specifying nodes/processors with PBS
Hi, I've successfully configured a local galaxy instance to submit jobs to our local cluster, but have a question regarding setting PBS options such as the number of nodes and processors to use. I've followed the guide here: http://wiki.g2.bx.psu.edu/Admin/Config/Performance/Cluster?highlight=%28pbs%... Is there any way to do this in a tool-specific manner. For example, for running tophat to do a mapping run, I would want it to use 24 CPUs, but for simply converting a file type I would want only 1 CPU requested. Is there a way to specify this somewhere? Thanks, Jason -- Jason Greenbaum, Ph.D. Manager, Bioinformatics Core | jgbaum@liai.org La Jolla Institute for Allergy and Immunology
Yes, you can set these things for each tool independently. Check the file "universe_wsgi.ini" and check the section: [galaxy:tool_runners] There you can set the -pe options for GSUB in the same way you can do on the command line... Here's an example of my setup: gatk_variant_select = drmaa://-V -pe smp 1 gatk_validate_variants = drmaa://-V -pe smp 1 far = drmaa://-V -pe smp 1 velveth = drmaa://-V -pe mpi 4 velvetg = drmaa://-V -pe mpi 8 I set the -pe mpi option to 8 for VELVETH/G and to 1 for GATK for instance Thon On May 23, 2012, at 12:47 PM, "J. Greenbaum" <jgbaum@liai.org> wrote: Hi, I've successfully configured a local galaxy instance to submit jobs to our local cluster, but have a question regarding setting PBS options such as the number of nodes and processors to use. I've followed the guide here: http://wiki.g2.bx.psu.edu/Admin/Config/Performance/Cluster?highlight=%28pbs%... Is there any way to do this in a tool-specific manner. For example, for running tophat to do a mapping run, I would want it to use 24 CPUs, but for simply converting a file type I would want only 1 CPU requested. Is there a way to specify this somewhere? Thanks, Jason -- Jason Greenbaum, Ph.D. Manager, Bioinformatics Core | jgbaum@liai.org La Jolla Institute for Allergy and Immunology ___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
Excellent! Thanks for the reply. -J -- Jason Greenbaum, Ph.D. Manager, Bioinformatics Core | jgbaum@liai.org La Jolla Institute for Allergy and Immunology ----- Original Message -----
From: "Anthonius deBoer" <thondeboer@me.com> To: "J. Greenbaum" <jgbaum@liai.org> Cc: galaxy-dev@lists.bx.psu.edu Sent: Wednesday, May 23, 2012 2:03:18 PM Subject: Re: [galaxy-dev] specifying nodes/processors with PBS
Yes, you can set these things for each tool independently. Check the file
"universe_wsgi.ini"
and check the section:
[galaxy:tool_runners]
There you can set the -pe options for GSUB in the same way you can do on the command line...
Here's an example of my setup:
gatk_variant_select = drmaa://-V -pe smp 1 gatk_validate_variants = drmaa://-V -pe smp 1 far = drmaa://-V -pe smp 1 velveth = drmaa://-V -pe mpi 4 velvetg = drmaa://-V -pe mpi 8
I set the -pe mpi option to 8 for VELVETH/G and to 1 for GATK for instance
Thon On May 23, 2012, at 12:47 PM, "J. Greenbaum" <jgbaum@liai.org> wrote:
Hi,
I've successfully configured a local galaxy instance to submit jobs to our local cluster, but have a question regarding setting PBS options such as the number of nodes and processors to use. I've followed the guide here:
http://wiki.g2.bx.psu.edu/Admin/Config/Performance/Cluster?highlight=%28pbs%...
Is there any way to do this in a tool-specific manner. For example, for running tophat to do a mapping run, I would want it to use 24 CPUs, but for simply converting a file type I would want only 1 CPU requested. Is there a way to specify this somewhere?
Thanks,
Jason
-- Jason Greenbaum, Ph.D.
Manager, Bioinformatics Core | jgbaum@liai.org
La Jolla Institute for Allergy and Immunology
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client. To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
participants (2)
-
Anthonius deBoer -
J. Greenbaum