1 / 10

DOE2000 P.I. Meeting and Snowfest March 8-10, 1999

DOE2000 P.I. Meeting and Snowfest March 8-10, 1999. Group M-1 (also known as Group 5) Al Geist ORNL (chair) Rob Armstrong SNL Steve Benson ANL Les Cottrell SLAC Brett Didieu PNNL Steve Elbert NSF Ian Foster ANL Jarek Nieplocha PNNL

binta
Download Presentation

DOE2000 P.I. Meeting and Snowfest March 8-10, 1999

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. DOE2000 P.I. Meeting and SnowfestMarch 8-10, 1999 Group M-1 (also known as Group 5) Al Geist ORNL (chair) Rob Armstrong SNL Steve Benson ANL Les Cottrell SLAC Brett Didieu PNNL Steve Elbert NSF Ian Foster ANL Jarek Nieplocha PNNL Bahram Parvin LBNL Val Wilson NASA

  2. Projects in DOE2K Program Pilots MMC DC Collaboration Research Software Infrastructure (CIF) Shared Virtual Reality Collaboration Management Security Electronic Notebooks Floor Control QoS ACTS Toolkit Numerics Frameworks Runtime

  3. Process Each member discussed their opinions about the projects and ideas outside the projects such as data management, MICS programmatic needs, software economy, social vs. security issues. We discussed “farther out” topic areas in question B After hearing everyone else’s opinions on projects and knowing what topics we considered important for B, we went back and prioritized the research projects for A by each picking our top two choices.

  4. Question A Assume the entire DOE 2K program becomes part of the MICS “base budget” in 01. What do you see as the highest priorities for future activities in this budget. • Some general themes that arose as each member • expressed his feelings about the projects. • Collaboration Management -- commercial solutions preferred • Security -- two felt commercial solutions preferred, one said that security not as important to their problems. • Notebooks, CIF, and ACTS came up most often positively

  5. Highest Priorities for Future Activities (1of 2) • Electronic Notebook (3 first place votes) • popularity among broad user group means that continued improvements will have the big impact across research community • customized to be an ORB • components based so it can be easily customized work w/ PSE • Common Component Architecture (4 second place votes) • frequency indicates that it is perceived as an important activity • lack of first place votes indicates that CCA is not mature enough to bet the farm on.

  6. Highest Priorities for Future Activities (2 of 2) • Software Infrastructure (2 first place, 1 second place votes) • resource discovery and information split off as important part of CIF • ACTS Runtime (1 first place, 2 second place votes) • group felt Infrastructure and Runtime were very similar • integration between them is encouraged • ACTS Numerics (1 first place, 1 second place vote)

  7. Question B Assuming SSI, what are the most important topics areas that we need to include, beyond the scope of supporting specific SSI projects and above and beyond the expected activities covered under question A? Seven topic areas that fit the above scope were suggested.

  8. Seven Important Topic Areas Data management and analysis -disaster at Tbyte -> Pbyte scale New programming models - MPI effectiveness?, shmem, hierarchical New data distributions - exploit NUMA, easily try different ones Level of concurrency - how can applications and tools cope? Scalable OS services - in particular I/O and file systems Single CPU performance - algorithms efficiently use microarchitecture Distributed Computing and Grid research (on the edge) New paradigms - heterogeneous systems, usage models fault tolerance and recovery intelligent resource discovery and scheduling moving code to data, rather than data to code. ultra-high bandwidth

  9. Question C What key technical work do you see (funded by other agencies) that would best complement the plans identified in A and B? What are the organizational mechanisms we can use to get the key groups to collaborate?

  10. Key Technical Work (nonDOE funded) NASA - Information Power Grid DARPA - Quorum (includes Globus and Tau) NSF - compilers, OS research, digital libraries, human/computer interface NIH - bio-informatics How to get key groups to collaborate • Cross funding - for example Globus funded through NCSA • Be reviewers for other agency proposals. • Money is a pressure to push two to collaborate • the hope is that the collaboration will remain when • the money pressure is released.

More Related