[Swift-devel] [Bug 150] multiple workers on one compute node

[bugzilla-daemon at mcs.anl.gov]

http://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=150





------- Comment #1 from wilde at mcs.anl.gov  2008-07-22 13:04 -------
The following note from TeraGrid Support suggests that we might be able to set
form GRAM, at least from WS-GRAM:

-------- Original Message --------
Subject: Re: Using > 1 CPU per compute node on Abe under GRAM]
Date: Tue, 22 Jul 2008 12:40:16 -0500
From: consult at ncsa.uiuc.edu
To: Michael Wilde <wilde at mcs.anl.gov>

FROM: Estabrook, John
(Concerning ticket No. 159049)

There is no policy nor setting to prevent this; the quoted info at

http://www.globus.org/toolkit/docs/4.0/execution/wsgram/WS_GRAM_Job_Desc_Extensions.html#r-wsgram-extensions-constructs-nodes

shows how to make the request, noting that it requires a current version of
GRAM.

------------------------
John Estabrook
NCSA Consulting Services
------------------------

Michael Wilde <help at teragrid.org> writes:
>Hi Help Team,
>
>Can you forward this to the Abe folks at NCSA?
>
>We have a user thats trying to get his single-node GRAM jobs to run up 
>to 8-per-compute-node (ie one-jobs-per-core) on Abe.
>
>Is this possible, or prevented by some combination of Abe PBS and GRAM 
>jobmanager setting?
>
>If its possible, please let us know what setting to use for WS-GRAM and 
>pre-WS-GRAM, if you could.
>
>If its not, not to worry - we have other strategies to turn these into 
>multi-node jobs.
>
>There's more info in the thread below.
>
>Thanks,
>
>Mike


-- 
Configure bugmail: http://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.