We merge cold context by default to save profile size. However trimming cold context after merging doesn't save size much, so default to off to reflect how it's commonly used.
Differential Revision: https://reviews.llvm.org/D109166
This change improves the warning for truncated context by: 1) deduplicate them as one call without probe can appear in many different context leading to duplicated warnings , 2) rephrase the message to make it easier to understand. The term "untracked frame" can be confusing.
Differential Revision: https://reviews.llvm.org/D109115
Adding the compiler support of MD5 CS profile based on pervious context split work D107299. A MD5 CS profile is about 40% smaller than the string-based extbinary profile. As a result, the compilation is 15% faster.
There are a few conversion from real names to md5 names that have been made on the sample loader and context tracker side to get it work.
Reviewed By: wenlei, wmi
Differential Revision: https://reviews.llvm.org/D108342
This change aims at supporting LBR only sample perf script which is used for regular(Non-CS) profile generation. A LBR perf script includes a batch of LBR sample which starts with a frame pointer and a group of 32 LBR entries is followed. The FROM/TO LBR pair and the range between two consecutive entries (the former entry's TO and the latter entry's FROM) will be used to infer function profile info.
An example of LBR perf script(created by `perf script -F ip,brstack -i perf.data`)
```
40062f 0x40062f/0x4005b0/P/-/-/9 0x400645/0x4005ff/P/-/-/1 0x400637/0x400645/P/-/-/1 ...
4005d7 0x4005d7/0x4005e5/P/-/-/8 0x40062f/0x4005b0/P/-/-/6 0x400645/0x4005ff/P/-/-/1 ...
...
```
For implementation:
- Extended a new child class `LBRPerfReader` for the sample parsing, reused all the functionalities in `extractLBRStack` except for an extension to parsing leading instruction pointer.
- `HybridSample` is reused(just leave the call stack empty) and the parsed samples is still aggregated in `AggregatedSamples`. After that, range samples, branch sample, address samples are computed and recorded.
- Reused `ContextSampleCounterMap` to store the raw profile, since it's no need to aggregation by context, here it just registered one sample counter with a fake context key.
- Unified to use `show-raw-profile` instead of `show-unwinder-output` to dump the intermediate raw profile, see the comments of the format of the raw profile. For CS profile, it remains to output the unwinder output.
Profile generation part will come soon.
Differential Revision: https://reviews.llvm.org/D108153
Currently context strings contain a lot of duplicated function names and that significantly increase the profile size. This change split the context into a series of {name, offset, discriminator} tuples so function names used in the context can be replaced by the index into the name table and that significantly reduce the size consumed by context.
A follow-up improvement made in the compiler and profiling tools is to avoid reconstructing full context strings which is time- and memory- consuming. Instead a context vector of `StringRef` is adopted to represent the full context in all scenarios. As a result, the previous prevalent profile map which was implemented as a `StringRef` is now engineered as an unordered map keyed by `SampleContext`. `SampleContext` is reshaped to using an `ArrayRef` to represent a full context for CS profile. For non-CS profile, it falls back to use `StringRef` to represent a contextless function name. Both the `ArrayRef` and `StringRef` objects are underpinned by real array and string objects that are stored in producer buffers. For compiler, they are maintained by the sample reader. For llvm-profgen, they are maintained in `ProfiledBinary` and `ProfileGenerator`. Full context strings can be generated only in those cases of debugging and printing.
When it comes to profile format, nothing has changed to the text format, though internally CS context is implemented as a vector. Extbinary format is only changed for CS profile, with an additional `SecCSNameTable` section which stores all full contexts logically in the form of `vector<int>`, which each element as an offset points to `SecNameTable`. All occurrences of contexts elsewhere are redirected to using the offset of `SecCSNameTable`.
Testing
This is no-diff change in terms of code quality and profile content (for text profile).
For our internal large service (aka ads), the profile generation is cut to half, with a 20x smaller string-based extbinary format generated.
The compile time of ads is dropped by 25%.
Differential Revision: https://reviews.llvm.org/D107299
The change adds a switch to allow sample loader to use global pre-inliner's decision instead. The pre-inliner in llvm-profgen makes inline decision globally based on whole program profile and function byte size as cost proxy.
Since pre-inliner also adjusts/merges context profile based on its inline decision, honoring its inline decision in sample loader would lead to better post-inline profile quality especially for thinlto where cross module profile merging isn't possible without pre-inliner.
Minor fix in profile reader is also included. When pre-inliner is use, we now also turn off the default merging and trimming logic unless it's explicitly asked.
Differential Revision: https://reviews.llvm.org/D108677
This is a follow up diff for BinarySizeContextTracker to track zero size for fully optimized inlinee. When an inlinee is fully optimized away, we won't be able to get its size through symbolizing instructions, hence we will treat the corresponding context size as unknown. However by traversing the inlined probe forest, we know what're original inlinees regardless of optimization. If a context show up in inlined probes, but not during symbolization, we know that it's fully optimized away hence its size is zero instead of unknown. It should provide more accurate size cost estimation for pre-inliner to make better inline decisions in llvm-profgen.
Differential Revision: https://reviews.llvm.org/D108350
This change enables llvm-profgen to use accurate context-sensitive post-optimization function byte size as a cost proxy to drive global preinline decisions.
To do this, BinarySizeContextTracker is introduced to track function byte size under different inline context during disassembling. In preinliner, we can not query context byte size under switch `context-cost-for-preinliner`. The tracker uses a reverse trie to keep size of functions under different context (callee as parent, caller as child), and it can give best/longest possible matching context size for given input context.
The new size cost is off by default. There're a few TODOs that needs to addressed: 1) avoid dangling string from `Offset2LocStackMap`, which will be addressed in split context work; 2) using inlinee's entry probe to make sure we have correct zero size for inlinee that's completely optimized away after inlining. Some tuning is also needed.
Differential Revision: https://reviews.llvm.org/D108180
Change to use unique pointer of profiled binary to unblock asan.
At same time, I realized we can decouple to move the profiled binary loading out of PerfReader, so I made some other related refactors.
Reviewed By: hoy
Differential Revision: https://reviews.llvm.org/D108254
As we decided to support only one binary each time, this patch cleans up the related code dealing with multiple binaries. We can use `llvm-profdata` to merge profile from multiple binaries.
Reviewed By: hoy, wenlei
Differential Revision: https://reviews.llvm.org/D108002
Currently we use a centralized string map(StringMap<FunctionSamples> ProfileMap) to store the profile while populating the sample, which might cause the memory usage bottleneck. I saw in an extreme case, there are thousands of samples whose context stack depth is >= 100. The memory consumption can be greater than 100GB.
As here the context is used for inlining, we can assume we won't have so many of inlinees keeping inlined at the same root function, so this change tried to cap the context stack and merge the samples for peak memory reduction and this is done after recursion compression.
The default value is -1 meaning no depth limit, in the future we can tune to a smaller one.
Reviewed By: hoy, wenlei
Differential Revision: https://reviews.llvm.org/D107800
One performance issue happened in profile generation and it turned out the line 525 loop is the bottleneck.
Moving the code outside of loop scope can fix this issue. The run time is improved from 30+mins to ~30s.
Reviewed By: hoy, wenlei
Differential Revision: https://reviews.llvm.org/D107529
Migrate pseudo probe decoding logic in llvm-profgen to MC, so other LLVM-base program could reuse existing codes. Redesign object layout of encoded and decoded pseudo probes.
Reviewed By: hoy
Differential Revision: https://reviews.llvm.org/D106861
This change tried to integrate a new count based aggregated type of perf script. The only difference of the format is that an aggregated count is added at the head of the original sample which means the same samples are repeated to the given count times. This is used to reduce the perf script size.
e.g.
```
2
4005dc
400634
400684
7f68c5788793
0x4005c8/0x4005dc/P/-/-/0 ....
```
Implemented by a dedicated PerfReader `AggregatedHybridPerfReader`.
Differential Revision: https://reviews.llvm.org/D107192
This change supports to run without parsing MMap binary loading events instead it always assumes binary is loaded at the preferred address. This is used when we have assured no binary load address changes or we have pre-processed the addresses resolution. Warn if there's interior mmap event but without leading mmap events.
Reviewed By: hoy
Differential Revision: https://reviews.llvm.org/D107097
In order to support different types of perf scripts, this change tried to refactor `PerfReader` by adding the base class `PerfReaderBase` and current HybridPerfReader is derived from it for CS profile generation. Common functions like, passMM2PEvents, extract_lbrs, extract_callstack, etc. can be reused.
Next step is to add LBR only reader(for non-CS profile) and aggregated perf scripts reader(do a pre-aggregation of scripts).
Reviewed By: hoy, wenlei
Differential Revision: https://reviews.llvm.org/D107014
[[noreturn]] can be used since Oct 2016 when the minimum compiler requirement was bumped to GCC 4.8/MSVC 2015.
Note: the definition of LLVM_ATTRIBUTE_NORETURN is kept for now.
The linker or post-link optimizer can create an ELF image with multiple executable segments each of which will be loaded separately at run time. This breaks the assumption of llvm-profgen that currently only supports one base load address. What it ends up with is that the subsequent mmap events will be treated as an overwrite of the first mmap event which will in turn screw up address mapping. While it is non-trivial to support multiple separate load addresses and given that on x64 those segments will always be loaded at consecutive addresses (though via separate mmap
sys calls), I'm adding an error checking logic to bail out if that's violated and keep using a single load address which is the address of the first executable segment.
Also changing the disassembly output from printing section offset to printing the virtual address instead, which matches the behavior of objdump.
Differential Revision: https://reviews.llvm.org/D103178
In a callback case, a return from internal code, say A, to external runtime can happen. The external runtime can then call back to another internal routine, say B. Making an artificial branch that looks like a return from A to B can confuse the unwinder to treat the instruction before B as the call instruction.
Reviewed By: wenlei, wmi
Differential Revision: https://reviews.llvm.org/D104546
Add a parameter of IsFSDiscriminator to function
getBaseDiscriminatorFromDiscriminator().
This function currently checks the internal flag of
--enable-fs-discriminator. This is not good because we might
change the default value of the internal flag.
Note that we have a default parameter. This is just
because create_afdo_tool has a call-site to it.
I will remove the default parameter in a later patch.
Differential Revision: https://reviews.llvm.org/D104584
As a follow-up to https://reviews.llvm.org/D104129, I'm cleaning up the danling probe related code in both the compiler and llvm-profgen.
I'm seeing a 5% size win for the pseudo_probe section for SPEC2017 and 10% for Ciner. Certain benchmark such as 602.gcc has a 20% size win. No obvious difference seen on build time for SPEC2017 and Cinder.
Reviewed By: wenlei
Differential Revision: https://reviews.llvm.org/D104477
We were using 0 as an indicator of invalid offset when computing disjoint ranges. In reality, 0 can be an valid code offset which stands for the first function in .text section. I'm using UINT64_MAX as an invalid code offset instead.
Reviewed By: wenlei
Differential Revision: https://reviews.llvm.org/D104497
If we have seen an inwards transition from external code to internal code, but not a following outwards transition, the inwards transition is likely due to interrupt which is usually unpaired. Ignore current and subsequent entries since they are likely from an unrelated pre-interrupt context.
LBR records from different interrupt context are unrelated and they should not be mixed together. Currenlty the OS does this for task-scheduling interrupt but not for all interrupts.
Reviewed By: wenlei, wlei
Differential Revision: https://reviews.llvm.org/D104276
We were using a `StringMap` object to store all profiles to be emitted. The object is basically an unordered hash table, therefore updating it in the process of trasvering it may cause issue since the underlying bucket array could change.
I'm also moving the `csspgo-preinliner` switch around so that no context tri will be constructed (by the constructor of `CSPreInliner`) when the switch is off.
