GMTK segfault occurs when using --resolution option

[EricR86]

Original report (BitBucket issue) by Eric Roberts (Bitbucket: ericr86, GitHub: ericr86).

---

From Google Code Issue #26

Imported Labels: bug, imported, Priority-High, Usability

From epa...@uw.edu on September 26, 2012 19:14:22

Error is described more here https://j.ee.washington.edu/trac/gmtk/ticket/343 .

The output/e//emt files will have an error similar to: segway-1.2.0.dev_r9225-py2.7.egg/segway/data/segway-wrapper.sh: line 38: 3501 Segmentation fault (core dumped) "$@" > "$stdout_filename"

Original issue: http://code.google.com/p/segway-genome/issues/detail?id=26

[EricR86]

Original comment by Eric Roberts (Bitbucket: ericr86, GitHub: ericr86).

---

From hoffman...@gmail.com on September 28, 2012 11:06:36

Paul will patch GMTK to fix this bug.

Owner: epa...@cs.washington.edu
Labels: -Priority-Medium Priority-High Usability

[EricR86]

Original comment by Eric Roberts (Bitbucket: ericr86, GitHub: ericr86).

---

From epa...@cs.washington.edu on March 25, 2013 13:33:05

See also: https://j.ee.washington.edu/trac/gmtk/ticket/6

[EricR86]

Original comment by Michael Hoffman (Bitbucket: hoffman, GitHub: michaelmhoffman).

---

@ericr86 can you verify whether this is still an issue?

[EricR86]

Original comment by Eric Roberts (Bitbucket: ericr86, GitHub: ericr86).

---

• Edited issue description* set assignee to "ericr86 (Bitbucket: ericr86, GitHub: ericr86)"
• set assignee_account_id to "557058:c80ca578-03a1-4ac6-b3ee-50372a3fceee"

[EricR86]

Original comment by Eric Roberts (Bitbucket: ericr86, GitHub: ericr86).

---

The bug does not exist on the ticket 6 branch of GMTK (GMTK 1.0.1 (Mercurial id: 52bb468453b2 (ticket6) checkin date: Wed May 07 17:08:57 2014 -0700)). I downloaded a script to reproduce the bug from here: https://j.ee.washington.edu/trac/gmtk/raw-attachment/ticket/6/weight_penalty_segfault.tar.gz

However on the default and ticket 161-2 branch the bug still exists. Their versions and their output are listed below:

$ gmtkTriangulate -version
GMTK 1.0.0rc (Mercurial id: c91e45e5729f (ticket161-2) checkin date: Mon Aug 26 22:37:35 2013 -0700)
$ gmtkTriangulate -version
GMTK 1.0.1 (Mercurial id: fed7fec16fe6 checkin date: Tue May 06 12:36:23 2014 -0700)

Produce identical output:

$ ./reproduce_bug.sh 
____ PROGRAM ENDED SUCCESSFULLY WITH STATUS 0 AT Tuesday June 10 2014, 10:15:47 EDT ____
windowBytes = 4 MiB = 4194304 B
fileBufferSize = 4194304
FileSource window = 5165 frames, 4 MiB
cookedBuffer = 20661 frames, 16 MiB
Reading keyword 'MEAN_IN_FILE' from inline.
Reading keyword 'COVAR_IN_FILE' from inline.
Reading keyword 'MC_IN_FILE' from inline.
Reading keyword 'DPMF_IN_FILE' from inline.
Reading keyword 'MX_IN_FILE' from inline.
Reading keyword 'DT_IN_FILE' from inline.
Reading keyword 'DETERMINISTIC_CPT_IN_FILE' from inline.
Reading keyword 'DENSE_CPT_IN_FILE' from inline.
Reading keyword 'DT_IN_FILE' from inline.
Reading keyword 'DETERMINISTIC_CPT_IN_FILE' from inline.
Finished reading in all parameters and structures
Creating Junction Tree
./train.sh: line 15: 18277 Segmentation fault      (core dumped) gmtkEMtrain -strFile dual_accumulator.str -triFile dual_accumulator.str.trifile -inputMasterFile dual_accumulator.master -outputMasterFile out.master -objsNotToTrain params.notrain -of1 data/strict-orbitrap-processed-spectrum.pfile -fmt1 pfile -ni1 0 -nf1 200 -of2 data/strict-orbitrap-processed-peptide.pfile -fmt2 pfile -ni2 3 -nf2 0 -fdiffact1 rl -dirichletPriors T -verb 58 -cbeam 0.5e10 -maxEmIters 5 $@

[EricR86]

Original comment by Michael Hoffman (Bitbucket: hoffman, GitHub: michaelmhoffman).

---

• changed state from "open" to "resolved"

Thanks, posted comment updating https://j.ee.washington.edu/trac/gmtk/ticket/6#comment:28 and we should assume this will be fixed soon so I'm going to mark it closed.