Terminating with 'std::bad_alloc'

[Jonathan Lawton]

Hello,

I am attempting to run corset on 78 samples of two conditions. I attempted to run the standard pipeline, but given the large number of reads that I have, it seemed like it would take weeks. I am now attempting to run the version with parallelized .bam reading. However, when I run the parallel loop suggested in the wiki page, many of the samples terminate early with the following error:

###

terminate called after throwing an instance of 'std::bad_alloc'
  what():  std::bad_alloc

###

Has anyone else encountered this problem? Any advice?

Thanks!
Jonathan

[Jonathan Lawton]

Hi, I figured this one out. It is not a corset-specific error. Rather, it's a java error resulting from running out of memory to allocate to a process. I am operating on a shared server environment and was fighting for RAM with a colleague. Leaving this up in case it will maybe help someone else.

On another note, the parallelized version of the file-reading process probably saved me weeks of time. I know the author mentioned in previous posts that they do some post-processing after reading each file in the conventional pipeline, but I think this takes way too long for large datasets.

[Nadia Davidson]

Thanks for sharing your experience for others, and apologies for not seeing this earlier.