Reviewed By: wenlei
Differential Revision: https://reviews.llvm.org/D104267
Previously dangling samples were represented by INT64_MAX in sample profile while probes never executed were not reported. This was based on an observation that dangling probes were only at a smaller portion than zero-count probes. However, with compiler optimizations, dangling probes end up becoming at large portion of all probes in general and reporting them does not make sense from profile size point of view. This change flips sample reporting by reporting zero-count probes instead. This enabled dangling probe to be represented by none (missing entry in profile). This has a couple benefits:
1. Reducing sample profile size in optimize mode, even when the number of non-executed probes outperform the number of dangling probes, since INT64_MAX takes more space over 0 to encode.
2. Binary size savings. No need to encode dangling probe anymore, since missing probes are treated as dangling in the profile reader.
3. Reducing compiler work to track dangling probes. However, for probes that are real dead and removed, we still need the compiler to identify them so that they can be reported as zero-count, instead of mistreated as dangling probes.
4. Improving counts quality by respecting the counts already collected on the non-dangling copy of a probe. A probe, when duplicated, gets two copies at runtime. If one of them is dangling while the other is not, merging the two probes at profile generation time will cause the real samples collected on the non-dangling one to be discarded. Not reporting the dangling counterpart will keep the real samples.
5. Better readability.
6. Be consistent with non-CS dwarf line number based profile. Zero counts are trusted by the compiler counts inferencer while missing counts will be inferred by the compiler.
Note that the current patch does include any work for #3. There will be follow-up changes.
For #1, I've seen for a large Facebook service, the text profile is reduced by 7%. For extbinary profile, the size of LBRProfileSection is reduced by 35%.
For #4, I have seen general counts quality for SPEC2017 is improved by 10%.
Reviewed By: wenlei, wlei, wmi
Differential Revision: https://reviews.llvm.org/D104129
This change provides the option to merge and aggregate cold context by the last k frames instead of context-less name. By default K = 1 means the context-less one.
This is for better perf tuning. The more selective merging and trimming will rely on llvm-profgen's preinliner.
Reviewed By: wenlei, hoy
Differential Revision: https://reviews.llvm.org/D104131
Make extended binary the default output format for CSSPGO. This avoids having to pass flag every time when generating profile. It also matches llvm-profdata where binary profile is the default (should we switch to extbinary as default for llvm-profdata?).
We plan to compress name table for context profile, which depends on the built-in compression of extbinary.
Differential Revision: https://reviews.llvm.org/D103650
llvm-profgen uses profile summary based cold threshold to merge and trim cold context profile. This is to strike a good balance between profile size and performance.
We've been using 99.9% as the cutoff to save profile size without affecting performance. This change switch to use 99.9% instead of 99.9999% as default cold threshold cutoff for llvm-profgen.
Redundant switch csprof-cold-thres is also removed and tests cleaned up.
Differential Revision: https://reviews.llvm.org/D103071
Fixing an issue where samples collected for an untrackable frame is not reported. An untrackable frame refers to a frame whose caller is untrackable due to missing debug info or pseudo probe. Though the frame is connected to its parent frame through the frame pointer chain at runtime, the compiler cannot build the connection without debug info or pseudo probe. In such case we just need to report the untrackable frame as the base frame and all of its child frames.
With more samples reported I'm seeing this improves the performance of an internal benchmark by 2.5%.
Reviewed By: wenlei, wlei
Differential Revision: https://reviews.llvm.org/D102961
This makes it possible for targets to define their own MCObjectFileInfo.
This MCObjectFileInfo is then used to determine things like section alignment.
This is a follow up to D101462 and prepares for the RISCV backend defining the
text section alignment depending on the enabled extensions.
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D101921
This untangles the MCContext and the MCObjectFileInfo. There is a circular
dependency between MCContext and MCObjectFileInfo. Currently this dependency
also exists during construction: You can't contruct a MOFI without a MCContext
without constructing the MCContext with a dummy version of that MOFI first.
This removes this dependency during construction. In a perfect world,
MCObjectFileInfo wouldn't depend on MCContext at all, but only be stored in the
MCContext, like other MC information. This is future work.
