Commit Graph

132 Commits

Author SHA1 Message Date
Rong Xu 998b97f6f1 [llvm-profdata] Add overlap command to compute similarity b/w two profile files
Add overlap functionality to llvm-profdata tool to compute the similarity
between two profile files.

Differential Revision: https://reviews.llvm.org/D60977

llvm-svn: 359612
2019-04-30 21:19:12 +00:00
Fangrui Song 8b0a15b0ef [llvm-profdata] Deleted unused Cutoffs added by D16005
llvm-svn: 356248
2019-03-15 10:43:51 +00:00
Rong Xu a6ff69f6dd [PGO] Context sensitive PGO (part 2)
Part 2 of CSPGO changes (mostly related to ProfileSummary).
Note that I use a default parameter in setProfileSummary() and getSummary().
This is to break the dependency in clang. I will make the parameter explicit
after changing clang in a separated patch.

Differential Revision: https://reviews.llvm.org/D54175

llvm-svn: 355131
2019-02-28 19:55:07 +00:00
Fangrui Song ef5987592e Fix some include order and file headers issues. NFC
llvm-svn: 354550
2019-02-21 07:42:31 +00:00
Petar Jovanovic 40a7f63c37 [PGO] Fix the type of the formated variable
Change the format type of Value to PRIu64 since it is a uint64_t.
The problem was detected on mips boards building 32-bit binaries,
where it was printing junk values and causing test failure.

Patch by Milos Stojanovic.

Differential Revision: https://reviews.llvm.org/D57583

llvm-svn: 353194
2019-02-05 18:09:28 +00:00
Chandler Carruth 2946cd7010 Update the file headers across all of the LLVM projects in the monorepo
to reflect the new license.

We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.

Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.

llvm-svn: 351636
2019-01-19 08:50:56 +00:00
Rong Xu 52aa224aff [llvm-profdata] add value-cutoff functionality in show command
This patch improves llvm-profdata show command:
(1) add -value-cutoff=<N> option: Show only those functions whose max count
    values are greater or equal to N.
(2) add -list-below-cutoff option: Only output names of functions whose max
    count value are below the cutoff.
(3) formats value-profile counts and prints out percentage.

Differential Revision: https://reviews.llvm.org/D56342

llvm-svn: 350673
2019-01-08 22:41:48 +00:00
Rong Xu 7162e16e6b [PGO] Revert r350579 to fix commit message.
Will re-commit it using the correct commit message.

llvm-svn: 350670
2019-01-08 22:37:12 +00:00
Rong Xu 6f366c3a04 [PGO] Use SourceFileName rather module name in PGOFuncName
In LTO or Thin-lto mode (though linker plugin), the module
names are of temp file names which are different for
different compilations. Using SourceFileName avoids the issue.
This should not change any functionality for current PGO as
all the current callers of getPGOFuncName() is before LTO.

llvm-svn: 350579
2019-01-07 23:25:56 +00:00
Richard Smith 3164fcfd27 Add flag to llvm-profdata to allow symbols in profile data to be remapped, and
add a tool to generate symbol remapping files.

Summary:
The new tool llvm-cxxmap builds a symbol mapping table from a file containing
a description of partial equivalences to apply to mangled names and files
containing old and new symbol tables.

Reviewers: davidxl

Subscribers: mgorny, llvm-commits

Differential Revision: https://reviews.llvm.org/D51470

llvm-svn: 342168
2018-09-13 20:22:02 +00:00
Richard Smith c6ba9ca169 Make llvm-profdata show -text work as advertised in the documentation.
Per LLVM's CommandGuide, llvm-profdata show -text is supposed to produce
textual output that can be passed as input to further llvm-profdata
invocations. This previously didn't work for two reasons:

1) -text was not sufficient to enable the machine-readable text format output;
instead, -text was effectively ignored if -counts was not also specified. (With
this patch, -counts is instead ignored if -text is specified, because the
machine-readable text format always includes counts.)

2) When the input data was an IR-level profile, the :ir marker was missing from
the output, resulting in a text format output that would not be usable as
profiling data due to function hash mismatches.

Differential Revision: https://reviews.llvm.org/D51188

llvm-svn: 340592
2018-08-24 01:34:45 +00:00
Wei Mi d9be2c7e64 [NFC] Change sample profile format enum name SPF_Raw_Binary to SPF_Binary.
Some out-of-tree targets depend on the enum name SPF_Binary. Keep the name
can avoid unnecessary churn to those targets.

llvm-svn: 334476
2018-06-12 05:53:49 +00:00
Wei Mi 432db3b43b Fix a typo in rL334447.
llvm-svn: 334475
2018-06-12 04:43:09 +00:00
Wei Mi a0c0857e7a [SampleFDO] Add a new compact binary format for sample profile.
Name table occupies a big chunk of size in current binary format sample profile.
In order to reduce its size, the patch changes the sample writer/reader to
save/restore MD5Hash of names in the name table. Sample annotation phase will
also use MD5Hash of name to query samples accordingly.

