Failed tests -- what might be the build problem?

690 views
Skip to first unread message

Thomas Sharpless

unread,
Jan 29, 2016, 1:03:42 PM1/29/16
to Ceres Solver
Building latest ceres solver release on Win10 with MSVC12 (VS2013 Community).
Using current Eigen distro + LAPACK & BLAS prebuilt from http://www.fi.muni.cz/~xsvobod2/misc/lapack/
Not using glog (would like to but I can't get it to link though it seems to build OK).

Everything builds without error but RUN_TESTS reports 15 failures, all of which seem to involve solver algorithms.
(summary below) The reported error in all cases was  ***Exception: Other  0.02 sec.

I hope someone more familiar with building ceres can recognize what might have gone wrong, and suggest how I might build it right.  My intended application is small scale bundle adjustment; so I don't care much about performance or capacity optimization -- I just need a solver that works.
1>  77% tests passed, 15 tests failed out of 66
1>  
1>  Total Test time (real) = 149.26 sec
1>  
1>  The following tests FAILED:
1>   10 - bundle_adjustment_test (OTHER_FAULT)
1>   11 - c_api_test (OTHER_FAULT)
1>   28 - gradient_problem_solver_test (OTHER_FAULT)
1>   33 - iterative_schur_complement_solver_test (OTHER_FAULT)
1>   35 - levenberg_marquardt_strategy_test (OTHER_FAULT)
1>   36 - dogleg_strategy_test (OTHER_FAULT)
1>   37 - line_search_preprocessor_test (OTHER_FAULT)
1>   40 - minimizer_test (OTHER_FAULT)
1>   54 - schur_complement_solver_test (OTHER_FAULT)
1>   58 - solver_test (OTHER_FAULT)
1>   59 - symmetric_linear_solver_test (OTHER_FAULT)
1>   61 - trust_region_minimizer_test (OTHER_FAULT)
1>   62 - trust_region_preprocessor_test (OTHER_FAULT)
1>   63 - unsymmetric_linear_solver_test (OTHER_FAULT)
1>   66 - system_test (OTHER_FAULT)
1>  Errors while running CTest
 

Sameer Agarwal

unread,
Jan 29, 2016, 1:06:45 PM1/29/16
to Ceres Solver
Thomas,
Can you run one the failing tests from the command line and share the output. 
Say symmetric_linear_solver_test ?
Sameer


--
You received this message because you are subscribed to the Google Groups "Ceres Solver" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ceres-solver...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ceres-solver/d78db4f9-34ce-4446-bc0f-a0f843dd142b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Thomas Sharpless

unread,
Jan 29, 2016, 1:26:27 PM1/29/16
to Ceres Solver
Sameer,
I'd be happy to do that but have no clue how.  Can you give brief instructions?
-- Tom

Sameer Agarwal

unread,
Jan 29, 2016, 1:27:26 PM1/29/16
to Ceres Solver
there should be a bin directory containing a binary called symmetric_linear_solver_test that you should be able run from the command line.
Sameer


--
You received this message because you are subscribed to the Google Groups "Ceres Solver" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ceres-solver...@googlegroups.com.

Thomas Sharpless

unread,
Jan 29, 2016, 1:45:50 PM1/29/16
to ceres-...@googlegroups.com
OK, found that, ran it, here's the output
C:\Users\tks\Projects\ceres-solver-master\build2-msvc2013\bin\Debug>symmetric_linear_solver_test.exe
[==========] Running 2 tests from 2 test cases.
[----------] Global test environment set-up.
[----------] 1 test from ConjugateGradientTest
[ RUN      ] ConjugateGradientTest.Solves3x3IdentitySystem
[       OK ] ConjugateGradientTest.Solves3x3IdentitySystem (0 ms)
[----------] 1 test from ConjugateGradientTest (1 ms total)

[----------] 1 test from ConjuateGradientTest
[ RUN      ] ConjuateGradientTest.Solves3x3SymmetricSystem
[       OK ] ConjuateGradientTest.Solves3x3SymmetricSystem (0 ms)
[----------] 1 test from ConjuateGradientTest (1 ms total)

[----------] Global test environment tear-down
[==========] 2 tests from 2 test cases ran. (9 ms total)
[  PASSED  ] 2 tests.


