Meme parallel fails on multiple nodes

[Some Xyz]

Dear all,

When I try to run Meme parallel on 8 nodes with 16 cores on each. It running for 6 iteration after after that It giving  following message :

meme:14413 terminated with signal 11 at PC=428d49 SP=7fff26d2f2d0.  Backtrace:

/lib64/libc.so.6(__libc_start_main+0xfd)[0x33e341ecdd]

meme[0x402ee9]

meme[0x428d49]

meme[0x4290a4]

meme[0x424bf2]

In my cluster we are using sge gridengine, following script is used for job submission :

#!/bin/bash

#

# Script jobs            # Comment

#

#$ -q all.q          # Use specific queue

#$ -N meme_CSC            # Job Name

#$ -cwd              # Change to current working Directory

#$ -S /bin/bash      # Job run shell

#$ -V                # Export environment variables into script

#$ -j y              # Merge standard out and error into one file

#$ -pe mpich 128      # MPI enviroment use for 2 cores

meme K27L_K4I_expanded.fa -oc K27L_expanded1 -p $NSLOTS -mod zoops -dna -nmotifs 10 -minw 5 -maxw 30 -maxsize 20000000;

But when I try to run on single node with 16 processor it doesn't giving error.

Please tell us why meme is failing ?

[CharlesEGrant]

My first guess would be that the head node is getting killed for using too much memory. What is the default amount of memory allocated to a node? You may need to tell SGE that your job requires more memory.

You might try something like this:

#$ -l h_vmem=4G