GeezerVison

View reviewed changes

corona10 left a comment •

edited

Loading

Uh oh!

There was an error while loading. Please reload this page.

Copy link

Copy Markdown

Member

There was a problem hiding this comment.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Two things need to be checked.

~~I failed to build the binary with this patch, This can be due to the BOLT bug but I would like to know which BOLT version you used.~~ -> solved

BOLT-INFO: Allocation combiner: 30 empty spaces coalesced (dyn count: 63791805).
 #0 0x0000563eb3e8d705 PrintStackTraceSignalHandler(void*) Signals.cpp:0:0
 #1 0x0000563eb3e8b2d4 SignalHandler(int) Signals.cpp:0:0
 #2 0x00007fc228930520 (/lib/x86_64-linux-gnu/libc.so.6+0x42520)
 #3 0x0000563eb4ebd106 llvm::bolt::BinaryFunction::translateInputToOutputAddress(unsigned long) const (/usr/local/bin/llvm-bolt+0x1c52106)
 #4 0x0000563eb3f52658 llvm::bolt::DWARFRewriter::updateUnitDebugInfo(llvm::DWARFUnit&, llvm::bolt::DebugInfoBinaryPatcher&, llvm::bolt::DebugAbbrevWriter&, llvm::bolt::DebugLocWriter&, llvm::bolt::DebugRangesSectionWriter&, llvm::Optional<unsigned long>) (/usr/local/bin/llvm-bolt+0xce7658)
 #5 0x0000563eb3f5688b llvm::bolt::DWARFRewriter::updateDebugInfo()::'lambda0'(unsigned long, llvm::DWARFUnit*)::operator()(unsigned long, llvm::DWARFUnit*) const DWARFRewriter.cpp:0:0
 #6 0x0000563eb3f5c45a llvm::bolt::DWARFRewriter::updateDebugInfo() (/usr/local/bin/llvm-bolt+0xcf145a)
 #7 0x0000563eb3f1aef8 llvm::bolt::RewriteInstance::updateMetadata() (/usr/local/bin/llvm-bolt+0xcafef8)
 #8 0x0000563eb3f428e6 llvm::bolt::RewriteInstance::run() (/usr/local/bin/llvm-bolt+0xcd78e6)
 #9 0x0000563eb355ccf8 main (/usr/local/bin/llvm-bolt+0x2f1cf8)
#10 0x00007fc228917d90 __libc_start_call_main ./csu/../sysdeps/nptl/libc_start_call_main.h:58:16
#11 0x00007fc228917e40 call_init ./csu/../csu/libc-start.c:128:20
#12 0x00007fc228917e40 __libc_start_main ./csu/../csu/libc-start.c:379:5
#13 0x0000563eb35dbd75 _start (/usr/local/bin/llvm-bolt+0x370d75)
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace.
Stack dump:
0.	Program arguments: /usr/local/bin/llvm-bolt python -o python.bolt -data=python.fdata -update-debug-sections -reorder-blocks=ext-tsp -reorder-functions=hfsort+ -split-functions=3 -icf=1 -inline-all -split-eh -reorder-functions-use-hot-size -peepholes=all -jump-tables=aggressive -inline-ap -indirect-call-promotion=all -dyno-stats -use-gnu-stack -frame-opt=hot
make: *** [Makefile:800: bolt-opt] Segmentation fault (core dumped

While profiling, I met the test failure, would you like to check that the optimized binary pass all std python test? (e.g python -m test), I met the related issue with the last attempts and it was solved by profiling through python -m test -> solved

./python.bolt_inst -m test --pgo --timeout=1200 || true
0:00:00 load avg: 2.17 Run tests sequentially (timeout: 20 min)
0:00:00 load avg: 2.17 [ 1/44] test_array
0:00:01 load avg: 2.17 [ 2/44] test_base64
0:00:02 load avg: 2.07 [ 3/44] test_binascii
0:00:02 load avg: 2.07 [ 4/44] test_binop
0:00:02 load avg: 2.07 [ 5/44] test_bisect
0:00:02 load avg: 2.07 [ 6/44] test_bytes
0:00:06 load avg: 2.07 [ 7/44] test_bz2
0:00:06 load avg: 2.07 [ 8/44] test_cmath
0:00:07 load avg: 2.07 [ 9/44] test_codecs
0:00:08 load avg: 1.99 [10/44] test_collections
0:00:09 load avg: 1.99 [11/44] test_complex
0:00:10 load avg: 1.99 [12/44] test_dataclasses
0:00:10 load avg: 1.99 [13/44] test_datetime
0:00:14 load avg: 1.83 [14/44] test_decimal
0:00:18 load avg: 1.76 [15/44] test_difflib
0:00:19 load avg: 1.76 [16/44] test_embed
0:00:21 load avg: 1.76 [17/44] test_float
0:00:22 load avg: 1.76 [18/44] test_fstring
0:00:23 load avg: 1.70 [19/44] test_functools
0:00:23 load avg: 1.70 [20/44] test_generators
0:00:24 load avg: 1.70 [21/44] test_hashlib
0:00:25 load avg: 1.70 [22/44] test_heapq
0:00:26 load avg: 1.70 [23/44] test_int
0:00:26 load avg: 1.70 [24/44] test_itertools
0:00:32 load avg: 1.64 [25/44] test_json
0:00:36 load avg: 1.59 [26/44] test_long
0:00:39 load avg: 1.54 [27/44] test_lzma
0:00:39 load avg: 1.54 [28/44] test_math
0:00:42 load avg: 1.50 [29/44] test_memoryview
0:00:43 load avg: 1.50 [30/44] test_operator
0:00:44 load avg: 1.50 [31/44] test_ordered_dict
0:00:46 load avg: 1.50 [32/44] test_patma
0:00:46 load avg: 1.50 [33/44] test_pickle
0:00:52 load avg: 1.46 [34/44] test_pprint
0:00:52 load avg: 1.42 [35/44] test_re
0:00:53 load avg: 1.42 [36/44] test_set
0:01:00 load avg: 1.39 [37/44] test_sqlite3
0:01:05 load avg: 1.36 [38/44] test_statistics
0:01:10 load avg: 1.33 [39/44] test_struct
0:01:11 load avg: 1.33 [40/44] test_tabnanny
0:01:12 load avg: 1.30 [41/44] test_time
0:01:15 load avg: 1.30 [42/44] test_unicode
test test_unicode failed
0:01:17 load avg: 1.28 [43/44] test_xml_etree -- test_unicode failed (1 failure)
0:01:19 load avg: 1.28 [44/44] test_xml_etree_c

Total duration: 1 min 21 sec
Tests result: FAILURE

I will share further investigation into this patch.
FYI, this is my environment.

- OS: Ubuntu 22.04 LTS
- BOLT revision e9b213131ae9c57f4f151d3206916676135b31b0
- gcc (Ubuntu 11.2.0-19ubuntu1) 11.2.0

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

All reactions

corona10 commented Aug 13, 2022

Copy link

Copy Markdown

Member

Hmm, I will try to build BOLT from LLVM 14.0.6

All reactions

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

corona10 commented Aug 13, 2022 •

edited

Loading

Uh oh!

There was an error while loading. Please reload this page.

Copy link

Copy Markdown

Member

I found why the BOLT was failed, I will downgrade the gcc version into 10.


DWARF 5 has become the default in GCC 11

All reactions

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

corona10 requested changes

View reviewed changes

corona10 left a comment •

edited

Loading

Uh oh!

There was an error while loading. Please reload this page.

Copy link

Copy Markdown

Member

There was a problem hiding this comment.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for work! All pipeline works correctly.

Please update https://github.com/python/cpython/blob/main/Doc/using/configure.rst too.
(If possible https://github.com/python/cpython/blob/main/Doc/whatsnew/3.12.rst too, I will update the whats new if you are too busy)
But please emphasize that this feature is experimental optimization support.

I am going to measure the performance enhancement soon through the pyperformance and also for the l1 i-cache miss ratio.

Looks like https://github.com/pyston/python-macrobenchmarks does not support Python 3.1[1-2] yet right? Please let me know if I know wrong.

plus
https://github.com/python/cpython/blob/main/Misc/ACKS Add your name in this file too :)

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

All reactions

Comment thread

Makefile.pre.in Resolved

Uh oh!

There was an error while loading. Please reload this page.

Comment thread

configure.ac Resolved

Uh oh!

There was an error while loading. Please reload this page.

bedevere-bot removed the awaiting review label

bedevere-bot commented Aug 13, 2022

Copy link

Copy Markdown

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

All reactions

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

bedevere-bot added the awaiting changes label

corona10 commented Aug 13, 2022 •

edited

Loading

Uh oh!

There was an error while loading. Please reload this page.

Copy link

Copy Markdown

Member

@gvanrossum @kmod cc @markshannon

Interesting result!
The following benchmark was measured on AWS c5n.metal / gcc-10. (base commit: f235178)
I wish to re-measure the benchmark from the FasterCPython project machine also.
I am going to measure the L1 i-cache miss ratio soon where the perf tool is available.

Benchmark	CPython 3.12 ./configure --enable-optimizations --with-lto	CPython 3.12 ./configure --enable-optimizations --with-lto --enable-bolt
2to3	269 ms	255 ms: 1.05x faster
chameleon	7.39 ms	7.02 ms: 1.05x faster
chaos	74.1 ms	68.8 ms: 1.08x faster
crypto_pyaes	82.3 ms	77.2 ms: 1.07x faster
deltablue	3.65 ms	3.41 ms: 1.07x faster
django_template	38.6 ms	35.3 ms: 1.09x faster
dulwich_log	67.6 ms	58.7 ms: 1.15x faster
fannkuch	385 ms	380 ms: 1.02x faster
float	73.2 ms	72.4 ms: 1.01x faster
genshi_text	24.3 ms	23.3 ms: 1.04x faster
genshi_xml	56.4 ms	52.8 ms: 1.07x faster
go	140 ms	136 ms: 1.03x faster
hexiom	6.40 ms	6.25 ms: 1.02x faster
html5lib	65.0 ms	60.7 ms: 1.07x faster
json_dumps	11.1 ms	10.4 ms: 1.07x faster
json_loads	28.7 us	26.3 us: 1.09x faster
logging_format	7.29 us	6.69 us: 1.09x faster
logging_silent	101 ns	97.6 ns: 1.03x faster
logging_simple	6.48 us	6.01 us: 1.08x faster
mako	10.6 ms	9.91 ms: 1.07x faster
meteor_contest	106 ms	102 ms: 1.04x faster
nbody	86.4 ms	87.7 ms: 1.02x slower
nqueens	91.3 ms	88.1 ms: 1.04x faster
pathlib	19.0 ms	16.8 ms: 1.13x faster
pickle_dict	32.2 us	32.6 us: 1.01x slower
pickle_list	4.69 us	4.62 us: 1.02x faster
pickle_pure_python	297 us	282 us: 1.05x faster
pidigits	177 ms	176 ms: 1.01x faster
pyflate	423 ms	416 ms: 1.02x faster
python_startup	8.72 ms	8.15 ms: 1.07x faster
python_startup_no_site	6.35 ms	5.97 ms: 1.06x faster
raytrace	312 ms	293 ms: 1.06x faster
regex_compile	139 ms	131 ms: 1.06x faster
regex_dna	180 ms	185 ms: 1.03x slower
regex_effbot	2.99 ms	2.82 ms: 1.06x faster
regex_v8	21.4 ms	20.4 ms: 1.05x faster
richards	48.6 ms	46.3 ms: 1.05x faster
scimark_fft	348 ms	338 ms: 1.03x faster
scimark_lu	120 ms	117 ms: 1.02x faster
scimark_monte_carlo	67.0 ms	65.4 ms: 1.02x faster
scimark_sor	116 ms	113 ms: 1.02x faster
spectral_norm	101 ms	102 ms: 1.01x slower
sqlalchemy_declarative	143 ms	135 ms: 1.06x faster
sqlalchemy_imperative	19.0 ms	17.0 ms: 1.12x faster
sqlite_synth	2.50 us	2.29 us: 1.09x faster
sympy_expand	507 ms	465 ms: 1.09x faster
sympy_integrate	21.7 ms	20.5 ms: 1.06x faster
sympy_sum	176 ms	164 ms: 1.08x faster
sympy_str	311 ms	286 ms: 1.09x faster
telco	7.02 ms	6.36 ms: 1.10x faster
tornado_http	125 ms	113 ms: 1.10x faster
unpickle	15.7 us	15.1 us: 1.04x faster
unpickle_list	4.74 us	4.56 us: 1.04x faster
unpickle_pure_python	229 us	219 us: 1.05x faster
xml_etree_parse	158 ms	155 ms: 1.02x faster
xml_etree_iterparse	103 ms	101 ms: 1.02x faster
xml_etree_generate	91.0 ms	84.3 ms: 1.08x faster
xml_etree_process	61.9 ms	58.4 ms: 1.06x faster
Geometric mean	(ref)	1.05x faster

