Hey @yurivict,

Thanks for the report. I also can't reproduce this one, so the issue template would have been really helpful: what's the Armadillo version? What's the Boost version? What's the compiler version? When I run the tests with valgrind, both on 3.2.2 and on the master branch, I get:

==3865714== Memcheck, a memory error detector
==3865714== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==3865714== Using Valgrind-3.15.0 and LibVEX; rerun with -h for copyright info
==3865714== Command: bin/mlpack_test -t LinearSVMTest
==3865714==
Running 20 test cases...

*** No errors detected
==3865714==
==3865714== HEAP SUMMARY:
==3865714==     in use at exit: 84,304 bytes in 2,765 blocks
==3865714==   total heap usage: 19,308,675 allocs, 19,305,910 frees, 1,111,205,703 bytes allocated
==3865714==
==3865714== 2,576 bytes in 7 blocks are possibly lost in loss record 2,758 of 2,759
==3865714==    at 0x4837B65: calloc (vg_replace_malloc.c:762)
==3865714==    by 0x40118B1: allocate_dtv (dl-tls.c:286)
==3865714==    by 0x40121FD: _dl_allocate_tls (dl-tls.c:532)
==3865714==    by 0x4AF1BDD: allocate_stack (allocatestack.c:621)
==3865714==    by 0x4AF1BDD: pthread_create@@GLIBC_2.2.5 (pthread_create.c:669)
==3865714==    by 0x4FE4F65: ??? (in /usr/lib/x86_64-linux-gnu/libgomp.so.1.0.0)
==3865714==    by 0x4FDBA3C: GOMP_parallel (in /usr/lib/x86_64-linux-gnu/libgomp.so.1.0.0)
==3865714==    by 0x1A7F748: std::enable_if<ens::IsArmaType<arma::SpMat<double> >::value, arma::Mat<double>::elem_type>::type ens::ParallelSGD<ens::ConstantStep>::Optimize<mlpack::svm::LinearSVMFunction<arma::Mat<double> >, arma::Mat<double>, arma::SpMat<double>>(mlpack::svm::LinearSVMFunction<arma::Mat<double> >&, arma::Mat<double>&) (in /home/ryan/src/mlpack-3.2.2/build/bin/mlpack_test)
==3865714==    by 0x1A7FB04: double mlpack::svm::LinearSVM<arma::Mat<double> >::Train<ens::ParallelSGD<ens::ConstantStep> >(arma::Mat<double> const&, arma::Row<unsigned long> const&, unsigned long, ens::ParallelSGD<ens::ConstantStep>) (in /home/ryan/src/mlpack-3.2.2/build/bin/mlpack_test)
==3865714==    by 0xE57B90: LinearSVMTest::LinearSVMPSGDSimpleTest::test_method() (in /home/ryan/src/mlpack-3.2.2/build/bin/mlpack_test)
==3865714==    by 0x1A6C9D8: LinearSVMTest::LinearSVMPSGDSimpleTest_invoker() (in /home/ryan/src/mlpack-3.2.2/build/bin/mlpack_test)
==3865714==    by 0x4BEC48D: boost::detail::function::function_obj_invoker0<boost::detail::forward, int>::invoke(boost::detail::function::function_buffer&) (in /usr/lib/x86_64-linux-gnu/libboost_unit_test_framework.so.1.67.0)
==3865714==    by 0x4BEC0C4: boost::execution_monitor::catch_signals(boost::function<int ()> const&) (in /usr/lib/x86_64-linux-gnu/libboost_unit_test_framework.so.1.67.0)
==3865714==
==3865714== LEAK SUMMARY:
==3865714==    definitely lost: 0 bytes in 0 blocks
==3865714==    indirectly lost: 0 bytes in 0 blocks
==3865714==      possibly lost: 2,576 bytes in 7 blocks
==3865714==    still reachable: 81,728 bytes in 2,758 blocks
==3865714==         suppressed: 0 bytes in 0 blocks
==3865714== Reachable blocks (those to which a pointer was found) are not shown.
==3865714== To see them, rerun with: --leak-check=full --show-leak-kinds=all
==3865714==
==3865714== For lists of detected and suppressed errors, rerun with: -s
==3865714== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)

and that one error I believe is resolved in the latest ensmallen version (and doesn't seem to be related to the issue you are seeing anyway).

If I can get enough information to accurately reproduce it, it's probably not hard to debug, but right now I don't have enough to go on. :(

Uh oh!

tests hang #2300

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions