Dreaded (because unresolved) bad_alloc error

[Aurelien Mazurie]

Greetings,
I run into a 'std::bad_alloc()' error when processing a 500MB BAM file with cufflinks on a 8BG machine. The exact error message is

. . .
> Processed 17748 loci.                        [*************************] 100%
> Map Properties:
>	Total Map Mass: 14607725.69
>	Read Type: 54bp single-end
>	Fragment Length Distribution: Gaussian (default)
>	              Estimated Mean: 205.82
>	           Estimated Std Dev: 73.30
[16:12:46] Assembling transcripts and estimating abundances.
> Processing Locus EG_1:2534992-2535117        [************************ ]  99%terminate called after throwing an instance of 'std::bad_alloc'
  what():  std::bad_alloc

I saw many posts about this 'bad_alloc' error, which seams to be related to memory problems. For example,

http://seqanswers.com/forums/showpos...9&postcount=11

http://seqanswers.com/forums/showpost.php?p=19891&postcount=1

http://seqanswers.com/forums/showpost.php?p=29830&postcount=1

http://seqanswers.com/forums/showpost.php?p=35410&postcount=1

http://seqanswers.com/forums/showpost.php?p=33033&postcount=1

http://seqanswers.com/forums/showpost.php?p=33333&postcount=1

However none of those posts received an answer. Does anybody have an idea why this error occurs and how to fix it? I tried to run the same process on a Mac with 4 cores and 8GB of RAM, and a Linux 64bits with 4 cores and 12GB of RAM, with the same result.

Best,
Aurelien

[tonybolger]

I saw many posts about this 'bad_alloc' error, which seams to be related to memory problems.

AFAIK, this always indicates that an attempt at allocating memory failed.

This can be a result of running out of memory (due to actual need or a memory leak), an attempt to allocate a stupidly large amount of memory via a broken calculation, or heap fragmentation (so there's enough in total, just not enough in one chunk). You also could be hitting ulimits - check if there are any.

Try monitoring the consumed resources as it runs (e.g. using top), or if that isn't enlightening, maybe see if you can enable detailed logging. Last resort would be a debugger or some other tool to track memory allocations, and see what you see.

[Aurelien Mazurie]

Aurelien,
Did you find solution for the problem above? I having the same problem.

-Eteleeb

I moved on other projects since then, and I can't remember how I fixed (or dealt with) this error anymore. I know I got the pipeline running on the same machines, so the solution wasn't to add memory.
Aurelien