Benchmark hidden because not significant (3): pickle, scimark_sparse_mat_mult, unpack_sequence

All reactions

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

corona10 self-assigned this

https://gist.github.com/corona10/5726d1528176677d4c694265edfc4bf5

corona10 commented Aug 14, 2022 •

edited

Loading

Uh oh!

There was an error while loading. Please reload this page.

Copy link

Copy Markdown

Member

Another benchmark from Azure VM(Ubuntu 20.04.4 LTS gcc 9.4.0):
https://gist.github.com/corona10/c2aa0108a5ffcc96be449c0ce033412d

But let's measure the benchmark from the Faster CPython machine after the PR is merged.

All reactions

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

corona10 reviewed

Aug 15, 2022

View reviewed changes

Comment thread

Makefile.pre.in Outdated


		bolt-opt: @PREBOLT_RULE@
		rm -f *.fdata
		@LLVM_BOLT@ $(BUILDPYTHON) -instrument -instrumentation-file-append-pid -instrumentation-file=$(abspath $(BUILDPYTHON).bolt) -o $(BUILDPYTHON).bolt_inst

corona10 Aug 15, 2022

Copy link

Copy Markdown

Member

There was a problem hiding this comment.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change

	@LLVM_BOLT@ $(BUILDPYTHON) -instrument -instrumentation-file-append-pid -instrumentation-file=$(abspath $(BUILDPYTHON).bolt) -o $(BUILDPYTHON).bolt_inst
	@LLVM_BOLT@ ./$(BUILDPYTHON) -instrument -instrumentation-file-append-pid -instrumentation-file=$(abspath $(BUILDPYTHON).bolt) -o $(BUILDPYTHON).bolt_inst

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

All reactions

Comment thread

Makefile.pre.in Outdated

		@LLVM_BOLT@ $(BUILDPYTHON) -instrument -instrumentation-file-append-pid -instrumentation-file=$(abspath $(BUILDPYTHON).bolt) -o $(BUILDPYTHON).bolt_inst
		./$(BUILDPYTHON).bolt_inst $(PROFILE_TASK) \|\| true
		@MERGE_FDATA@ $(BUILDPYTHON).*.fdata > $(BUILDPYTHON).fdata
		@LLVM_BOLT@ $(BUILDPYTHON) -o $(BUILDPYTHON).bolt -data=$(BUILDPYTHON).fdata -update-debug-sections -reorder-blocks=ext-tsp -reorder-functions=hfsort+ -split-functions=3 -icf=1 -inline-all -split-eh -reorder-functions-use-hot-size -peepholes=all -jump-tables=aggressive -inline-ap -indirect-call-promotion=all -dyno-stats -use-gnu-stack -frame-opt=hot

corona10 Aug 15, 2022

Copy link

Copy Markdown

Member

There was a problem hiding this comment.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change

	@LLVM_BOLT@ $(BUILDPYTHON) -o $(BUILDPYTHON).bolt -data=$(BUILDPYTHON).fdata -update-debug-sections -reorder-blocks=ext-tsp -reorder-functions=hfsort+ -split-functions=3 -icf=1 -inline-all -split-eh -reorder-functions-use-hot-size -peepholes=all -jump-tables=aggressive -inline-ap -indirect-call-promotion=all -dyno-stats -use-gnu-stack -frame-opt=hot
	@LLVM_BOLT@ ./$(BUILDPYTHON) -o $(BUILDPYTHON).bolt -data=$(BUILDPYTHON).fdata -update-debug-sections -reorder-blocks=ext-tsp -reorder-functions=hfsort+ -split-functions=3 -icf=1 -inline-all -split-eh -reorder-functions-use-hot-size -peepholes=all -jump-tables=aggressive -inline-ap -indirect-call-promotion=all -dyno-stats -use-gnu-stack -frame-opt=hot

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

All reactions

corona10 commented Aug 15, 2022 •

edited

Loading

Uh oh!

There was an error while loading. Please reload this page.

Copy link

Copy Markdown

Member

I success to get cache miss-related metadata and also I got the pyperformance result which is similar to my previous attempts and Kevin's report.
I didn't analyze whether the GCC version or OS version could affect the performance result.
But I can conclude that BOLT definitely makes CPython faster.

Environment

Hardware: AWS c5n.metal
Red Hat Enterprise Linux release 8.6 (Ootpa)
gcc: gcc (GCC) 8.5.0 20210514 (Red Hat 8.5.0-10)
LLVM version 14.0.6

Binary Size

Without BOLT: 79M
With BOLT: 36M

ICache miss

Experiment	instructions	L1-icache-misses	ratio
PGO + LTO	8,330,863,079,932	77,047,357,163	0.92%
PGO + LTO + BOLT	8,312,698,165,975	65,319,225,064	0.79%

Benchmark (1.01x faster)

osevan reacted with thumbs up emoji

leo-smi reacted with rocket emoji

All reactions

👍 1 reaction
🚀 1 reaction

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

kmod and others added 6 commits

August 16, 2022 17:31


          Update Makefile.pre.in

c12dbea

Co-authored-by: Dong-hee Na <[email protected]>


          Update configure.ac

ce25757

Co-authored-by: Dong-hee Na <[email protected]>


          Add myself to ACKS

ded38f0


          Add docs

cc17806


          Other review comments


          fix tab/space issue

83da8c4

aaupov mentioned this pull request

Aug 18, 2022

Optimizing module shared libraries with BOLT faster-cpython/ideas#449

Open

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

4938e5a

Using the patch adapted from python/cpython#95908

vmarkovtsev mentioned this pull request

[DEV-5385] Build CPython using BOLT athenianco/athenian-api#3040

Merged

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

6e7c370

Using the patch adapted from python/cpython#95908

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

25c9972

Using the patch adapted from python/cpython#95908

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

b71c8d7

Using the patch adapted from python/cpython#95908

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

77c21d0

Using the patch adapted from python/cpython#95908

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

f23b3d7

Using the patch adapted from python/cpython#95908

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

a276be0

Using the patch adapted from python/cpython#95908

vmarkovtsev mentioned this pull request

BOLT segfaults on optimizing CPython llvm/llvm-project#59025

Closed

osevan commented Dec 18, 2022 •

edited

Loading

Uh oh!

There was an error while loading. Please reload this page.

Copy link

Copy Markdown

Another question and important view of performance tuning.

Gcc pgo and clang pgo are different , and gcc pgo profiler like profile-generate, can get more deeply data for pgo, instead of clang profile-generate.

So, would be nice to make new flags with

--enable-lto-gcc --enable-pgo-gcc,but considering at gcc level reorder flag needing for BOLT at clang

bolting

And one compilechain completely in clang
--enable-lto-llvm --enable-pgo-llvm plus bolt

Thank you very much

All reactions

Sorry, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

f37d126

Using the patch adapted from python/cpython#95908

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

28da5b8

Using the patch adapted from python/cpython#95908

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request


          [DEV-5385] Build CPython using BOLT

040f302

Using the patch adapted from python/cpython#95908

vmarkovtsev added a commit to athenianco/athenian-api that referenced this pull request