Hi Dear Aaron and Mia, thank you very much for your quick replies about my "bad_alloc" problem with mauveAligner a couple of weeks ago. I'm sorry for the long delay in coming back to you, but I've been waiting for computer resources and I've been busy with other things in the meantime.
Yes, using the 64-bit version of mauveAligner solves the problem, and runs fine when more than 4GB of RAM are needed. So, I could get a multiple alignment of 67 bacterial genomes in 24hrs with a 48GB CPU, without splitting the dataset into smaller subsets. Thanks very much again, Sincerely, Nicolas. >From: Mia D Champion <[email protected]> >To: Nicolas Tourasse <[email protected]> >Sent: Sun, January 31, 2010 6:59:43 PM >Subject: Re: [Mauve-users] Memory allocation error in MauveAligner > >Dear Nicolas: > >I seem to recall running into a similar problem- > >I think your error may not be due to a direct issue with memory but, >instead a problematic alignment with one or more of the genomes. >I would suggest incrementing your genomes for alignment in smaller >batches to try to identify the culprit- >afterall, going from 20 successful alignments to 67 is quite a jump. > >Hope this helps, >Mia Aaron Darling wrote: > Hello Nicolas, > > On Sat, 2010-01-30 at 16:27 +0100, Nicolas Tourasse wrote: >> I got a "std::bad_alloc" exception when trying to align 67 bacterial >> genome sequences of 5 Mbp each. The error occurs at the 1st alignment >> step after reading the sequences and creating the sorted mer lists. The >> command line and the end of the output are shown below. >> It is likely to be a problem with memory allocation related to a C++ >> "new" statement. I would think it's a problem with a program statement, >> because I'm running mauveAligner on a supercomputer with 16GB RAM CPUs >> and the crash occurs when 4GB are used, so it's probably not a problem >> of available RAM on the computer. Maybe the allocation statement makes >> the program ask for more memory than it should use. In fact, the program >> works fine when I align only 20 of the genomes. >> >> Would you be able to help me with this problem? > > > Yes indeed, std::bad_alloc is a c++ error that occurs when memory > allocation fails with the new function. > > Although your supercomputer has plenty of memory, I have a strong > suspicion that you might be using a 32-bit version of mauveAligner, > which can not allocate more than 4GB of memory due to 32-bit address > space limitations. Mauve releases contain both 32-bit and 64-bit > builds. In the Linux release the 64-bit builds of mauveAligner and > progressiveMauve reside in the linux-x64 subdirectory within the main > release archive. It's possible to check whether a program is 32-bit or > 64-bit using the 'file' command, e.g.: > file mauveAligner > > In windows the 64-bit binaries are in a win64/ subdirectory. On Mac OS > X the 32-bit and 64-bit versions are bundled together into a single > "fat" binary, so Mac users shouldn't ever experience this issue. > > If that doesn't solve it then something has run seriously amok, and it > would be great if you could provide more information (off-list) about > your genome sequences. > > Cheers, > -Aaron > -- Tourasse Nicolas, PhD University of Oslo, Laboratory for Microbial Dynamics (LaMDa), School of Pharmacy, Department of Pharmaceutical Biosciences, P.O. Box 1068 Blindern, 0316 Oslo, Norway. Tel: (+47) 22 84 49 41 FAX: (+47) 22 84 49 44 e-mail: [email protected] website: http://lamda.uio.no/lamda.html ------------------------------------------------------------------------------ SOLARIS 10 is the OS for Data Centers - provides features such as DTrace, Predictive Self Healing and Award Winning ZFS. Get Solaris 10 NOW http://p.sf.net/sfu/solaris-dev2dev _______________________________________________ Mauve-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mauve-users
