GH-40394: [C++] Add support for mold by kou · Pull Request #40397 · apache/arrow

kou · 2024-03-07T06:56:08Z

Rationale for this change

mold is a faster linker. It will reduce build time.

mold supports only ELF. So it's available on only Linux and *BSD.

What changes are included in this PR?

Add new ARROW_USE_MOLD CMake option. It's OFF by default because it doesn't work with conda.

Are these changes tested?

Yes.

Are there any user-facing changes?

No.

GitHub Issue: [C++] Add support for mold #40394

kou · 2024-03-07T06:56:31Z

@github-actions crossbow submit -g cpp -g linux

github-actions · 2024-03-07T06:56:37Z

⚠️ GitHub issue #40394 has been automatically assigned in GitHub to PR creator.

pitrou · 2024-03-07T10:22:25Z

I get the following warnings when building:

[407/859] Linking CXX shared library debug/libarrow.so.1600.0.0
mold: warning: /home/antoine/arrow/dev/cpp/src/arrow/symbols.map: cannot assign version `global` to symbol `descriptor_table_Flight_2eproto`: symbol not found
mold: warning: /home/antoine/arrow/dev/cpp/src/arrow/symbols.map: cannot assign version `global` to symbol `descriptor_table_FlightSql_2eproto`: symbol not found
[591/859] Linking CXX shared library debug/libarrow_cuda.so.1600.0.0
mold: warning: /home/antoine/arrow/dev/cpp/src/arrow/symbols.map: cannot assign version `global` to symbol `descriptor_table_Flight_2eproto`: symbol not found
mold: warning: /home/antoine/arrow/dev/cpp/src/arrow/symbols.map: cannot assign version `global` to symbol `descriptor_table_FlightSql_2eproto`: symbol not found

and then I get various kinds of crashes when running the tests.

pitrou · 2024-03-07T10:26:38Z

arrow-acero-fetch-node-test throws an exception here:

#0  __pthread_kill_implementation (no_tid=0, signo=6, threadid=140737222772352) at ./nptl/pthread_kill.c:44
#1  __pthread_kill_internal (signo=6, threadid=140737222772352) at ./nptl/pthread_kill.c:78
#2  __GI___pthread_kill (threadid=140737222772352, signo=signo@entry=6) at ./nptl/pthread_kill.c:89
#3  0x00007ffff4314476 in __GI_raise (sig=sig@entry=6) at ../sysdeps/posix/raise.c:26
#4  0x00007ffff42fa7f3 in __GI_abort () at ./stdlib/abort.c:79
#5  0x00007ffff45f3f00 in __gnu_cxx::__verbose_terminate_handler () at ../../../../libstdc++-v3/libsupc++/vterminate.cc:95
#6  0x00007ffff45f243c in __cxxabiv1::__terminate (handler=<optimized out>) at ../../../../libstdc++-v3/libsupc++/eh_terminate.cc:48
#7  0x00007ffff45f248e in std::terminate () at ../../../../libstdc++-v3/libsupc++/eh_terminate.cc:58
#8  0x00007ffff45f2680 in __cxxabiv1::__cxa_throw (obj=0x55555561cd70, tinfo=0x7ffff47108d8 <typeinfo for std::system_error>, 
