Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Valgrind errors when running batch_rl #71

Open
sofian opened this issue Jul 24, 2013 · 0 comments
Open

Valgrind errors when running batch_rl #71

sofian opened this issue Jul 24, 2013 · 0 comments
Labels

Comments

@sofian
Copy link
Owner

sofian commented Jul 24, 2013

When running batch_rl I get a bunch of "Conditional jump or move depends on uninitialised value(s)" errors that come from a stack allocation at BatchRLTrainer::_doTrainEpisode(DataSet*). They might be linked to an uninitialized value in the DataSet that is passed as an argument to _doTrainEpisode.

==30259== Memcheck, a memory error detector
==30259== Copyright (C) 2002-2012, and GNU GPL'd, by Julian Seward et al.
==30259== Using Valgrind-3.8.1 and LibVEX; rerun with -h for copyright info
==30259== Command: ./build/computer/batch_rl -iter 1 training_set.txt 5 1 2
==30259== 
==30259== Conditional jump or move depends on uninitialised value(s)
==30259==    at 0x403778: float min<float>(float, float) (in /home/tats/Documents/workspace/qualia/examples/batch_rl/build/computer/batch_rl)
==30259==    by 0x403479: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:63)
==30259==    by 0x404DB0: DataSetTrainer::trainEpisode(DataSet*) (DataSetTrainer.cpp:45)
==30259==    by 0x404D09: DataSetTrainer::train(DataSet*, int) (DataSetTrainer.cpp:38)
==30259==    by 0x401C37: main (batch_rl.cpp:155)
==30259==  Uninitialised value was created by a stack allocation
==30259==    at 0x403332: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:41)
==30259== 
==30259== Conditional jump or move depends on uninitialised value(s)
==30259==    at 0x4037A3: float max<float>(float, float) (in /home/tats/Documents/workspace/qualia/examples/batch_rl/build/computer/batch_rl)
==30259==    by 0x4034B3: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:64)
==30259==    by 0x404DB0: DataSetTrainer::trainEpisode(DataSet*) (DataSetTrainer.cpp:45)
==30259==    by 0x404D09: DataSetTrainer::train(DataSet*, int) (DataSetTrainer.cpp:38)
==30259==    by 0x401C37: main (batch_rl.cpp:155)
==30259==  Uninitialised value was created by a stack allocation
==30259==    at 0x403332: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:41)
==30259== 
==30259== Conditional jump or move depends on uninitialised value(s)
==30259==    at 0x4E91845: __ieee754_exp_avx (e_exp.c:70)
==30259==    by 0x4E5A641: exp (w_exp.c:26)
==30259==    by 0x405DE2: NeuralNetwork::_propagateLayer(NeuralNetwork::Layer&, NeuralNetwork::Layer&) (NeuralNetwork.cpp:194)
==30259==    by 0x405AE4: NeuralNetwork::propagate() (NeuralNetwork.cpp:105)
==30259==    by 0x403BF1: QFunction::getValue(Observation const*, Action const*) (QFunction.cpp:53)
==30259==    by 0x40360F: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:97)
==30259==    by 0x404DB0: DataSetTrainer::trainEpisode(DataSet*) (DataSetTrainer.cpp:45)
==30259==    by 0x404D09: DataSetTrainer::train(DataSet*, int) (DataSetTrainer.cpp:38)
==30259==    by 0x401C37: main (batch_rl.cpp:155)
==30259==  Uninitialised value was created by a stack allocation
==30259==    at 0x403332: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:41)
==30259== 

...

==30259== Use of uninitialised value of size 8
==30259==    at 0x4E9191A: __ieee754_exp_avx (e_exp.c:91)
==30259==    by 0x4E5A641: exp (w_exp.c:26)
==30259==    by 0x405DE2: NeuralNetwork::_propagateLayer(NeuralNetwork::Layer&, NeuralNetwork::Layer&) (NeuralNetwork.cpp:194)
==30259==    by 0x405AE4: NeuralNetwork::propagate() (NeuralNetwork.cpp:105)
==30259==    by 0x403BF1: QFunction::getValue(Observation const*, Action const*) (QFunction.cpp:53)
==30259==    by 0x40360F: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:97)
==30259==    by 0x404DB0: DataSetTrainer::trainEpisode(DataSet*) (DataSetTrainer.cpp:45)
==30259==    by 0x404D09: DataSetTrainer::train(DataSet*, int) (DataSetTrainer.cpp:38)
==30259==    by 0x401C37: main (batch_rl.cpp:155)
==30259==  Uninitialised value was created by a stack allocation
==30259==    at 0x403332: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:41)
==30259== 

...

==30259== 
==30259== Conditional jump or move depends on uninitialised value(s)
==30259==    at 0x56A1FA7: __printf_fp (printf_fp.c:982)
==30259==    by 0x56A0B4B: vfprintf (vfprintf.c:1654)
==30259==    by 0x56C3AA3: vsprintf (iovsprintf.c:42)
==30259==    by 0x40309F: Q_MESSAGE(char const*, ...) (error.cpp:127)
==30259==    by 0x4036E6: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:112)
==30259==    by 0x404DB0: DataSetTrainer::trainEpisode(DataSet*) (DataSetTrainer.cpp:45)
==30259==    by 0x404D09: DataSetTrainer::train(DataSet*, int) (DataSetTrainer.cpp:38)
==30259==    by 0x401C37: main (batch_rl.cpp:155)
==30259==  Uninitialised value was created by a stack allocation
==30259==    at 0x403332: BatchRLTrainer::_doTrainEpisode(DataSet*) (BatchRLTrainer.cpp:41)
==30259== 

...

==30259== 
==30259== 
==30259== HEAP SUMMARY:
==30259==     in use at exit: 20 bytes in 3 blocks
==30259==   total heap usage: 234 allocs, 231 frees, 889,788 bytes allocated
==30259== 
==30259== LEAK SUMMARY:
==30259==    definitely lost: 20 bytes in 3 blocks
==30259==    indirectly lost: 0 bytes in 0 blocks
==30259==      possibly lost: 0 bytes in 0 blocks
==30259==    still reachable: 0 bytes in 0 blocks
==30259==         suppressed: 0 bytes in 0 blocks
==30259== Rerun with --leak-check=full to see details of leaked memory
==30259== 
==30259== For counts of detected and suppressed errors, rerun with: -v
==30259== ERROR SUMMARY: 901525 errors from 155 contexts (suppressed: 2 from 2)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

1 participant