[phylocom-user] problem solved, now a new one

Steven Kembel steve.kembel at gmail.com
Fri Apr 10 09:14:48 PDT 2009


Hi Chris,

That error is a new one to me. Looks like a memory issue. Comstruct  
requires much more RAM than comdist since comstruct stores results  
from randomizations in memory.

Could you try trimming your sample file down to just a couple of  
samples, run phylocom comstruct on that smaller sample file, and see  
what happens? Alternatively, try decreasing the number of runs with  
the -r argument and see if that changes anything. Decreasing the  
number of runs will decrease the memory required by comstruct in a  
linear fashion. If either of these approaches works, you will have to  
split up your sample file and run it in pieces, or find a computer  
with more memory, since you may be hitting limits based on the amount  
of RAM you have.

Cheers,
Steve

On Apr 10, 2009, at 8:46 AM, Chris Jones wrote:

> Hi again,
>
> Sorry to keep hitting the listserve...I figured out the previous  
> issue (bootstrap values were the problem).  But, now I know I'm stuck.
>
>
> comdist and comdistnn work fine, but now I get this error when  
> calling comstruct:
>
> File `/Users/chrisjones/Documents/NirK_NirS_MS/Label_files/ 
> nirK_phylocom_lab' has unix newlines
> phylocom(11835) malloc: *** mmap(size=2339078144) failed (error  
> code=12)
> *** error: can't allocate region
> *** set a breakpoint in malloc_error_break to debug
> Numerical Recipes run-time error...
> allocation failure 2 in matrix()
> ...now exiting to system...
>
> Anyone have a suggestion for a work around?
>
> Again, many thanks,
> Chris
>
> Chris Jones
> Ph. D. Student, Department of Microbiology
> Swedish University of Agricultural Sciences
> chris.jones at mikrob.slu.se
>
>
>
> _______________________________________________
> phylocom-user mailing list
> phylocom-user at lists.phylodiversity.net
> http://lists.phylodiversity.net/listinfo.cgi/phylocom-user-phylodiversity.net




More information about the phylocom-user mailing list