    dest=0x7ffff460fa28 <std::system_error::~system_error()>) at ../../../../libstdc++-v3/libsupc++/eh_throw.cc:98
#9  0x00007ffff45ec92b in std::__throw_system_error (__i=-1)
    at /home/conda/feedstock_root/build_artifacts/gcc_compilers_1706816862910/work/build/x86_64-conda-linux-gnu/libstdc++-v3/include/system_error:233
#10 0x000055555559dcc3 in std::call_once<arrow::acero::RegisterTestNodes()::<lambda()> >(std::once_flag &, struct {...} &&) (__once=..., __f=...)
    at /home/antoine/mambaforge/envs/pyarrow/x86_64-conda-linux-gnu/include/c++/12.3.0/mutex:860
#11 0x000055555559ac67 in arrow::acero::RegisterTestNodes () at /home/antoine/arrow/dev/cpp/src/arrow/acero/test_nodes.cc:360
#12 0x00005555555cc53a in arrow::acero::CheckFetch (options=...) at /home/antoine/arrow/dev/cpp/src/arrow/acero/fetch_node_test.cc:43
#13 0x00005555555cd976 in arrow::acero::FetchNode_Basic_Test::TestBody (this=0x55555561d320) at /home/antoine/arrow/dev/cpp/src/arrow/acero/fetch_node_test.cc:75
#14 0x00007ffff7d5f6d6 in testing::internal::HandleExceptionsInMethodIfSupported<testing::Test, void> (object=0x55555561d320, method=&virtual testing::Test::TestBody(), 
    location=0x7ffff7d26f83 "the test body") at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2665
#15 0x00007ffff7d39b6f in testing::Test::Run (this=0x55555561d320) at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2682
#16 0x00007ffff7d3a625 in testing::TestInfo::Run (this=0x55555561eba0) at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2861
#17 0x00007ffff7d3b011 in testing::TestSuite::Run (this=0x55555561d4b0) at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:3015
#18 0x00007ffff7d4b7d9 in testing::internal::UnitTestImpl::RunAllTests (this=0x55555561ca50) at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:5855
#19 0x00007ffff7d608c4 in testing::internal::HandleExceptionsInMethodIfSupported<testing::internal::UnitTestImpl, bool> (object=0x55555561ca50, 
    method=(bool (testing::internal::UnitTestImpl::*)(class testing::internal::UnitTestImpl * const)) 0x7ffff7d4b3c2 <testing::internal::UnitTestImpl::RunAllTests()>, 
    location=0x7ffff7d279e8 "auxiliary test code (environments or event listeners)") at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2665
#20 0x00007ffff7d49be3 in testing::UnitTest::Run (this=0x7ffff7d80ec0 <testing::UnitTest::GetInstance()::instance>)
    at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:5438
#21 0x00007ffff7d82a02 in RUN_ALL_TESTS () at /build/build-test/_deps/googletest-src/googletest/include/gtest/gtest.h:2490
#22 0x00007ffff7d82a4b in main (argc=1, argv=0x7fffffffc0f8) at /build/build-test/_deps/googletest-src/googletest/src/gtest_main.cc:52

