1 / 8

MRC Recap/Progress/Path forward

MRC Recap/Progress/Path forward. March – MRC Workshop Centralized services are of interest Not clear whether there is lab-wide requirement for MRC June – Division Director’s Retreat MRC presentation by Sandy Merola and Horst Simon Favorable response to idea to provide MRC services

deon
Download Presentation

MRC Recap/Progress/Path forward

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. MRC Recap/Progress/Path forward • March – MRC Workshop • Centralized services are of interest • Not clear whether there is lab-wide requirement for MRC • June – Division Director’s Retreat • MRC presentation by Sandy Merola and Horst Simon • Favorable response to idea to provide MRC services • July – Sandy presented proposal to Ops and Scientific Division Heads • Overhead/recharge funded MRC 2-year Start-up Program • Scientific Division Heads interested • Currently – ITSD starting to refine proposal and line up potential applicants for program • September – Next chance to secure funding for overhead supported program

  2. MRC Two Year Start-up Program • Intended to reduce costs to programs and improve use of scientific computing • Services provided to selected projects • Exact number of projects accommodated depends on demand, funding, show impact on lab science • Target minimal level is support 60 nodes – approximately 6 systems • Projects staged in over first year of program • Funded by overhead with possible modest recharge • Possible end states in FY’05 • Demand for computing will have led to an Institutional MRC • Remove subsidy; MRC support services are self-sufficient • Remove subsidy; MRC support is not feasible or needed

  3. Program Services • Services provided as part of program (overhead with possible modest recharge) • Computer room space • Test cluster • Pre-purchase consulting and procurement assistance • Initial setup and configuration • On-going system admin and security support • A la carte services – recharge • Backup and archive • Services not provided by program • Applications porting support

  4. User Perspective What do they get? • Computer room location • Expertise, consulting, setup, configuration at $0 • Consistent Sys admin and security support for remainder of program • Undecided whether modest recharge for this, weighing tradeoffs What does it cost them? • Electric/space post pilot • $0-modest recharge during pilot for system admin and security support (10 nodes/~ $6K) • Full recharge for support post pilot) (depending on support level and configuration, e.g. 10 nodes/~$12K-$24K) Why would they want to do this? • Less expensive than if they did it themselves (10 nodes/~$0-$25K for ¼ postdoc) even though it’s a hidden cost (see next slide) • Guarantees a properly working and secure system • Incentive to make change to cluster computing, increase science output

  5. Lab Perspective Economy-of-scale cost savings • Leveraging of technical expertise • Standardized approach to support saves $$ Hidden cost savings • Allows scientists to do science instead of systems administration • Minimizes the guesswork in purchase, configuration decisions • Results in more computational throughput Other benefits • Motivates paradigm shift to cluster and parallel computing • Access to MRC computing helps keep our scientists competitive with scientists at other labs • Funding based on analysis of current needs. Expanding this program beyond 4-6 new systems will require additional funds

  6. Next Steps • ITSD currently refining proposals and needs help lining up applicant projects • Gather information about applicants • Workshop survey • MRC/CSAC identified • ITSD identified • Prioritize applicants based on criteria • Check for success factors • Create natural groupings and associated budget • Present proposal in September

  7. How Can MRC-WG Help? • Help refine application • Help identify applicants • Conduct project interviews to gather information • Relay information to ITSD for prioritization • Note: • Competitive process • Program will fund limited number of applicants

  8. Other Information • Communicate expectations for projects participating in program, .e.g. • Must exceed minimal size • Root access • Conform to standard configuration • Existing and new clusters are candidates • Timeline • July 29 – Today’s talk • August 5 – MRC-WG members receive applications and supporting materials • August 23 – Deadline for applications to be returned to ITSD • Accurate and thoughtful input on applications critical • Success of project in prioritized list • Success of resulting proposal to lab management • This program should not preclude institutional system at some future point if that is “the right thing”

More Related