Duke-UNC Brain Imaging and Analysis Center
BIAC Forums | Profile | Register | Active Topics | Members | Search | FAQ
 All Forums
 Support Forums
 Cluster Support
 List of cluster users

Note: You must be registered in order to post a reply.
To register, click here. Registration is FREE!

Screensize:
UserName:
Password:
Format Mode:
Format: BoldItalicizedUnderlineStrikethrough Align LeftCenteredAlign Right Horizontal Rule Insert HyperlinkInsert EmailInsert Image Insert CodeInsert QuoteInsert List
   
Message:

* HTML is OFF
* Forum Code is ON
Smilies
Smile [:)] Big Smile [:D] Cool [8D] Blush [:I]
Tongue [:P] Evil [):] Wink [;)] Clown [:o)]
Black Eye [B)] Eight Ball [8] Frown [:(] Shy [8)]
Shocked [:0] Angry [:(!] Dead [xx(] Sleepy [|)]
Kisses [:X] Approve [^] Disapprove [V] Question [?]

 
Check here to subscribe to this topic.
   

T O P I C    R E V I E W
clithero Posted - Sep 07 2009 : 5:51:06 PM
Hi everyone,
I was wondering if would be possible to somewhere (protected, I understand) have a list of all cluster users (their cluster IDs) and the corresponding emails.
Cluster usage is going up (which is great!), but I believe that usage would be more efficient (in terms of number and length of jobs submitted) if everyone was able to communicate with other users.
Thanks,
John
8   L A T E S T    R E P L I E S    (Newest First)
dvsmith Posted - Nov 06 2009 : 6:11:18 PM
Thanks, Chris. I'm sure this will be very helpful.
petty Posted - Nov 05 2009 : 4:42:20 PM
You'll notice now that when you login to your dashboard on the biacweb site there is a new "Tools" tab.

Currently the only tool is a directory search of all biac users. You can use this to search by username to get in contact with any users out of the view of the public.

http://fourier.biac.duke.edu/biacweb/tools/search.php


josh.bizzell Posted - Sep 23 2009 : 11:37:21 AM
I've made the Huettel Lab wiki page linked above readable by anyone who has a BIAC wiki account.

There is also a link to the page from the following wiki page:
http://fourier.biac.duke.edu/wiki/doku.php/biac:cluster

- Josh
dvsmith Posted - Sep 22 2009 : 6:17:33 PM
Yeah, this page should be made publicly accessible until a better submission/management algorithm is developed at BIAC. Perhaps the information in it can be mirrored to the public part of the wiki...

lh115 Posted - Sep 22 2009 : 10:00:52 AM
Sorry for any inconvenience I may have caused.

These scripts look really helpful, I hadn't heard of them before. I don't currently have access to the Huettel wiki. Maybe we could post documentation on their use so myself and other users can better manage submissions.

Lars
dvsmith Posted - Sep 21 2009 : 10:14:35 PM
so when people have stuck jobs and their email is ambiguous, we should post to the forums? i doubt everyone uses their biac account as their primary email (and hence they probably do not check it regularly), so it might be good to at least have a name that goes with the user ID (which can be fairly cryptic sometimes) so we can have a better chance of figuring out someone's email.

currently, lh115 (whoever that is) has had what are presumably 1st levels running for about 2 hours. there are about 10 of them up there. 1st levels do not take that long, so obviously the jobs are stuck or either this person is misusing the cluster.
1149804 1 L1Pre-Mask-B-18Sep-all.sh lh115 09/21/2009 19:58:11 users.q

hopefully we can work on improving current short-comings like this in the future. justin meyer and mckell carter developed an elegant job submission script that can easily prevent a lot of problems (like stuck jobs).
http://www.biac.duke.edu/forums/topic.asp?TOPIC_ID=1184
http://fourier.biac.duke.edu/wiki/doku.php/huettel:using_the_biac_cluster
until better algorithms are developed, scripts like these should be compulsory.


yaxley Posted - Sep 09 2009 : 10:19:04 AM
Correct me if I'm wrong, but the primary reason for "communicating with cluster users" is to ensure that folks are being polite with their job submission. It is absurd that people are doing this menial task. The gridlock that often occurs with job submissions is an indictment of the current cluster implementation.

As we move forward with the new cluster, I suggest that a job submission algorithm be implemented that simply passes jobs to the cluster by rotating consecutively through each user that has submitted jobs. This is a simple rule that would ensure that no matter how many jobs any user submits, they will only have to wait until their turn in the queue of cluster users to see a job run. Is such an implementation reasonable/doable?
petty Posted - Sep 09 2009 : 09:07:50 AM
I think the best place for communication is actually on the forums ... there really isn't a secure place that everyone can access, that can be routinely kept up to date. And we wouldn't want to publish email addresses on the forums since they are all visible by everyone.

BIAC Forums © 2000-2010 Brain Imaging and Analysis Center Go To Top Of Page
This page was generated in 0.36 seconds. Snitz Forums 2000