Experiment shows compact binary format can reduce the size of sample profile by
2/3 compared with binary format generally.

Differential Revision: https://reviews.llvm.org/D47955

llvm-svn: 334447
2018-06-11 22:40:43 +00:00
Jonas Devlieghere e46b7565bb [llvm-profdata] Use WithColor for printing errors
Use convenience helpers in WithColor to print errors and warnings.

Differential revision: https://reviews.llvm.org/D45658

llvm-svn: 330262
2018-04-18 14:42:33 +00:00
Rui Ueyama 197194b6c9 Define InitLLVM to do common initialization all at once.
We have a few functions that virtually all command wants to run on
process startup/shutdown. This patch adds InitLLVM class to do that
all at once, so that we don't need to copy-n-paste boilerplate code
to each llvm command's main() function.

Differential Revision: https://reviews.llvm.org/D45602

llvm-svn: 330046
2018-04-13 18:26:06 +00:00
Vedant Kumar 188efda585 [llvm-profdata] Don't treat non-fatal merge errors as fatal
This fixes an issue seen on the coverage bot:
http://lab.llvm.org:8080/green/view/Experimental/job/clang-stage2-coverage-R/1930

Profile merging shouldn't fail if a single counter mismatch is detected.

llvm-svn: 318555
2017-11-17 21:18:32 +00:00
Vedant Kumar faaa42ad0a [llvm-profdata] Fix a dangling reference to an error string
llvm-svn: 318502
2017-11-17 02:58:23 +00:00
Adam Nemet 1142b2d7b7 [llvm-profdata] Report if profile data file is IR- or FE-level
Differential Revision: https://reviews.llvm.org/D39997

llvm-svn: 318159
2017-11-14 16:59:18 +00:00
Rafael Espindola 8c0ff9508d Bring r314809 back.
But now include a check for CPU_COUNT so we still build on 10 year old
versions of glibc.

Original message:

Use sched_getaffinity instead of std:🧵:hardware_concurrency.

The issue with std:🧵:hardware_concurrency is that it forwards
to libc and some implementations (like glibc) don't take thread
affinity into consideration.

With this change a llvm program that can execute in only 2 cores will
use 2 threads, even if the machine has 32 cores.

This makes benchmarking a lot easier, but should also help if someone
doesn't want to use all cores for compilation for example.

llvm-svn: 314931
2017-10-04 20:27:01 +00:00
Daniel Neilson bef94bcbae Revert D38481 due to missing cmake check for CPU_COUNT
Summary:
This reverts D38481. The change breaks systems with older versions of glibc. It
injects a use of CPU_COUNT() from sched.h without checking to ensure that the
function exists first.

Reviewers:

Subscribers:

llvm-svn: 314922
2017-10-04 18:19:03 +00:00
Rafael Espindola 6e182fbab4 Use sched_getaffinity instead of std:🧵:hardware_concurrency.
The issue with std:🧵:hardware_concurrency is that it forwards
to libc and some implementations (like glibc) don't take thread
affinity into consideration.

With this change a llvm program that can execute in only 2 cores will
use 2 threads, even if the machine has 32 cores.

This makes benchmarking a lot easier, but should also help if someone
doesn't want to use all cores for compilation for example.

llvm-svn: 314809
2017-10-03 16:25:15 +00:00
Xinliang David Li 801b5319c5 [ProfileData] Add new option to dump topn hottest functions
Differential Revision: http://reviews.llvm.org/D35155

llvm-svn: 307702
2017-07-11 20:30:43 +00:00
David Blaikie 98cce00371 llvm-profdata: Reduce memory usage by using Error callback rather than member
Reduces llvm-profdata memory usage on a large profile from 7.8GB to 5.1GB.

The ProfData API now supports reporting all the errors/warnings rather
than only the first, though llvm-profdata ignores everything after the
first for now to preserve existing behavior. (if there's a desire for
other behavior, happy to implement that - but might be as well left for
a separate patch)

Reviewers: davidxl

Differential Revision: https://reviews.llvm.org/D35149

llvm-svn: 307516
2017-07-10 03:04:59 +00:00
David Blaikie cf9d52c690 Prototype: Reduce llvm-profdata merge memory usage further
The InstrProfWriter already stores the name and hash of the record in
the nested maps it uses for lookup while merging - this data is
duplicated in the value within the maps.

Refactor the InstrProfRecord to use a nested struct for the counters
themselves so that InstrProfWriter can use this nested struct alone
without the name or hash duplicated there.