--
You received this message because you are subscribed to a topic in the Google Groups "Ceres Solver" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ceres-solver/6Y3NMH1_Xqc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ceres-solver...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ceres-solver/CABqdRUA%3DO7MYG2X4HH4AXD5%3D-wbXqoGXH9YoxrZVW7YigCnRng%40mail.gmail.com.

Sameer Agarwal

unread,
Jan 29, 2016, 1:46:35 PM1/29/16
to ceres-...@googlegroups.com, alex...@gmail.com
okay thats weird, because that test was failing in the log you sent out.
+Alex Stewart should have something to say here. 

Thomas Sharpless

unread,
Jan 29, 2016, 1:49:15 PM1/29/16
to ceres-...@googlegroups.com
And this from iterative_schur_complement_solver_test.  Does this mean ceres is OK and the fault is really in RUN_TESTS?

C:\Users\tks\Projects\ceres-solver-master\build2-msvc2013\bin\Debug>iterative_schur_complement_solver_test
[==========] Running 2 tests from 1 test case.
[----------] Global test environment set-up.
[----------] 2 tests from IterativeSchurComplementSolverTest
[ RUN      ] IterativeSchurComplementSolverTest.NormalProblem
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\block_sparse_matrix.cc:80 Allocating values array with 104 bytes.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\wall_time.cc:74

DenseQRSolver::Solve
                                   Delta   Cumulative
                           Setup :    0.00015      0.00015
                           Solve :    0.00041      0.00056
                        TearDown :    0.00001      0.00057
                           Total :    0.00001      0.00058


C:\Users\tks\Projects\ceres-solver-master\internal\ceres\detect_structure.cc:113 Schur complement static structure <1,1,1>.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\partitioned_matrix_view.cc:176 Template specializations not found for <1,1,1>
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\block_sparse_matrix.cc:80 Allocating values array with 16 bytes.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\block_random_access_diagonal_matrix.cc:66 Matrix Size [3,3] 3
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\schur_eliminator.cc:155 Template specializations not found for <1,1,1>
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\wall_time.cc:74

IterativeSchurComplementSolver::Solve
                                   Delta   Cumulative
                           Setup :    0.04588      0.04588
                           Solve :    0.00021      0.04610
                           Total :    0.00002      0.04611


C:\Users\tks\Projects\ceres-solver-master\internal\ceres\wall_time.cc:74

DenseQRSolver::Solve
                                   Delta   Cumulative
                           Setup :    0.00007      0.00007
                           Solve :    0.00040      0.00047
                        TearDown :    0.00002      0.00049
                           Total :    0.00001      0.00050


C:\Users\tks\Projects\ceres-solver-master\internal\ceres\detect_structure.cc:113 Schur complement static structure <1,1,1>.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\partitioned_matrix_view.cc:176 Template specializations not found for <1,1,1>
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\block_sparse_matrix.cc:80 Allocating values array with 16 bytes.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\block_random_access_diagonal_matrix.cc:66 Matrix Size [3,3] 3
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\schur_eliminator.cc:155 Template specializations not found for <1,1,1>
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\wall_time.cc:74

IterativeSchurComplementSolver::Solve
                                   Delta   Cumulative
                           Setup :    0.05039      0.05039
                           Solve :    0.00029      0.05068
                           Total :    0.00002      0.05070


[       OK ] IterativeSchurComplementSolverTest.NormalProblem (194 ms)
[ RUN      ] IterativeSchurComplementSolverTest.ProblemWithNoFBlocks
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\block_sparse_matrix.cc:80 Allocating values array with 40 bytes.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\wall_time.cc:74

DenseQRSolver::Solve
                                   Delta   Cumulative
                           Setup :    0.00002      0.00002
                           Solve :    0.00014      0.00016
                        TearDown :    0.00001      0.00018
                           Total :    0.00001      0.00019


C:\Users\tks\Projects\ceres-solver-master\internal\ceres\detect_structure.cc:113 Schur complement static structure <1,1,0>.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\partitioned_matrix_view.cc:176 Template specializations not found for <1,1,0>
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\block_sparse_matrix.cc:80 Allocating values array with 16 bytes.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\iterative_schur_complement_solver.cc:88 No parameter blocks left in the schur complement.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\wall_time.cc:74

IterativeSchurComplementSolver::Solve
                                   Delta   Cumulative
                           Total :    0.04644      0.04644


C:\Users\tks\Projects\ceres-solver-master\internal\ceres\wall_time.cc:74

DenseQRSolver::Solve
                                   Delta   Cumulative
                           Setup :    0.00014      0.00014
                           Solve :    0.00038      0.00052
                        TearDown :    0.00003      0.00055
                           Total :    0.00002      0.00057


C:\Users\tks\Projects\ceres-solver-master\internal\ceres\detect_structure.cc:113 Schur complement static structure <1,1,0>.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\partitioned_matrix_view.cc:176 Template specializations not found for <1,1,0>
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\block_sparse_matrix.cc:80 Allocating values array with 16 bytes.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\iterative_schur_complement_solver.cc:88 No parameter blocks left in the schur complement.
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\wall_time.cc:74

IterativeSchurComplementSolver::Solve
                                   Delta   Cumulative
                           Total :    0.04535      0.04535


[       OK ] IterativeSchurComplementSolverTest.ProblemWithNoFBlocks (189 ms)
[----------] 2 tests from IterativeSchurComplementSolverTest (391 ms total)

[----------] Global test environment tear-down
[==========] 2 tests from 1 test case ran. (397 ms total)
[  PASSED  ] 2 tests.

C:\Users\tks\Projects\ceres-solver-master\build2-msvc2013\bin\Debug>

Thomas Sharpless

unread,
Jan 29, 2016, 1:56:27 PM1/29/16
to ceres-...@googlegroups.com, alex...@gmail.com
Well, everything that happens inside Visual Studio is more or less weird :)


Alex Stewart

unread,
Jan 29, 2016, 2:03:47 PM1/29/16
to Thomas Sharpless, ceres-...@googlegroups.com
Tom,

What is the contents of:

<CERES_BUILD_DIR>/Testing/Temporary/LastTest.log

?  It should contain the output from all tests after you run make test.

-Alex

Thomas Sharpless

unread,
Jan 29, 2016, 2:46:50 PM1/29/16
to Alex Stewart, ceres-...@googlegroups.com
Hi Alex

That log shows no output from any of the failed tests.  Example:

35/66 Testing: levenberg_marquardt_strategy_test
35/66 Test: levenberg_marquardt_strategy_test
Command: "C:/Users/tks/Projects/ceres-solver-master/build2-msvc2013/bin/Debug/levenberg_marquardt_strategy_test.exe" "--test_srcdir" "C:/Users/tks/Projects/ceres-solver-master/data"
Directory: C:/Users/tks/Projects/ceres-solver-master/build2-msvc2013/internal/ceres
"levenberg_marquardt_strategy_test" start time: Jan 29 12:38 Eastern Standard Time
Output:
----------------------------------------------------------
<end of output>
Test time =   0.02 sec
----------------------------------------------------------
Test Failed.
"levenberg_marquardt_strategy_test" end time: Jan 29 12:38 Eastern Standard Time
"levenberg_marquardt_strategy_test" time elapsed: 00:00:00
----------------------------------------------------------

Thomas Sharpless

unread,
Jan 29, 2016, 2:49:38 PM1/29/16
to Alex Stewart, ceres-...@googlegroups.com
But running the same command from the CLI gave this:

C:\Users\tks\Projects\ceres-solver-master\build2-msvc2013\bin\Debug>levenberg_marquardt_strategy_test.exe "--test_srcdir" "C:/Users/tks/Projects/ceres-solver-master/data"
[==========] Running 2 tests from 1 test case.
[----------] Global test environment set-up.
[----------] 2 tests from LevenbergMarquardtStrategy
[ RUN      ] LevenbergMarquardtStrategy.AcceptRejectStepRadiusScaling
[       OK ] LevenbergMarquardtStrategy.AcceptRejectStepRadiusScaling (0 ms)
[ RUN      ] LevenbergMarquardtStrategy.CorrectDiagonalToLinearSolver
C:\Users\tks\Projects\ceres-solver-master\internal\ceres\levenberg_marquardt_strategy.cc:114 Linear solver failure. Failed to compute a step:
[       OK ] LevenbergMarquardtStrategy.CorrectDiagonalToLinearSolver (13 ms)
[----------] 2 tests from LevenbergMarquardtStrategy (28 ms total)

[----------] Global test environment tear-down
[==========] 2 tests from 1 test case ran. (41 ms total)
[  PASSED  ] 2 tests.

Reply all
Reply to author
Forward
0 new messages