Notes for the ITC-Research Computing Standing Committee Meeting
February 9, 2009 at 10:45 AM
2015 Ivy Road, 1st Floor Conference Room

Members: Alice, Andrew, Hamp, Jim, Joe S., Mark, Mike, Robin, Terry, Tim S., Tom and members of Systems, Research Computing and ACHS as desired

Attending: Alice, Hamp, Jim, Joe S., Mark, Mike, Robin, Terry, Tim S., Tom, Ed Hall, Katherine Holcomb

Chair: Tom
Recorder: Alice

I. Corrections to minutes from last meeting on December 8, 2008? (Click to review the agenda from this meeting)

II. Old Business:

• Update on UVACSE APB and related.
• Andrew has contacted Hamp and Jim about moving to a queuing process that would allow pre-emption (although users will have to learn to use better descriptors for the queues).
• Storage needs assessment -- LSP dialog, survey (Joe)
• Storage needs and strategies were discussed at the most recent Computational Science group meeting – the CS department has a proposed storage strategy that might be useful for ITC and UVa.

• Cluster node purchase
• We have three researchers with funding for 29 new nodes (21 with InfiniBand) – if more buyers show up, we can order more nodes from Dell. We should have the orders out in a few days – but have had quite a hassle over the ETF rules.
• We are reconfiguring the 12 Elder nodes gradually this week – and can do pre-emption. We will install the new nodes when they come, using the Elder head node.

III. New Business:

• PBS Renewal / Migration (Terry)
• The ITC budget for next year does not include $6,000 for renewal of the PBS license (and this license also covers the CS and Astronomy departments). So what to do?
• We could switch to Maui and Torque – or to SGE – neither is vastly different from PBS, although Maui/Torque is closer to PBS.
• After some discussion about the pros/cons of the alternatives, it sounded like Maui/Torque is preferred.
• How much effort would it take on the backend to make the switch? Maybe 1-2 days.
• Let’s try it on a test cluster – play with it and see what it takes and what the impacts could be. And we can use the floating license manager to get PBS counts on the clusters. We will look at all this before the next meeting. We could make the switch over the summer.
• Katherine and Ed will talk to Andrew about this – and Katherine will contact James River about the costs for PBS for the coming year
• Remote Matlab use on 64-bit Windows computers (Ed)
• Jim says that there will be a solution – but there is no timeline for it yet.
• Meanwhile we can put something up on the web that addresses this issue
• Any other Items
• Katherine brought up compiler upgrades: she recommends that we wait for Intel 11.1 – but we could go ahead and upgrade PGI to 8.0. Since compiler upgrades tend to be involved, a staggered schedule tends to ease the pain. Reasons to upgrade PGI include a good debugger and performance improvements. Katherine will test more on her machine. The upgrade can be scheduled whenever it is convenient for Hamp and BillP. Hamp will also check the logs to see how much the PCI licenses are being used.
• Joe reported interest in having ITC do the distribution for ChemDraw. This would be OK if we can put a pointer to a vendor download link – and if we can make it clear that this software is not ITC funded or supported.

IV. Adjourn by 11:45 and next meeting

• Next Scheduled Meeting of ITC ResComp Standing Committee: March 9, 2009 in ITC-2015 Ivy Road, Room #102 (First Floor Conference Room)
  Please send suggestions, additions, corrections to: Tom or Alice