This work is incomplete, but enough to demonstrate the value (around a
50% decrease in memory usage for a large test case (10GB -> 5GB)).
Though most of that decrease is probably from removing the
SoftInstrProfError as well, but I haven't implemented a replacement for
it yet. (it needs to go with the counters, because the operations on the
counters - merging, etc, are where the failures are - unlike the
name/hash which are totally unused by those counter-related operations
and thus easy to split out)

Ongoing discussion about removing SoftInstrProfError as a field of the
InstrProfRecord is happening on the thread that added it - including
the possibility of moving back towards an earlier version of that
proposed patch that passed SoftInstrProfError through the various APIs,
rather than as a member of InstrProfRecord.

Reviewers: davidxl

Differential Revision: https://reviews.llvm.org/D34838

llvm-svn: 307298
2017-07-06 19:00:12 +00:00
Vedant Kumar b5794ca90c [ProfileData] PR33517: Check for failure of symtab creation
With PR33517, it became apparent that symbol table creation can fail
when presented with malformed inputs. This patch makes that sort of
error detectable, so llvm-cov etc. can fail more gracefully.

Specifically, we now check that function names within the symbol table
aren't empty.

Testing: check-{llvm,clang,profile}, some unit test updates.
llvm-svn: 305765
2017-06-20 01:38:56 +00:00
Teresa Johnson cd2aa0d2e4 Fix a couple of typos in memory intrinsic optimization output (NFC)
s/instrinsic/intrinsic

llvm-svn: 303782
2017-05-24 17:55:25 +00:00
Rong Xu 60faea19f8 Resubmit r297897: [PGO] Value profile for size of memory intrinsic calls
R297897 inadvertently enabled annotation for memop profiling. This new patch
fixed it.

llvm-svn: 297996
2017-03-16 21:15:48 +00:00
Eric Liu 971de62291 Revert "[PGO] Value profile for size of memory intrinsic calls"
This commit reverts r297897 and r297909.

llvm-svn: 297951
2017-03-16 13:16:35 +00:00
Rong Xu 4ed52798ce [PGO] Value profile for size of memory intrinsic calls
This patch adds the value profile support to profile the size parameter of
memory intrinsic calls: memcpy, memcmp, and memmov.

Differential Revision: http://reviews.llvm.org/D28965

llvm-svn: 297897
2017-03-15 21:47:27 +00:00
Rong Xu 0cf1f56a8c [PGO] Refactor profile dumping function for ease of adding other profile kind
Refactor the dumping function so that we can add other value profile kind easily.

Differential Revision: https://reviews.llvm.org/D30752

llvm-svn: 297399
2017-03-09 19:03:57 +00:00
Chris Bieneman 8e47604975 [CMake] llvm-profdata depends on intrinsics_gen
llvm-profdata.cpp has the following include chain:

llvm/ProfileData/SampleProfReader.h
llvm/IR/Function.h
llvm/IR/Argument.h
llvm/IR/Attributes.h
llvm/IR/Attributes.gen

This means llvm-profdata needs to depend on intrinsics_gen.

llvm-svn: 287394
2016-11-18 23:04:15 +00:00
Rong Xu fe90d86c8e [PGO] Fix a use-after-move. NFC.
llvm-svn: 284664
2016-10-19 23:31:59 +00:00
Rong Xu 2c684cfd94 [PGO] Fix bogus warning for merging empty llvm profile file
Profile runtime can generate an empty raw profile (when there is no function in
the shared library). This empty profile is treated as a text format profile.  A
test format profile without the flag of "#IR" is thought to be a clang
generated profile.  So in llvm profile merging, we will get a bogus warning of
"Merge IR generated profile with Clang generated profile."

The fix here is to skip the empty profile (when the buffer size is 0) for
profile merge.

Reviewers: vsk, davidxl

Differential Revision: http://reviews.llvm.org/D25687

llvm-svn: 284659
2016-10-19 22:51:17 +00:00
Mehdi Amini 732afdd09a Turn cl::values() (for enum) from a vararg function to using C++ variadic template
The core of the change is supposed to be NFC, however it also fixes
what I believe was an undefined behavior when calling:

 va_start(ValueArgs, Desc);

with Desc being a StringRef.

