HPCwire: Move Up the Scale

The Leading Source for Global News and Information Covering the Ecosystem of High Productivity Computing

Leading HPC
Solution Providers

HPCwire >> Topic >> Systems

August 17, 2007

Move Up the Scale

Scientists Benefit From CPU-Hour Reimbursement Program to Improve Code Performance

NERSC's large-scale reimbursement program has provided nearly six million computing hours so far this year to 21 projects that have taken advantage of the opportunity to scale their runs in preparation for using the new Cray XT4 system.

The program has set aside nine million computing hours on Seaborg this year so far to encourage researchers to improve their codes for large runs on the new Cray, Franklin, once it enters production later this year.

The incentive has attracted projects from a variety of scientific disciplines including astrophysics, life sciences, fusion, chemistry and climate research. Scientists say the program has enabled them to pinpoint and resolve issues before running jobs on Franklin.

"We have participated in the Scaling Reimbursement program in order to tackle the problem of testing the predictive power of new empirical force fields for biomolecular simulation. Our system of interest is the Abeta peptide and various sub-peptides which are associated in the formation of amyloid plaques in Alzheimer's Disease," said Nicolas Lux Fawzi, a UC Berkeley scientist on a research team led by Teresa Head-Gordon at Berkeley Lab. "We have used the CPU time in the program to demonstrate that we can run large parallel simulations on 1,024 processors using the replica exchange technique to generate the complete equilibrated ensemble for our system at a range of temperatures.

"We have very much enjoyed the chance to work with NERSC as part of the scaling program. We've received some excellent input from the NERSC staff on how to evaluate and improve the scaling of the code."

Franklin, which arrived at NERSC earlier this year, has more than 19,000 processors and can deliver 10 times more computing power than any existing NERSC system. Franklin can provide a sustained performance of at least 16 trillion calculations per second, with a theoretical peak speed of more than 100 teraflop/s.

The reimbursement program is open to all NERSC users and requires researchers to run 1,024- to 1,500-processor jobs on Seaborg, depending on whether they have participated in the program in the past. The target is to carry out a run on at least 2,418 CPUs. Each project can get a maximum reimbursement of 500,000 hours this year. Scientists also have to use the Integrated Performance Monitoring (IPM) software to gather performance information about each run.

"The quantum Monte Carlo methods developed in the Lester group are naturally amenable to parallel computing. Historically, our production jobs have run on several hundred processors with near perfect parallel efficiency," said Brian Austin, a researcher in a group led by William A. Lester, Jr., a UC Berkeley chemistry professor and Berkeley Lab researcher. "The advent of near-petascale computers such as Franklin will bring jobs with thousands of processors into the norm. In this regime, subtle changes to our mode of parallel communication have dramatic effects that were unnoticeable at previous scales: communication time increased from 2 percent to almost 50 percent as the number of processors increased from 512 to 2,048. The reimbursement program has been essential to our exploration and resolution of these issues."

Don Lamb, a University of Chicago scientist who leads a supernovae research project, also is an active participant in the reimbursement program. Lamb is a recipient of this year's Innovative and Novel Computational Impact on Theory and Experiment (INCITE) awards, a program by the DOE Office of Science to support large-scale projects in national labs, universities and industry.

Lamb's research team members said the IPM wasn't easy to use initially, but they overcame those issues with the support of the NERSC staff.

"We had to pay more attention to exit codes issued by FLASH than we had previously, since non-zero exit codes force IPM to throw away all output. But the NERSC staff was helpful and understanding of IPM issues, never letting missing IPM output interfere with reimbursement. Where available, the IPM output supplied interesting profiling information," said Carlo Graziani, a researcher on Lamb's team.

Other scientists whose projects are among the top 10 recipients of reimbursed hours are George Vahala (fusion plasma), Doug Toussaint (quantum chromodynamics), Stephen Gray (chemistry -- nanoscale electrodynamics), Cameron Geddess (accelerator physics), Wei-li Lee (fusion plasmas) and Paola Cessi (climate research).

More information about the reimbursement program can be found at http://www.nersc.gov/hypermail/all-announcements/0755.html.

-----

Source: NERSC News (http://www.nersc.gov/news/nerscnews)

Article Tools

Share & Save Options

Discussion

There are 0 discussion items posted.

Join the Discussion

You must be a registered user in order to comment on HPCwire stories.

Members enjoy:

	Access to Special Content, including Podcasts, Webcasts, and Whitepapers and more!
	Ability to take an active role in the discussions that are shaping the HPC technology discipline!
	Special offers available only to HPCwire members!

Become a Registered User Today!

Registered Users Log in Here to Comment

Email Address:	Password (case sensitive)

Remember me	Forgot Password?

RSS Feeds

Feeds by Topic

Feeds by Industry

Feeds by Content Type

Feature Articles

The New Science of Visual Analytics

When Jim Thomas set out to find new ways to deal with the mountains of information our society generates, he didn't just create a new organization, he created a new science. In this article we'll take a look at how the National Visualization and Analytics Center is transforming the problem of finding needles in haystacks into an opportunity for a more secure future.
Read More...

The Week in Review

ORNL Jaguar doubles its performance; the SC08 Cluster Challenge is gearing up; the University of Central Florida uses Army dollars to purchase an IBM super; and IBM's RoadRunner prepares to break the petaflop barrier. John West recaps those stories and more in our weekly wrap-up.
Read More...

Revaluating FPGAs for 64-bit Floating-Point Calculations

We now have generally available 2.5 GHz quad-core Opterons and Virtex-5 LX330, SX95T and recently announced SX240T FPGAs. In addition to this, Xilinx is releasing a new version of their floating-point cores that reduces the amount of logic and DSP slices needed for building floating-point function units. Taken together it is time to revisit Opteron floating-point performance versus FPGA performance.
Read More...

Read more HPCwire features...

Top Headlines

Parallel Processing Calls for a Fortress Mentality

May 14 | InfoWorld | Sun Microsystems is taking the lessons learned from Java and applying them to the application development challenges of the high performance computing realm. Read more...

IBM Set to Test the Fastest Computer in the World

May 14 | Computerworld | IBM is assembling the final pieces of what they hope will soon become the world's most powerful supercomputer. Read more...

IBM Shifts Cell to 65 Nanometers

May 13 | EETimes | IBM Corp. has announced the next-generation version of its Cell processor, the first specifically geared for computer servers. Read more...

Record-Setting Simulations on Ranger Reconstruct the Reionization Era

May 12 | Texas Advanced Computing Center | In the coming months, Dr. Michael L. Norman of UCSD will use Ranger, the world's most powerful supercomputer for open-science research, to perform the largest cosmological simulation to date. Read more...

2 AMD Executives Out in Restructuring Amid Slump

May 12 | BusinessWeek | Two executives have left AMD, including the head of the slumping chip maker's microprocessor division, as the company tries to engineer a dramatic turnaround to fend off larger rival Intel Corp. Read more...

Read more headlines...

Multimedia

Podcast: Interview with Ben Bennett of ClearSpeed Technology

Today, HPC organizations are requiring substantially more floating point performance to solve real-world problems. In this podcast, Ben Bennett, ClearSpeed General Manager, discusses how acceleration technology can improve the overall performance of standard x86-based systems...