Minutes of the ITC-Research Computing Standing Committee

Meeting on April 28, 2003

 

Members:  Alice, Bill, Brian, Dawn, Ed, Hamp, Kathy G., Katherine H., Jim, Mark S., Martha, Michael, Robin, Sue Ellen, Steve, Terry, Tim S., Tom S.

 

Attending: Alice, Bill, Ed, Hamp, Kathy G., Katherine H., Jim, Mark S., Martha, Robin, Steve, Tim S., Tim T.

 

Chair:  Tim T.

Recorder:  Alice

 

I.   Corrections to minutes from the last meeting on March 31, 2003

 

1.  Item II.4 (concerning network security of workstations in the Unixlabs) was corrected to remove reference to using the SMBClient which the lab workstations do not support.  Those minutes are available at: http://holmes.acc.virginia.edu/ResNotes/passprotected/itc-rescomp.min04-28-03.html

 

II.   Ongoing Discussion Topics:

 

1.  Aspen Xeon Cluster (Birch) rollout:

·    Bill has amended the PBSPro script to echo cluster name and temp space location.

·    Need for additional disk space on both clusters:  there was discussion about the various options for providing users with more disk space.  Should we add file servers and allow users to have their “own longtmp” as a staging space?  Or should we add a large drive to each front-end and make a bigger “bigtmp”?  The latter would be less expensive.  There are also policy aspects to this – do we want to essentially give some users more home directory space?  No, whatever we do, it needs to be treated as a temporary space (like longtmp) and called longtmp.  We will explore the options and compare the costs.

·    Need to upgrade Totalview to support the Intel compiler – but this would require fixing a kernel problem.  There was discussion about  the pros/cons of having a custom kernel vs. staying with the RedHat kernel – perhaps RedHat will come out with a kernel that fixes this problem.  We decided to wait awhile – but go ahead with the cluster rollout since the cluster can be used without Totalview – even though we do want to encourage users to debug using Totalview.

·    Intel vs. PGI compilers:  After the Xeon cluster goes public, Katherine and Ed can begin testing and comparison to see if we can settle on just one.

·    Testing and support of test users – any other outstanding issues or concerns?  Close to rollout.  Katherine trying to get VASP code working on Birch cluster.

 

2.  Aspen AMD Cluster (aspen):  Hamp will set a date (after the end of the spring semester) for a day of downtime to upgrade the AMD cluster.

 

3.  SP Migration:  Have been in contact with some users – working with Lester Andrews and Gaussian users.  Will send an e-mail to all current SP users when we have announced the availability of the Xeon cluster.

 

III.   New Topics:

 

1.  The Aspen clusters need names that users can remember and use to login – and so that they and we can communicate with each other clearly and easily.  It was proposed and agreed that we will use a tree theme – aspen for lc0, birch for lc1, and cedar for the next one.

 

2.  It was decided to have a May meeting on Monday, May 19, at 10:30 AM in 117 Astronomy.