Differential Revision: https://reviews.llvm.org/D25342

llvm-svn: 283671
2016-10-08 19:41:06 +00:00
Xinliang David Li deda33cdbd [Profile] dump ic value profile value/site-count histogram
Differential Revision: http://reviews.google.com/D24783

llvm-svn: 282017
2016-09-20 21:04:22 +00:00
Justin Bogner 253eb17b4e llvm-profdata: Clarify the top level help
It wasn't very obvious that you're supposed to call help on the
subcommands. This should help.

llvm-svn: 277678
2016-08-03 23:10:51 +00:00
Benjamin Kramer a81f4728f3 [llvm-profdata] Bring back reading profile data from STDIN.
This feature was lost in r276197.

llvm-svn: 276407
2016-07-22 12:39:55 +00:00
Benjamin Kramer 929e7dbb7a [profdata] Remove constructor that MSVC 2013 pretends to not understand.
No functionality change intended.

llvm-svn: 276284
2016-07-21 14:29:11 +00:00
Xinliang David Li 9a1bfcfa16 Reapply r276185
Fix the test case that should not depend on dir iteration order.

llvm-svn: 276197
2016-07-20 22:24:52 +00:00
Xinliang David Li ce3f385eeb Revert r276185 -- build bot failure
llvm-svn: 276194
2016-07-20 21:50:38 +00:00
Xinliang David Li d0b867e3e5 [Profile] support directory reading in profile merging
Differential Revision:  http://reviews.llvm.org/D22560

llvm-svn: 276185
2016-07-20 21:31:29 +00:00
Vedant Kumar e3a0bf5048 Retry: [llvm-profdata] Speed up merging by using a thread pool
Add a "-j" option to llvm-profdata to control the number of threads used.
Auto-detect NumThreads when it isn't specified, and avoid spawning threads when
they wouldn't be beneficial.

I tested this patch using a raw profile produced by clang (147MB). Here is the
time taken to merge 4 copies together on my laptop:

  No thread pool: 112.87s user 5.92s system 97% cpu 2:01.08 total
  With 2 threads: 134.99s user 26.54s system 164% cpu 1:33.31 total

Changes since the initial commit:

  - When handling odd-length inputs, call ThreadPool::wait() before merging the
    last profile. Should fix a race/off-by-one (see r275937).

Differential Revision: https://reviews.llvm.org/D22438

llvm-svn: 275938
2016-07-19 01:17:20 +00:00
Vedant Kumar 21ab20e005 Revert "[llvm-profdata] Speed up merging by using a thread pool"
This reverts commit r275921. It broke the ppc64be bot:

  http://lab.llvm.org:8011/builders/clang-ppc64be-linux-multistage/builds/3537

I'm not sure why it broke, but based on the output, it looks like an
off-by-one (one profile left un-merged).

llvm-svn: 275937
2016-07-19 00:57:09 +00:00
Vedant Kumar 0bd9907581 [llvm-profdata] Speed up merging by using a thread pool
Add a "-j" option to llvm-profdata to control the number of threads
used. Auto-detect NumThreads when it isn't specified, and avoid spawning
threads when they wouldn't be beneficial.

I tested this patch using a raw profile produced by clang (147MB). Here is the
time taken to merge 4 copies together on my laptop:

  No thread pool: 112.87s user 5.92s system 97% cpu 2:01.08 total
  With 2 threads: 134.99s user 26.54s system 164% cpu 1:33.31 total

Differential Revision: https://reviews.llvm.org/D22438

llvm-svn: 275921
2016-07-18 22:02:39 +00:00
Benjamin Kramer 1afc1de406 Apply another batch of fixes from clang-tidy's performance-unnecessary-value-param.
Contains some manual fixes. No functionality change intended.

llvm-svn: 273047
2016-06-17 20:41:14 +00:00
Richard Smith 2ad6d48b0c Search for llvm-symbolizer binary in the same directory as argv[0], before
looking for it along $PATH. This allows installs of LLVM tools outside of
$PATH to find the symbolizer and produce pretty backtraces if they crash.

llvm-svn: 272232
2016-06-09 00:53:21 +00:00
Vedant Kumar cef4360ac4 Retry^4 "[llvm-profdata] Add option to ingest filepaths from a file"
Changes since the initial commit:
- Use echo instead of printf. This should side-step the character
  escaping issues on Windows.

Differential Revision: http://reviews.llvm.org/D20980

llvm-svn: 272068
2016-06-07 22:47:31 +00:00
Vedant Kumar 8d0e861e9b Revert "Retry^2 "[llvm-profdata] Add option to ingest filepaths from a file""
This reverts commit r271953. It's still breaking on Windows, though the
list initialization issue is fixed:

http://bb.pgr.jp/builders/ninja-clang-i686-msc19-R/builds/3751

llvm-svn: 271963
2016-06-06 23:43:56 +00:00
Vedant Kumar f051269a7f Retry^2 "[llvm-profdata] Add option to ingest filepaths from a file"
Changes since the initial commit:
- Normalize file paths read from the file to prevent Windows path
  separators from escaping parts of the path.
- Since we need to store the normalized file paths in WeightedFile,
  don't do tricky things to keep the source MemoryBuffer alive.
- Don't use list-initialization for a std::string in WeightedFile.

Differential Revision: http://reviews.llvm.org/D20980

llvm-svn: 271953
2016-06-06 23:17:22 +00:00