pitrou · 2024-03-07T10:27:51Z

arrow-acero-util-test hangs here:

#0  __futex_abstimed_wait_common64 (private=0, cancel=true, abstime=0x0, op=393, expected=0, futex_word=0x555555621c20) at ./nptl/futex-internal.c:57
#1  __futex_abstimed_wait_common (cancel=true, private=0, abstime=0x0, clockid=0, expected=0, futex_word=0x555555621c20) at ./nptl/futex-internal.c:87
#2  __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x555555621c20, expected=expected@entry=0, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#3  0x00007ffff4365a41 in __pthread_cond_wait_common (abstime=0x0, clockid=0, mutex=0x555555621b70, cond=0x555555621bf8) at ./nptl/pthread_cond_wait.c:503
#4  ___pthread_cond_wait (cond=0x555555621bf8, mutex=0x555555621b70) at ./nptl/pthread_cond_wait.c:627
#5  0x00007ffff7421774 in std::condition_variable::wait<arrow::internal::ThreadPool::WaitForIdle()::<lambda()> >(std::unique_lock<std::mutex> &, struct {...}) (this=0x555555621bf8, __lock=..., __p=...) at /home/antoine/mambaforge/envs/pyarrow/x86_64-conda-linux-gnu/include/c++/12.3.0/condition_variable:102
#6  0x00007ffff741f612 in arrow::internal::ThreadPool::WaitForIdle (this=0x55555561e0d0) at /home/antoine/arrow/dev/cpp/src/arrow/util/thread_pool.cc:497
#7  0x00005555555cad6a in arrow::acero::MakePrimedThreadPool (num_threads=11) at /home/antoine/arrow/dev/cpp/src/arrow/acero/task_util_test.cc:67
#8  0x00005555555ccf33 in arrow::acero::TaskScheduler_StressTwo_Test::TestBody (this=0x55555561a930) at /home/antoine/arrow/dev/cpp/src/arrow/acero/task_util_test.cc:193
#9  0x00007ffff7d686d1 in testing::internal::HandleSehExceptionsInMethodIfSupported<testing::Test, void> (object=0x55555561a930, method=&virtual testing::Test::TestBody(), location=0x7ffff7d26f83 "the test body") at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2607
#10 0x00007ffff7d5f691 in testing::internal::HandleExceptionsInMethodIfSupported<testing::Test, void> (object=0x55555561a930, method=&virtual testing::Test::TestBody(), location=0x7ffff7d26f83 "the test body") at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2643
#11 0x00007ffff7d39b6f in testing::Test::Run (this=0x55555561a930) at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2682
#12 0x00007ffff7d3a625 in testing::TestInfo::Run (this=0x55555561c2f0) at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2861
#13 0x00007ffff7d3b011 in testing::TestSuite::Run (this=0x55555560b180) at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:3015
#14 0x00007ffff7d4b7d9 in testing::internal::UnitTestImpl::RunAllTests (this=0x55555561aa50) at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:5855
#15 0x00007ffff7d69690 in testing::internal::HandleSehExceptionsInMethodIfSupported<testing::internal::UnitTestImpl, bool> (object=0x55555561aa50, method=(bool (testing::internal::UnitTestImpl::*)(class testing::internal::UnitTestImpl * const)) 0x7ffff7d4b3c2 <testing::internal::UnitTestImpl::RunAllTests()>, location=0x7ffff7d279e8 "auxiliary test code (environments or event listeners)") at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2607
#16 0x00007ffff7d60878 in testing::internal::HandleExceptionsInMethodIfSupported<testing::internal::UnitTestImpl, bool> (object=0x55555561aa50, method=(bool (testing::internal::UnitTestImpl::*)(class testing::internal::UnitTestImpl * const)) 0x7ffff7d4b3c2 <testing::internal::UnitTestImpl::RunAllTests()>, location=0x7ffff7d279e8 "auxiliary test code (environments or event listeners)") at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:2643
#17 0x00007ffff7d49be3 in testing::UnitTest::Run (this=0x7ffff7d80ec0 <testing::UnitTest::GetInstance()::instance>) at /build/build-test/_deps/googletest-src/googletest/src/gtest.cc:5438
#18 0x00007ffff7d82a02 in RUN_ALL_TESTS () at /build/build-test/_deps/googletest-src/googletest/include/gtest/gtest.h:2490
#19 0x00007ffff7d82a4b in main (argc=1, argv=0x7fffffffc108) at /build/build-test/_deps/googletest-src/googletest/src/gtest_main.cc:52

pitrou · 2024-03-07T10:28:11Z

I think the bottom line is that, for now, we cannot enable mold by default.

kou · 2024-03-07T13:27:05Z

The warnings are caused by sharing cpp/src/arrow/symbols.map with libarrow*.so.
descriptor_table_Flight_2eproto exists only in libarrow_flight.so and descriptor_table_FlightSql_2eproto exists only in libarrow_flight_sql.so.
I used * to suppress the warnings.

kou · 2024-03-07T13:28:52Z

These test problems aren't happen on my environment. (g++ (Debian 13.2.0-13) 13.2.0)
But mold doesn't work with conda. I'll disable this by default.

mold doesn't work with conda.

