeManager` object, is automatically launched to verify the CFG checksum of every function in the current compilation against the corresponding checksum from the profile. Mismatched checksums will cause a function profile to be slipped. A `SampleProfileProber` pass is scheduled before any of the `SampleProfileLoader` instances so that the CFG checksums as well as probe mappings are available during the profile loading time. The `PseudoProbeManager` object is set up right after the profile reading is done. In the future a CFG-based fuzzy matching could be done in `PseudoProbeManager`. Samples will be applied only to pseudo probe instructions as well as probed callsites once the checksum verification goes through. Those instructions are processed in the same way that regular instructions would be processed in the line-number-based scenario. In other words, a function is processed in a regular way as if it was reduced to just containing pseudo probes (block probes and callsites). **Adjustment to profile format ** A CFG checksum field is being added to the existing AutoFDO profile formats. So far only the text format and the extended binary format are supported. For the text format, a new line like ``` !CFGChecksum: 12345 ``` is added to the end of the body sample lines. For the extended binary profile format, we introduce a metadata section to store the checksum map from function names to their CFG checksums. Differential Revision: https://reviews.llvm.org/D92347
2020-12-17 04:54:50 +08:00
//===- PseudoProbe.cpp - Pseudo Probe Helpers -----------------------------===//
//
// 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 file implements the helpers to manipulate pseudo probe IR intrinsic
// calls.
//
//===----------------------------------------------------------------------===//
#include "llvm/IR/PseudoProbe.h"
#include "llvm/IR/DebugInfoMetadata.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Instruction.h"
using namespace llvm;
namespace llvm {
Optional<PseudoProbe> extractProbeFromDiscriminator(const Instruction &Inst) {
assert(isa<CallBase>(&Inst) && !isa<IntrinsicInst>(&Inst) &&
"Only call instructions should have pseudo probe encodes as their "
"Dwarf discriminators");
if (const DebugLoc &DLoc = Inst.getDebugLoc()) {
const DILocation *DIL = DLoc;
auto Discriminator = DIL->getDiscriminator();
if (DILocation::isPseudoProbeDiscriminator(Discriminator)) {
PseudoProbe Probe;
Probe.Id =
PseudoProbeDwarfDiscriminator::extractProbeIndex(Discriminator);
Probe.Type =
PseudoProbeDwarfDiscriminator::extractProbeType(Discriminator);
Probe.Attr =
PseudoProbeDwarfDiscriminator::extractProbeAttributes(Discriminator);
[CSSPGO] Introducing distribution factor for pseudo probe. Sample re-annotation is required in LTO time to achieve a reasonable post-inline profile quality. However, we have seen that such LTO-time re-annotation degrades profile quality. This is mainly caused by preLTO code duplication that is done by passes such as loop unrolling, jump threading, indirect call promotion etc, where samples corresponding to a source location are aggregated multiple times due to the duplicates. In this change we are introducing a concept of distribution factor for pseudo probes so that samples can be distributed for duplicated probes scaled by a factor. We hope that optimizations duplicating code well-maintain the branch frequency information (BFI) based on which probe distribution factors are calculated. Distribution factors are updated at the end of preLTO pipeline to reflect an estimated portion of the real execution count. This change also introduces a pseudo probe verifier that can be run after each IR passes to detect duplicated pseudo probes. A saturated distribution factor stands for 1.0. A pesudo probe will carry a factor with the value ranged from 0.0 to 1.0. A 64-bit integral distribution factor field that represents [0.0, 1.0] is associated to each block probe. Unfortunately this cannot be done for callsite probes due to the size limitation of a 32-bit Dwarf discriminator. A 7-bit distribution factor is used instead. Changes are also needed to the sample profile inliner to deal with prorated callsite counts. Call sites duplicated by PreLTO passes, when later on inlined in LTO time, should have the callees’s probe prorated based on the Prelink-computed distribution factors. The distribution factors should also be taken into account when computing hotness for inline candidates. Also, Indirect call promotion results in multiple callisites. The original samples should be distributed across them. This is fixed by adjusting the callisites' distribution factors. Reviewed By: wmi Differential Revision: https://reviews.llvm.org/D93264
2020-12-12 04:18:31 +08:00
Probe.Factor =
PseudoProbeDwarfDiscriminator::extractProbeFactor(Discriminator) /
(float)PseudoProbeDwarfDiscriminator::FullDistributionFactor;
[CSSPGO] Consume pseudo-probe-based AutoFDO profile This change enables pseudo-probe-based sample counts to be consumed by the sample profile loader under the regular `-fprofile-sample-use` switch with minimal adjustments to the existing sample file formats. After the counts are imported, a probe helper, aka, a `PseudoProbeManager` object, is automatically launched to verify the CFG checksum of every function in the current compilation against the corresponding checksum from the profile. Mismatched checksums will cause a function profile to be slipped. A `SampleProfileProber` pass is scheduled before any of the `SampleProfileLoader` instances so that the CFG checksums as well as probe mappings are available during the profile loading time. The `PseudoProbeManager` object is set up right after the profile reading is done. In the future a CFG-based fuzzy matching could be done in `PseudoProbeManager`. Samples will be applied only to pseudo probe instructions as well as probed callsites once the checksum verification goes through. Those instructions are processed in the same way that regular instructions would be processed in the line-number-based scenario. In other words, a function is processed in a regular way as if it was reduced to just containing pseudo probes (block probes and callsites). **Adjustment to profile format ** A CFG checksum field is being added to the existing AutoFDO profile formats. So far only the text format and the extended binary format are supported. For the text format, a new line like ``` !CFGChecksum: 12345 ``` is added to the end of the body sample lines. For the extended binary profile format, we introduce a metadata section to store the checksum map from function names to their CFG checksums. Differential Revision: https://reviews.llvm.org/D92347
2020-12-17 04:54:50 +08:00
return Probe;
}
}
return None;
}
Optional<PseudoProbe> extractProbe(const Instruction &Inst) {
if (const auto *II = dyn_cast<PseudoProbeInst>(&Inst)) {
PseudoProbe Probe;
Probe.Id = II->getIndex()->getZExtValue();
Probe.Type = (uint32_t)PseudoProbeType::Block;
Probe.Attr = II->getAttributes()->getZExtValue();
[CSSPGO] Introducing distribution factor for pseudo probe. Sample re-annotation is required in LTO time to achieve a reasonable post-inline profile quality. However, we have seen that such LTO-time re-annotation degrades profile quality. This is mainly caused by preLTO code duplication that is done by passes such as loop unrolling, jump threading, indirect call promotion etc, where samples corresponding to a source location are aggregated multiple times due to the duplicates. In this change we are introducing a concept of distribution factor for pseudo probes so that samples can be distributed for duplicated probes scaled by a factor. We hope that optimizations duplicating code well-maintain the branch frequency information (BFI) based on which probe distribution factors are calculated. Distribution factors are updated at the end of preLTO pipeline to reflect an estimated portion of the real execution count. This change also introduces a pseudo probe verifier that can be run after each IR passes to detect duplicated pseudo probes. A saturated distribution factor stands for 1.0. A pesudo probe will carry a factor with the value ranged from 0.0 to 1.0. A 64-bit integral distribution factor field that represents [0.0, 1.0] is associated to each block probe. Unfortunately this cannot be done for callsite probes due to the size limitation of a 32-bit Dwarf discriminator. A 7-bit distribution factor is used instead. Changes are also needed to the sample profile inliner to deal with prorated callsite counts. Call sites duplicated by PreLTO passes, when later on inlined in LTO time, should have the callees’s probe prorated based on the Prelink-computed distribution factors. The distribution factors should also be taken into account when computing hotness for inline candidates. Also, Indirect call promotion results in multiple callisites. The original samples should be distributed across them. This is fixed by adjusting the callisites' distribution factors. Reviewed By: wmi Differential Revision: https://reviews.llvm.org/D93264
2020-12-12 04:18:31 +08:00
Probe.Factor = II->getFactor()->getZExtValue() /
(float)PseudoProbeFullDistributionFactor;
[CSSPGO] Consume pseudo-probe-based AutoFDO profile This change enables pseudo-probe-based sample counts to be consumed by the sample profile loader under the regular `-fprofile-sample-use` switch with minimal adjustments to the existing sample file formats. After the counts are imported, a probe helper, aka, a `PseudoProbeManager` object, is automatically launched to verify the CFG checksum of every function in the current compilation against the corresponding checksum from the profile. Mismatched checksums will cause a function profile to be slipped. A `SampleProfileProber` pass is scheduled before any of the `SampleProfileLoader` instances so that the CFG checksums as well as probe mappings are available during the profile loading time. The `PseudoProbeManager` object is set up right after the profile reading is done. In the future a CFG-based fuzzy matching could be done in `PseudoProbeManager`. Samples will be applied only to pseudo probe instructions as well as probed callsites once the checksum verification goes through. Those instructions are processed in the same way that regular instructions would be processed in the line-number-based scenario. In other words, a function is processed in a regular way as if it was reduced to just containing pseudo probes (block probes and callsites). **Adjustment to profile format ** A CFG checksum field is being added to the existing AutoFDO profile formats. So far only the text format and the extended binary format are supported. For the text format, a new line like ``` !CFGChecksum: 12345 ``` is added to the end of the body sample lines. For the extended binary profile format, we introduce a metadata section to store the checksum map from function names to their CFG checksums. Differential Revision: https://reviews.llvm.org/D92347
