Notes for the ITC-Research Computing Standing Committee Meeting

July 13, 2009 at 10:45 AM
Carruthers, Conference Room A

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, Andrew, Hamp, Jim, Mark, Robin, Tim S., Tom, Kathy Gerber, Katherine Holcomb, Ed Hall, Tony Townsend, Dale Castle

Chair: Tom,

Recorder: Alice

I. Corrections to minutes from last meeting on April 13, 2009? (Click to review the agenda from this meeting)

II. Old Business:

• Announcement about “new” HPC policies (Katherine)
• Katherine sent out a draft announcement and did not receive any responses – she will send it out again.
• MRI/ARI proposals
• Neither one is a ‘go’; the MRI proposal did not get past the VPR review process; and the ARI’s matching requirements were not a good fit for our situation.
• SGE testing
• No news
• PBS purchase
• Looks like we need to renew for another year – Katherine will get a quote.
• Nodes for Leonard Scott?
• His proposal has been stymied by OSP’s objections to shared nodes – trying to work around this with a rate for ‘parking nodes’ in the cluster.
• Additional memory for Dogwood
• Have cost estimates for adding memory to Dogwood – but after some discussion it was not clear that there is a demand for additional memory on Dogwood – so decided to wait on this until there are requests for more memory.
• Additional disks for Elder
• Dependent on outcome of glustre testing – see Lustre testing item below.
• Additional Elder nodes
• Andrew added to itc-rescomp?
• Yes – done.
• Default-shell change to bash
• Dale and Mark tested a set of bash startup scripts – looks fine. It is easy to change the default login shell for new accounts – and old users can change to it if they want to (as long as they have the right path). Bash and ksh have a few incompatibilities. We will make this change for new accounts – Katherine will send out information.
• Dropping OpenMPI?
• Drop it? Keep it but don’t maintain it? Having multiple MPIs and lots of compilers means building an MPI for each compiler – and inhibits moving codes around. Katherine will send out some information.
• Lustre testing
• Bill Pemberton has experimented with glustre (which does not require kernel mods) – will write this up – looks like it could satisfy our requirements. We should have more complete results by next week.
• Web page maintenance
• Katherine has put the cluster pages on uvacse and is working on making it self-contained – also rewriting the Unix Basics documents – these are still in the test area – and will be moved to production. Will also archive the ‘out of date’ material. Katherine will send out a link for review. Ed Hall wants to move some cluster-related Matlab stuff too.
• New storage service/pricing
• The new rate proposal is still in Cost Accounting -- waiting for approval. There was discussion about pricing and storage levels – as well as about the existing archiving system (HSM). Need up-to-date documentation and more outreach about these storage options. Katherine is willing to rewrite the HSM documentation – and then UVACSE/Andrew will highlight the HSM and the new storage service and pricing.
• Research software budget
• Next time
• Service Unit allocation
• Andrew working with Mike and Murielle on this – needed more than ever now that we are adding more resources to the grid.

III. New Business:

• 32-bit Virtual Machine
• This is under discussion – so that we could have a place to build 32-bit versions of things.
• Gaussian 98 (G09?)
• Gaussian 98 will not compile under 64-bits – and the 32-bit version had lots of warnings and will not run. Have a user with plotting issues with G03. Maybe this would be a good time to say ‘no’ to Gaussian 98. Andrew will take this issue to the APB and get approval – and then that recommendation can go to the ITC Transitions group (who are primarily interested in having orderly transitions with prior notifications). FYI – G09 is on the horizon.

IV. Adjourn by 11:45 and next meeting

• Next Scheduled Meeting of ITC ResComp Standing Committee: August 10, 2009 in ITC-2015 Ivy Road, Room #102 (First Floor Conference Room)

Please send suggestions, additions, corrections to: Tom or Alice