kou · 2024-03-07T13:32:25Z

@github-actions crossbow submit -g cpp -g linux

github-actions · 2024-03-07T13:35:14Z

Revision: c24db80

Submitted crossbow builds: ursacomputing/crossbow @ actions-cbd198594d

Task	Status
almalinux-8-amd64
almalinux-8-arm64
almalinux-9-amd64
almalinux-9-arm64
amazon-linux-2023-amd64
amazon-linux-2023-arm64
centos-7-amd64
centos-8-stream-amd64
centos-8-stream-arm64
centos-9-stream-amd64
centos-9-stream-arm64
debian-bookworm-amd64
debian-bookworm-arm64
debian-bullseye-amd64
debian-bullseye-arm64
debian-trixie-amd64
debian-trixie-arm64
test-alpine-linux-cpp
test-build-cpp-fuzz
test-conda-cpp
test-conda-cpp-valgrind
test-cuda-cpp
test-debian-11-cpp-amd64
test-debian-11-cpp-i386
test-fedora-39-cpp
test-ubuntu-20.04-cpp
test-ubuntu-20.04-cpp-bundled
test-ubuntu-20.04-cpp-minimal-with-formats
test-ubuntu-20.04-cpp-thread-sanitizer
test-ubuntu-22.04-cpp
test-ubuntu-22.04-cpp-20
test-ubuntu-22.04-cpp-no-threading
test-ubuntu-24.04-cpp
test-ubuntu-24.04-cpp-gcc-14
ubuntu-focal-amd64
ubuntu-focal-arm64
ubuntu-jammy-amd64
ubuntu-jammy-arm64
ubuntu-noble-amd64
ubuntu-noble-arm64

kou · 2024-03-08T05:35:51Z

Hmm. It seems that mold doesn't work with sanitizers. I disabled ARROW_USE_MOLD if one of ARROW_USE_{ASAN,TSAN,UBSAN} is specified.

cpp/cmake_modules/DefineOptions.cmake

kou · 2024-03-11T01:52:13Z

I'll merge this tomorrow if nobody objects this.

assignUser

Cool!

conbench-apache-arrow · 2024-03-13T09:26:21Z

After merging your PR, Conbench analyzed the 7 benchmarking runs that have been run so far on merge-commit 60dbaab.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 4 possible false positives for unstable benchmarks that are known to sometimes produce them.

apacheGH-40394: [C++] Add support for mold

ee81123

kou requested review from assignUser and raulcd as code owners March 7, 2024 06:56

github-actions bot added Component: C++ awaiting committer review Awaiting committer review labels Mar 7, 2024

This comment was marked as outdated.

Sign in to view

kou added 2 commits March 7, 2024 22:24

Suppress warnings

2cd39cd

Use --ld-path for clang

770696a

Don't enable mold by default

c24db80

mold doesn't work with conda.

Don't use mold with sanitizer checks

b27fd84

Fix lint error

c3dbedf

assignUser reviewed Mar 8, 2024

View reviewed changes

cpp/cmake_modules/DefineOptions.cmake Outdated Show resolved Hide resolved

Add warning message for ARROW_USE_MOLD and ARROW_USE{ASAN,TSAN,UBSAN}

5bd138f

github-actions bot added awaiting changes Awaiting changes and removed awaiting committer review Awaiting committer review labels Mar 9, 2024

assignUser approved these changes Mar 13, 2024

View reviewed changes

assignUser merged commit 60dbaab into apache:main Mar 13, 2024

assignUser removed the awaiting changes Awaiting changes label Mar 13, 2024

assignUser mentioned this pull request Mar 13, 2024

[C++] Add support for mold #40394

Closed

assignUser linked an issue Mar 13, 2024 that may be closed by this pull request

[C++] Add support for mold #40394

Closed

github-actions bot added the awaiting committer review Awaiting committer review label Mar 13, 2024

kou deleted the cpp-mold branch March 13, 2024 04:36

eitsupi mentioned this pull request Aug 19, 2025

Use mold if available pola-rs/r-polars#1508

Closed

Conversation

kou commented Mar 7, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

Uh oh!

kou commented Mar 7, 2024

Uh oh!

github-actions bot commented Mar 7, 2024

Uh oh!

This comment was marked as outdated.

pitrou commented Mar 7, 2024

Uh oh!

pitrou commented Mar 7, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pitrou commented Mar 7, 2024

Uh oh!

pitrou commented Mar 7, 2024

Uh oh!

kou commented Mar 7, 2024

Uh oh!

kou commented Mar 7, 2024

Uh oh!

kou commented Mar 7, 2024

Uh oh!

github-actions bot commented Mar 7, 2024

Uh oh!

kou commented Mar 8, 2024

Uh oh!

Uh oh!

kou commented Mar 11, 2024

Uh oh!

assignUser left a comment

Choose a reason for hiding this comment

Uh oh!

conbench-apache-arrow bot commented Mar 13, 2024

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

kou commented Mar 7, 2024 •

edited

Loading

pitrou commented Mar 7, 2024 •

edited

Loading