FEATURES & COMMENTARY LIVEwire
Dallas, Texas — The HPC User Forum, representing leading supercomputer users in government, industry and academia, has reported progress on a plan to create better performance tests for this most powerful class of computers. Speaking at the SC2000 supercomputing conference held here, HPC User Forum officials said improved tests are needed to advance scientific research, industrial engineering and classified government work, all of which rely heavily on supercomputers. The HPC (High-Performance Computing) User Forum is organized by IDC on behalf of HPC users.
“Despite the strategic and economic importance of supercomputers, or HPC systems, there is no widely accepted standard for evaluating their performance today,” said Debra Goldfarb, IDC group vice president, Worldwide Systems and Servers. “It is not uncommon for user organizations to find that the actual performance of a new, multi-million-dollar supercomputer is a small fraction of the stellar results produced on today’s limited tests. Although these Linpack and peak performance test scores are nearly meaningless beyond ‘chest-thumping’ publicity, in the absence of anything better they continue to be emphasized in many procurements.”
Team Engaged To Develop Benchmarks.
IDC analyst Earl Joseph II said the HPC User Forum has engaged a group headed by Robert Lucas, head of HPC Research at the Department of Energy’s National Energy Research Computing Center (NERSC), to develop a new benchmark suite to test the overall scalable performance of HPC systems, and an additional benchmark initiative to test performance on a broad range of specific application codes. In a mid-September meeting, HPC User Forum members endorsed the preliminary plan outlined by Lucas.
“Bob Lucas and the other members of this NERSC and University of Tennessee team are world-class experts at benchmarking HPC systems,” Joseph said. “They have years of leading experience and were involved in the development of the NAS Parallel Benchmarks and ‘Top500.'”
‘Short List’ Identified For Overall Scalable Performance Benchmarks.
Lucas reported that since the September meeting his team has identified a “short list” of candidate tests for the overall scalable performance benchmark suite. This suite is intended to measure processor performance, as well as internal memory speed, disks, and external networks.
“We will now bring this selection process quickly to a conclusion in collaboration with the HPC user community,” said Lucas. “For the processor (CPU) benchmark we intend to collaborate with the ‘Top500’ organization. Our objective is to supplement, not replace, the Linpack test on which the ‘Top500’ rankings are based,” Lucas added.
Development Schedules Outlined.
For the scalable benchmark suite, Lucas said, his team plans by the second quarter of 2001 to complete the following: finalize the selection of tests; engage HPC leaders to define an initial set of orthogonal benchmark kernels; develop pencil and paper descriptions and representative codes; and establish a Web site and publish initial results.
For the full application benchmark suite, in the same timeframe the NERSC and University of Tennessee team plan to identify an initial application and a domain expert partner; develop a generic methodology for benchmarking full applications; and deploy the prototype with the partner.
Benefits to Stakeholders.
Goldfarb said the new tests will benefit all stakeholders:
* The U.S. will be better able to maintain industrial and scientific competitiveness.
* Worldwide industries that use supercomputers (automotive, aerospace, chemical, genomics/bioinformatics, petroleum, pharmaceutical and weather) will be better able to address their most complex, economically important problems.
* Major supercomputing centers will be better able to support large, diverse user communities.
* All user organizations will be better able to know which third-party applications perform best on which HPC systems.
* Applications developers will be better able to know which HPC systems are most suitable for their needs.
* HPC vendors will not need to increase benchmarking resources; they can focus on the applications domains that are important for their markets.
About the HPC User Forum
The HPC User Forum ( http://www.idc.com/hpc ) is an initiative organized by IDC at the request of high-performance computing users in industry, government and academia. The Forum believes that ensuring the long-term health of the HPC industry is crucial for scientific and industrial progress.
About NERSC
The National Energy Research Scientific Computing Center (NERSC) at Lawrence Berkeley National Laboratory is one of the nation’s most powerful unclassified computing resources and is a world leader in accelerating scientific discovery through computation. Funded by the U.S. Department of Energy’s Office of Science, NERSC currently serves about 2,000 researchers at national laboratories, universities, and industries across the country.
About the “Top500” List
The list of the “World’s Top500 Supercomputing Sites” is compiled, based on results of the standard Linpack benchmark test, and published semi-annually by Jack Dongarra, Hans Meuer and Erich Strohmaier from the universities of Tennessee and Mannheim (Germany).
About IDC
IDC is the world’s leading provider of information technology (IT) industry analysis, market data and insight, and strategic and tactical guidance to builders, providers and users of information technology. IDC provides global research with local content through more than 600 analysts in 43 countries worldwide. IDC is a division of IDG, the world’s leading IT media, research, and exposition company.
—
The following brief background notes and supplementary points were compiled by Tim Staub, associate editor:
hpc@idc User Forum Background
Concept from PITAC
-Need to revitalize HPC industry
-Need for industry, government and univercity HPC users to join together
Announced last year at SC’99
Quarterly planning meetings
-User driven agenda
First full user forum meeting Sept 18th/19th
-Over 85 members and growing
-Approved performance approach
Initial HPC User Issue Areas:
“Sizing” Computers
Learn what other users are doing
Collaboration among users
Leverage Technology
Leverage HPC Vendors
-and academia & ISV’s
Map Applications to Architectures
The Economics of Supercomputing
What are the Goals?
IDC2000 Initiative has two parts:
#1 Develop and promote a more meaningful industry standand metric
-for use in comparing computers
-for price/performance calculations
-to offset the problems using peak
#2 Creat a matrix with many metrics and “real” applications results
– An indicator for actual procurement selection
– Shows advantages of all types of HPC computers
-Provide a tool for users to more quickly identify which computer architectures fit their application needs
Why Worry About HPC Performance Metrics?
HPC Users identified it as one of the top issues
– Causes troubles in purchases and funding
In 2000 the range in prices as a ratio to peak performance is OVER EIGHT TIMES
-list to peak SMP performance is: $6,300 to $49,000 US dollars per peak GFLOPS
Price as a ratio to peak performance is becoming less meaningful each year
-a metric with a range of eight fold is almost useless for comparing computers
-yet it is broadly used (and abused)
In summary:
IDC2000 Initiative:
-#1 Create a new HPC industry standard benchmark
-#2 Create a matrix of application benchmark results anolg with feeds and speeds
-Results from User Forum Members
-On real application jobs
-Watch the progress at: http://www.idc.com/hpc
============================================================