Hi Wei-Jen,

1.

It seems that you are using mpiexec 0.83 from
https://www.osc.edu/~djohnson/mpiexec/#Download

This mpiexec flavor does not seem to have the option I
told you to use.

In the source code of mpiexec 0.83, there is this:

#if 0  /* No this doesn't work, but if I don't check it in I'll forget 
where I put it.  */
     fprintf(stderr,
       "  -output <prefix> : send process output to separate files\n");
#endif

So the feature is in the source code, but is not activated.


2.

To change the k-mer length, use -k <value> (for 31, use -k 31).

For a k-mer length longer than 31, you need to compile Ray with
MAXKMERLENGTH=<maximum> (for -k 55, you need to compile with
MAXKMERLENGTH=64.

Wei-Jen Chang a écrit :
> Hi Sebastien,
>
> It still did not work. Below is the message I received. I have asked our
> administrator to take a look. Maybe he has a better idea how to deal
> with it.
>
> Btw, I want to ask how to increase the kmer limit in Ray? I know the
> default number is 21.
>
> Thanks,
>
> WJ
>
> Usage: mpiexec [<args>] <executable> [<exe args>]...
>     or: mpiexec [<args>] -config[=]<file>
>     or: mpiexec [<args>] -server
>    -n <numproc> : use only some of the allocated processors
>       Default behavior allocates one process per allocated processor.
>    -verbose : be verbose about mpiexec operation
>    -nostdin : do not listen to stdin, allowing process to go into background
>    -allstdin : send stdin to all processes (default just proc #0)
>    -nostdout : do not redirect stdout/stderr, but let pbs accumulate it
>    -comm (gm|mx|p4|ib|rai|pmi|lam|shmem|emp|portals|none) : choose MPI
> (default m
> pich2-pmi)
>      -mpich-p4-[no-]shmem : for MPICH/P4, specify if the library was
>                             compiled with shared memory support (default
> yes)
>    -pernode : allocate only one process per compute node
>    -npernode <nprocs> : allocate no more than <nprocs> processes per
> compute node
>    -nolocal : do not run any MPI processes on the local node
>    -transform-hostname[=]<sed expression> : use alternate names for MPI
>    -transform-hostname-program[=]<executable> : use this script or program
>                                                 to generate alternate names
>    -tv : debug using totalview (ensure it is in your path)
>    -kill : kill other processes if any one process exits
>    -config[=]<file> : use heterogenous node specification file ("-" for
> stdin)
>    -server : do not run any tasks, just serve other concurrent mpiexec
> clients
>    -version : show version information
> Version 0.83, configure options: '--with-pbs' '--with-default-comm=pmi'
>
>
> On Wed, Aug 15, 2012 at 11:02 AM, Sébastien Boisvert
> <sebastien.boisver...@ulaval.ca <mailto:sebastien.boisver...@ulaval.ca>>
> wrote:
>
>     [please keep denovoassembler-users cc'ed]
>
>     Hello,
>
>     There are two dominant MPI implementations: Open-MPI and MPICH2.
>     Most others are derived from these two.
>
>     If you got a usage error, you are probably using MPICH2.
>
>     With MPICH2, you can generate one log file per MPI rank with this
>     command:
>
>     mpiexec -np 64 \
>       -outfile-pattern Ray_test_08.%r.stdout \
>       -errfile-pattern Ray_test_08.%r.stderr \
>
>       Ray -k 21 -p \
>       /home/bio443/mtherkel/__SRR131208_1.fastq \
>       /home/bio443/mtherkel/__SRR131208_2.fastq \
>       -o Ray_test_08
>
>
>     Sébastien
>
>     Wei-Jen Chang a écrit :
>
>         Hi Sebastien,
>
>         I might have misunderstood you. I ran a line like this:
>
>         mpiexec -np 64 -output-filename Ray_test_08 Ray -k 21 -p
>         /home/bio443/mtherkel/__SRR131208_1.fastq
>         /home/bio443/mtherkel/__SRR131208_2.fastq -o Ray_test_08
>
>         and received an error message (usage error). Did I put the
>         -output-filename wrongfully?
>
>         Thanks,
>
>
>         WJ
>
>         On Tue, Aug 14, 2012 at 11:24 AM, Sébastien Boisvert
>         <Sebastien.Boisvert.3@ulaval.__ca
>         <mailto:sebastien.boisver...@ulaval.ca>
>         <mailto:Sebastien.Boisvert.3@__ulaval.ca
>         <mailto:sebastien.boisver...@ulaval.ca>>>
>
>         wrote:
>
>              Hello Wei-Jen,
>
>
>              If you are using Open-MPI, you can use the command below to
>         generate
>              one standard output file per MPI rank (there is 1 MPI rank per
>              processor core).
>
>                    mpiexec -n 64 \
>                    -output-filename Test21 \
>
>                     Ray -k 21 \
>                    -o Test21 \
>                    -p SRR131208_1.fastq SRR131208_2.fastq
>
>              You will get 64 standard output files named
>              Test21.1.00, Test21.1.01, ...,  Test21.1.63.
>
>              With this, it will be easier to find the problem.
>
>              Also, you can compile Ray with
>
>              make ASSERT=y DEBUG=y PREFIX=DebugBuild
>              make install
>
>              To turn on runtime assertions.
>
>
>
>
>
> --
> Associate Professor of Biology
> Hamilton College
> Clinton, NY 13323
> (315) 859-4296


------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Denovoassembler-users mailing list
Denovoassembler-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/denovoassembler-users

Reply via email to