Menu
▾
▴
#360 Barista memory leaks
Examine barista-tutorial.html for further explaination on how to set up barista.
Valgrind output:
==12860== Memcheck, a memory error detector
==12860== Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al.
==12860== Using Valgrind-3.12.0.SVN and LibVEX; rerun with -h for copyright info
==12860== Command: crux barista --separate-searches crux-output/tide-search.decoy.txt concat.fasta demo.ms2 crux-output/tide-search.target.txt --overwrite T
==12860==
INFO: Beginning barista.
WARNING: The output directory 'crux-output' already exists.
Existing files will be overwritten.
WARNING: The file 'crux-output/barista.log.txt' already exists and will be overwritten.
INFO: CPU: D-173-250-156-211.dhcp4.washington.edu
INFO: Wed Dec 2 17:32:14 PST 2015
INFO: database source: concat.fasta
INFO: sqt source: crux-output/tide-search.target.txt
INFO: ms2 source: demo.ms2
INFO: output_directory: crux-output
INFO: enzyme: trypsin
INFO: decoy prefix: decoy_
INFO: digesting database concat.fasta
INFO: parsing files:
INFO: reading file demo.ms2
INFO: parsing file crux-output/tide-search.target.txt
INFO: PSM number 0
INFO: parsing file crux-output/tide-search.decoy.txt
INFO: Number of spectra: 332
INFO: Number of PSMs: total 332 positives 166 negatives 166
INFO: Number of peptides: total 188 positives 88 negatives 100
INFO: Number of proteins: total 89 positives 41 negatives 48
INFO: loading and normalizing data
INFO: trainset size 89
INFO: q<0.01: max non-parsimonious so far 9
INFO: q<0.01: max non-parsimonious so far 13
INFO: q<0.01: max non-parsimonious so far 15
INFO: q<0.01: max non-parsimonious so far 17
INFO: q<0.01: max non-parsimonious so far 19
INFO: finished training, making parsimonious protein set
number of meta groups 89
number of subset groups 0
INFO: total proteins parsimonious at q<0.01: 19
INFO: peptides at q<0.01: 21
INFO: psms at q<0.01: 65
INFO: Elapsed time: 1.2e+03 s
INFO: Finished crux barista.
INFO: Return Code:0
==12860==
==12860== HEAP SUMMARY:
==12860== in use at exit: 26,824,232 bytes in 56,333 blocks
==12860== total heap usage: 7,674,400 allocs, 7,618,067 frees, 1,450,847,717 bytes allocated
==12860==
==12860== LEAK SUMMARY:
==12860== definitely lost: 31,176 bytes in 340 blocks
==12860== indirectly lost: 26,745,590 bytes in 55,676 blocks
==12860== possibly lost: 0 bytes in 0 blocks
==12860== still reachable: 25,327 bytes in 128 blocks
==12860== suppressed: 22,139 bytes in 189 blocks
==12860== Rerun with --leak-check=full to see details of leaked memory
==12860==
==12860== For counts of detected and suppressed errors, rerun with: -v
==12860== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
Should also mention that the majority of the time taken to run valgrind (30 mins+) happened after the line that reads "INFO: psms at q<0.01: 65". So the issue is probably somewhere in clean up.
Last edit: Jared Moore 2016-01-10
I'm investigating the 26,761,300 lost in Barista::crux_set_command_line_options
In debug mode (with line numbers)
This is from running valgrind on a main with only the creation of an SQTParser. It's at least causing some errors.
Committed revision 16982.