This also shifts/adds more information to the MCContext making it more
available to the different targets. Namely:
- TargetTriple
- ObjectFileType
- SubtargetInfo
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D101462
The change adds support for triming and merging cold context when mergine CSSPGO profiles using llvm-profdata. This is similar to the context profile trimming in llvm-profgen, however the flexibility to trim cold context after profile is generated can be useful.
Differential Revision: https://reviews.llvm.org/D100528
Report dangling probes for frames that have real samples collected. Dangling probes are the probes associated to an empty block. When reported, sample count on a dangling probe will not be trusted by the compiler and we will rely on the counts inference algorithm to get the probe a reasonable count. This actually fixes a bug where previously only those dangling probes with samples collected were reported.
This patch also fixes two existing issues. Pseudo probes are stored in `Address2ProbesMap` and their pointers are used in `PseudoProbeInlineTree`. Previously `std::vector` was used to store probes and the pointers to probes may get obsolete as the vector grows. I'm changing `std::vector` to `std::list` instead.
The other issue is that all outlined functions shared the same inline frame previously due to the unchanged `Index` value as the dummy inlineSite identifier.
Good results seen for SPEC2017 in general regarding profile quality.
Reviewed By: wenlei, wlei
Differential Revision: https://reviews.llvm.org/D100235
CommandLine.h is indirectly included in ~50% of TUs when building
clang, and VirtualFileSystem.h is large.
(Already remarked by jhenderson on D70769.)
No behavior change.
Differential Revision: https://reviews.llvm.org/D100957
This patch fixed the following issues along side with some refactoring:
1. Fix bugs where StringRef for context string out live the underlying std::string. We now keep string table in profile generator to hold std::strings. We also do the same for bracketed context strings in profile writer.
2. Make sure profile output strictly follow (total sample, name) order. Previously, there's inconsistency between ProfileMap's key and FunctionSamples's name, leading to inconsistent ordering. This is now fixed by introducing context profile canonicalization. Assertions are also added to make sure ProfileMap's key and FunctionSamples's name are always consistent.
3. Enhanced error handling for profile writing to make sure we bubble up errors properly for both llvm-profgen and llvm-profdata when string table is not populated correctly for extended binary profile.
4. Keep all internal context representation bracket free. This avoids creating new strings for context trimming, merging and preinline. getNameWithContext API is now simplied accordingly.
5. Factor out the code for context trimming and merging into SampleContextTrimmer in SampleProf.cpp. This enables llvm-profdata to use the trimmer when merging profiles. Changes in llvm-profgen will be in separate patch.
Differential Revision: https://reviews.llvm.org/D100090
Use profiled call edges to augment the top-down order. There are cases that the top-down order computed based on the static call graph doesn't reflect real execution order. For example:
1. Incomplete static call graph due to unknown indirect call targets. Adjusting the order by considering indirect call edges from the profile can enable the inlining of indirect call targets by allowing the caller processed before them.
2. Mutual call edges in an SCC. The static processing order computed for an SCC may not reflect the call contexts in the context-sensitive profile, thus may cause potential inlining to be overlooked. The function order in one SCC is being adjusted to a top-down order based on the profile to favor more inlining.
3. Transitive indirect call edges due to inlining. When a callee function is inlined into into a caller function in LTO prelink, every call edge originated from the callee will be transferred to the caller. If any of the transferred edges is indirect, the original profiled indirect edge, even if considered, would not enforce a top-down order from the caller to the potential indirect call target in LTO postlink since the inlined callee is gone from the static call graph.
4. #3 can happen even for direct call targets, due to functions defined in header files. Header functions, when included into source files, are defined multiple times but only one definition survives due to ODR. Therefore, the LTO prelink inlining done on those dropped definitions can be useless based on a local file scope. More importantly, the inlinee, once fully inlined to a to-be-dropped inliner, will have no profile to consume when its outlined version is compiled. This can lead to a profile-less prelink compilation for the outlined version of the inlinee function which may be called from external modules. while this isn't easy to fix, we rely on the postlink AutoFDO pipeline to optimize the inlinee. Since the survived copy of the inliner (defined in headers) can be inlined in its local scope in prelink, it may not exist in the merged IR in postlink, and we'll need the profiled call edges to enforce a top-down order for the rest of the functions.
Considering those cases, a profiled call graph completely independent of the static call graph is constructed based on profile data, where function objects are not even needed to handle case #3 and case 4.
I'm seeing an average 0.4% perf win out of SPEC2017. For certain benchmark such as Xalanbmk and GCC, the win is bigger, above 2%.
The change is an enhancement to https://reviews.llvm.org/D95988.
Reviewed By: wmi, wenlei
Differential Revision: https://reviews.llvm.org/D99351
This change sets up a framework in llvm-profgen to estimate inline decision and adjust context-sensitive profile based on that. We call it a global pre-inliner in llvm-profgen.
It will serve two purposes:
1) Since context profile for not inlined context will be merged into base profile, if we estimate a context will not be inlined, we can merge the context profile in the output to save profile size.
2) For thinLTO, when a context involving functions from different modules is not inined, we can't merge functions profiles across modules, leading to suboptimal post-inline count quality. By estimating some inline decisions, we would be able to adjust/merge context profiles beforehand as a mitigation.
Compiler inline heuristic uses inline cost which is not available in llvm-profgen. But since inline cost is closely related to size, we could get an estimate through function size from debug info. Because the size we have in llvm-profgen is the final size, it could also be more accurate than the inline cost estimation in the compiler.
This change only has the framework, with a few TODOs left for follow up patches for a complete implementation:
1) We need to retrieve size for funciton//inlinee from debug info for inlining estimation. Currently we use number of samples in a profile as place holder for size estimation.
2) Currently the thresholds are using the values used by sample loader inliner. But they need to be tuned since the size here is fully optimized machine code size, instead of inline cost based on not yet fully optimized IR.
Differential Revision: https://reviews.llvm.org/D99146
Switch to use cold threshold from profile summary for cold context merging and trimming, instead of relying on hard coded values. Minor refactoring included for switch names, etc.
Differential Revision: https://reviews.llvm.org/D98921
This changes adds attribute field for metadata of context profile. Currently we have an inline attribute that indicates whether the leaf frame corresponding to a context profile was inlined in previous build.
This will be used to help estimating inlining and be taken into account when trimming context. Changes for that in llvm-profgen will follow. It will also help tuning.
Differential Revision: https://reviews.llvm.org/D98823
Previously we didn't support to keep the unique linkage name(-funique-internal-linkage-name) in llvm-profgen. As discussed in https://reviews.llvm.org/D96932, we choose to do canonicalization for it.
Now since "selected" is set as the default parameter of getCanonicalFnName in `D96932`, we don't need to add any attribute here for the previous usage and only fix the missing usage in the pseudo probe decoding.
Differential Revision: https://reviews.llvm.org/D98226
For ThinLTO's prelink compilation, we need to put external inline candidates into an import list attached to function's entry count metadata. This enables ThinLink to treat such cross module callee as hot in summary index, and later helps postlink to import them for profile guided cross module inlining.
For AutoFDO, the import list is retrieved by traversing the nested inlinee functions. For CSSPGO, since profile is flatterned, a few things need to happen for it to work:
- When loading input profile in extended binary format, we need to load all child context profile whose parent is in current module, so context trie for current module includes potential cross module inlinee.
- In order to make the above happen, we need to know whether input profile is CSSPGO profile before start reading function profile, hence a flag for profile summary section is added.
- When searching for cross module inline candidate, we need to walk through the context trie instead of nested inlinee profile (callsite sample of AutoFDO profile).
- Now that we have more accurate counts with CSSPGO, we swtiched to use entry count instead of total count to decided if an external callee is potentially beneficial to inline. This make it consistent with how we determine whether call tagert is potential inline candidate.
Differential Revision: https://reviews.llvm.org/D98590
Previously we errored out when disassembling illegal instructions and there would be no profile generated. In fact illegal instructions are not uncommon and we'd better skip them and print "unknown" instead of erroring out. This matches the behavior of llvm-objdump (see disassembleObject in llvm-objdump.cpp).
Reviewed By: wlei, wenlei
Differential Revision: https://reviews.llvm.org/D97776
Wenlei He