Hello,
I am attempting to run MODELLER in parallel on a cluster with multiple nodes. The script I used is below:
from modeller import * from modeller.parallel import * from modeller.automodel import *
log.verbose() j = sge_pe_job() env = environ() a = automodel(env, alnfile='2NT2-mult.ali', knowns=('1m3gA','2wgpA','2hxpA'), sequence='2NT2', assess_methods=(assess.DOPE, assess.GA341)) a.starting_model = 1 a.ending_model = 25 a.use_parallel_job(j) a.make()
However, once I submit the job (using the parallel environment mpich), the job runs on the master node fine but fails on all the slave nodes. The error is also ambiguous:
within the slave1 output file (identical for all other output files except the slave0 one which runs fine): error: executing task of job 76003 failed:
Thanks in advance for the help!
-Charles Xue
On 7/27/10 10:39 PM, Charles Xue wrote: ... > However, once I submit the job (using the parallel environment mpich), > the job runs on the master node fine but fails on all the slave nodes. > The error is also ambiguous: > > within the slave1 output file (identical for all other output files > except the slave0 one which runs fine): > error: executing task of job 76003 failed:
That's not a Modeller error message. Looks like you have some problem with your SGE setup, so that a 'qrsh -inherit' call does not work properly. Do you have SGE correctly set up for tight integration?
Ben Webb, Modeller Caretaker
-
Charles Xue -
Modeller Caretaker