Re: [modeller_usage] Is it possible to rum model_single.py in multiple nodes ?
To: Mahesh Velusamy <>,
Subject: Re: [modeller_usage] Is it possible to rum model_single.py in multiple nodes ?
From: Modeller Caretaker <>
Date: Mon, 27 Jun 2016 11:56:38 -0700
On 6/25/16 1:34 PM, Mahesh Velusamy wrote:
I know there is a option to run modeller in multiple cores but How can run
model_single.py in multiple nodes (My wish is to rum 44 nodes where each
node has 16 PROCESSOR).
It's exactly the same. The building of each model is completely
independent of all the others, so they are trivial to parallelize. Most
simply you can just run the same script on each node, just with a
different starting random number seed:
If you want to use automodel's parallel job support, rather than using
local_slave use one of the other slaves that understands multiple nodes
(e.g. ssh_slave) or compute clusters (e.g. sge_pe_slave, sge_qsub_slave).