Evaluation of Early Systems Research Activities
Evaluation of Early Systems Activities
The EES project has five major
components.
Acquisition
- Identify and acquire access to or procure new architectures that hold
promise as platforms for solving CCS computational Grand Challenges
Beta testing
- Test and evaluate pre-production hardware and software, working with
vendors to eliminate bugs and improve performance.
- Recent results
- Acceptance of Paragon XPS/150 system with 1024 MP nodes.
-
Publications
Benchmarking and evaluation
- Measure performance of components and of full systems, to
- determine peak and achievable computation rates, and determine
what is required to achieve these rates.
- evaluate the suitability of the architecture for CCS Grand
Challenge application codes, and
- make comparative analyses with other advanced architectures.
- Recent results
-
Publications
- Ongoing efforts
- Evalution of MPI.
- Evaluation of parallel I/O on Intel Paragon MP node system.
- Evaluation of sensitivity of performance to changes in
latency and bandwidth.
- Comparison of SUNMOS versus NX on MP node systems.
- Evaluation of different programming paradigms or styles on the Intel
Paragon, the IBM SP2, and the Cray Research T3D.
Modeling
- Construct predictive performance models, to
- validate understanding of how computer system works and how it can be
used most efficiently, and
- determine how performance will scale when larger systems are used to
solve larger instances of the Grand Challenge problems.
-
Publications
Information Dissemination
- Provide guidance to CCS staff and users
on how to maximize performance when using CCS machines.
- Provide data to CCS and DOE on interesting advanced
architectures for the next generation of CCS computing resources.
- Recent results
- Evaluation of Early Systems WWW pages
Evaluation of Early Systems Project
Patrick H. Worley / (worleyph@ornl.gov)
Last Modified Tuesday, 12-Sep-2000 13:01:11 EDT.