2020-12-17 04:54:50 +08:00
return Probe;
}
if (isa<CallBase>(&Inst) && !isa<IntrinsicInst>(&Inst))
return extractProbeFromDiscriminator(Inst);
return None;
}
[CSSPGO] Introducing distribution factor for pseudo probe. Sample re-annotation is required in LTO time to achieve a reasonable post-inline profile quality. However, we have seen that such LTO-time re-annotation degrades profile quality. This is mainly caused by preLTO code duplication that is done by passes such as loop unrolling, jump threading, indirect call promotion etc, where samples corresponding to a source location are aggregated multiple times due to the duplicates. In this change we are introducing a concept of distribution factor for pseudo probes so that samples can be distributed for duplicated probes scaled by a factor. We hope that optimizations duplicating code well-maintain the branch frequency information (BFI) based on which probe distribution factors are calculated. Distribution factors are updated at the end of preLTO pipeline to reflect an estimated portion of the real execution count. This change also introduces a pseudo probe verifier that can be run after each IR passes to detect duplicated pseudo probes. A saturated distribution factor stands for 1.0. A pesudo probe will carry a factor with the value ranged from 0.0 to 1.0. A 64-bit integral distribution factor field that represents [0.0, 1.0] is associated to each block probe. Unfortunately this cannot be done for callsite probes due to the size limitation of a 32-bit Dwarf discriminator. A 7-bit distribution factor is used instead. Changes are also needed to the sample profile inliner to deal with prorated callsite counts. Call sites duplicated by PreLTO passes, when later on inlined in LTO time, should have the callees’s probe prorated based on the Prelink-computed distribution factors. The distribution factors should also be taken into account when computing hotness for inline candidates. Also, Indirect call promotion results in multiple callisites. The original samples should be distributed across them. This is fixed by adjusting the callisites' distribution factors. Reviewed By: wmi Differential Revision: https://reviews.llvm.org/D93264
2020-12-12 04:18:31 +08:00
void setProbeDistributionFactor(Instruction &Inst, float Factor) {
assert(Factor >= 0 && Factor <= 1 &&
"Distribution factor must be in [0, 1.0]");
if (auto *II = dyn_cast<PseudoProbeInst>(&Inst)) {
IRBuilder<> Builder(&Inst);
uint64_t IntFactor = PseudoProbeFullDistributionFactor;
if (Factor < 1)
IntFactor *= Factor;
auto OrigFactor = II->getFactor()->getZExtValue();
if (IntFactor != OrigFactor)
II->replaceUsesOfWith(II->getFactor(), Builder.getInt64(IntFactor));
} else if (isa<CallBase>(&Inst) && !isa<IntrinsicInst>(&Inst)) {
if (const DebugLoc &DLoc = Inst.getDebugLoc()) {
const DILocation *DIL = DLoc;
auto Discriminator = DIL->getDiscriminator();
if (DILocation::isPseudoProbeDiscriminator(Discriminator)) {
auto Index =
PseudoProbeDwarfDiscriminator::extractProbeIndex(Discriminator);
auto Type =
PseudoProbeDwarfDiscriminator::extractProbeType(Discriminator);
auto Attr = PseudoProbeDwarfDiscriminator::extractProbeAttributes(
Discriminator);
// Round small factors to 0 to avoid over-counting.
uint32_t IntFactor =
PseudoProbeDwarfDiscriminator::FullDistributionFactor;
if (Factor < 1)
IntFactor *= Factor;
uint32_t V = PseudoProbeDwarfDiscriminator::packProbeData(
Index, Type, Attr, IntFactor);
DIL = DIL->cloneWithDiscriminator(V);
Inst.setDebugLoc(DIL);
}
}
}
}
[CSSPGO] Consume pseudo-probe-based AutoFDO profile This change enables pseudo-probe-based sample counts to be consumed by the sample profile loader under the regular `-fprofile-sample-use` switch with minimal adjustments to the existing sample file formats. After the counts are imported, a probe helper, aka, a `PseudoProbeManager` object, is automatically launched to verify the CFG checksum of every function in the current compilation against the corresponding checksum from the profile. Mismatched checksums will cause a function profile to be slipped. A `SampleProfileProber` pass is scheduled before any of the `SampleProfileLoader` instances so that the CFG checksums as well as probe mappings are available during the profile loading time. The `PseudoProbeManager` object is set up right after the profile reading is done. In the future a CFG-based fuzzy matching could be done in `PseudoProbeManager`. Samples will be applied only to pseudo probe instructions as well as probed callsites once the checksum verification goes through. Those instructions are processed in the same way that regular instructions would be processed in the line-number-based scenario. In other words, a function is processed in a regular way as if it was reduced to just containing pseudo probes (block probes and callsites). **Adjustment to profile format ** A CFG checksum field is being added to the existing AutoFDO profile formats. So far only the text format and the extended binary format are supported. For the text format, a new line like ``` !CFGChecksum: 12345 ``` is added to the end of the body sample lines. For the extended binary profile format, we introduce a metadata section to store the checksum map from function names to their CFG checksums. Differential Revision: https://reviews.llvm.org/D92347
2020-12-17 04:54:50 +08:00
} // namespace llvm