[mira_talk] Re: mapping assembly -> memory problem
- From: Bastien Chevreux <bach@xxxxxxxxxxxx>
- To: mira_talk@xxxxxxxxxxxxx
- Date: Tue, 22 Dec 2009 16:05:06 +0100
On Dienstag 22 Dezember 2009 Charles Imbusch wrote:
> I'm trying to make use of mira's mapping assembly feature. As
> a reference I use genomic data (a fasta file around 400MB) and
> an EST dataset from the same species to be assembled.
> In a very early stage, mira quits with the message:
> Out of memory detected, exception message is: std::bad_alloc
> I would be very interested in what went wrong. I attached the log file
> and the parameters to this message. I worked with mira-3rc3e
> Any hint is highly appreciated.
Hi Charles,
you're running a 32 bit executable ("Compiled on varcadia32 ... i686"), this
limits the available space to 4 GB per process (even if your machine has 16
GB), and that is not enough for MIRA to build internal tables for these 400MB
backbone. Take the 64 bit executable, then you'll be able to use more RAM.
BTW, thanks for the log, this showed me how to really find out during runtime
whether a program is running in 32 bit regardless of the host.
I never tried mapping ESTs to more than small fungi, it may very well be that
MIRA building "contigs suitable for finishing" will eat away more memory than
you have.
On another note: mapping ESTs to a genome will not work well as MIRA doesn't
do splicing, I need to document that in the docs. You might want to use other
mappers for this task.
Regards,
Bastien
--
You have received this mail because you are subscribed to the mira_talk mailing
list. For information on how to subscribe or unsubscribe, please visit
http://www.chevreux.org/mira_mailinglists.html
Other related posts:
