Visit additional Tabor Communication Publications
June 02, 2010
WEST LAFAYETTE, Ind., June 1 -- Purdue University researchers, working with high-performance computing experts at Lawrence Livermore National Laboratory, have created an automated program to "debug" simulations used to more efficiently certify the nation's nuclear weapons.
The program, called AutomaDeD (pronounced like automated), finds errors in computer code for complex "parallel" programs.
"The simulations take several weeks to run, and then they have to be debugged to correct errors in the code," said Saurabh Bagchi, an associate professor in Purdue's School of Electrical and Computer Engineering. "The error might have occurred in the first hour of operation, and if you had known about it you could have stopped it then."
Because international treaties forbid the detonation of nuclear test weapons, certification is done using complex simulations. The simulations, which may contain as many as 100,000 lines of computer code, must accurately show reactions taking place on the scale of milliseconds, or thousandths of a second.
"Many times an error in a simulation code may not become evident until long after it occurs," said Bronis R. de Supinski, co-leader of the ASC Application Development Environment Performance Team at the U.S. Department of Energy's Lawrence Livermore National Laboratory. "These delays are challenging since they make the actual location of the bug unclear."
In parallel operations used for powerful simulation tools, a highly complex job is split into numerous smaller and more manageable processes that are handled by separate machines in large computer clusters. After the computers complete their individual processes, all of the parallel results are combined.
Conventional debugging programs, however, must be operated manually, with engineers navigating through a large number of processes.
"Debuggers have worked well for sequential applications," Bagchi said. "But when we extend these to large parallel applications, application developers are not very happy because it's very time consuming and difficult to do the manual debugging. It is just difficult for human cognitive abilities to keep track of what is going on simultaneously in many processes and determine what is anomalous."
So, to enable the automatic debugging of the simulations, the researchers created AutomaDeD, which stands for automata-based debugging for dissimilar parallel tasks.
"The idea is to use AutomaDeD as the simulation is running to automatically monitor what's happening," Bagchi said. "If things start going wrong, AutomaDeD would stop and flag which process and which part of the code in the process is likely anomalous."
Errors in software code cause "stalls" and "hangs" that slow or halt simulations or give incorrect results. Another problem with parallel programs is interference from software that previously ran on the same computer clusters but were not properly expunged before the new job started running.
Recent research findings show AutomatDeD was 90 percent accurate in identifying the time "phase" when stalls and hangs occurred; 80 percent accurate in identifying the specific tasks that were the sources for stalls and hangs; and 70 percent accurate in identifying the interference faults.
The findings will be detailed in a research paper to be presented on June 30 during the 40th Annual IEEE/IFIP International Conference on Dependable Systems and Networks in Chicago. The paper was written by Purdue doctoral student Ignacio Laguna, Bagchi, and Lawrence Livermore scientists Greg Bronevetsky, de Supinski, Dong H. Ahn and Martin Schulz. The primary developers of the program are Bronevetsky and Laguna.
The same debugging approach could be used to find errors in other parallel applications, such as those used in climate modeling and high-energy particle physics.
AutomaDeD works first by grouping the large number of processes into a smaller number of "equivalence classes" with similar traits. Grouping the processes into equivalence classes keeps the analysis simple enough that it can be done while the simulation is running.
AutomataDeD also works by splitting a simulation into numerous windows of time, called phases.
"So our tool lets you know if the error occurs for task 1 and task 5 in phase 153 and allows you to zoom in and find the specific part of the code that is problematic," Bagchi said.
Large computer clusters operated by Lawrence Livermore containing thousands of processors have been used for the debugging operations.
Purdue researchers did not work with the actual classified nuclear weapons software code but instead used generic "NAS parallel benchmarks," a set of programs designed to help evaluate the performance of parallel supercomputers developed by the NASA Advanced Supercomputing division.
The work began a year ago, and the new debugging program is currently being used by the federal lab to detect certain types of errors. The researchers are continuing to improve the program. The work is funded by the Department of Energy.
Also involved in the ongoing work is Karu Sankaralingam, an assistant professor of computer sciences and electrical and computer engineering at the University of Wisconsin at Madison.
Related Web sites:
Surabh Bagchi: https://engineering.purdue.edu/ECE/People/profile?resource_id=3261
The abstract on the research in this release is available at http://www.purdue.edu/newsroom/research/2010/100601BagchiNuclear.html.
Source: Purdue University
Contributing commentator, Andrew Jones, offers a break in the news cycle with an assessment of what the national "size matters" contest means for the U.S. and other nations...
Today at the International Supercomputing Conference in Leipzing, Germany, Jack Dongarra presented on a proposed benchmark that could carry a bit more weight than its older Linpack companion. The high performance conjugate gradient (HPCG) concept takes into account new architectures for new applications, while shedding the floating point....
Not content to let the Tianhe-2 announcement ride alone, Intel rolled out a series of announcements around its Knights Corner and Xeon Phi products--all of which are aimed at adding some options and variety for a wider base of potential users across the HPC spectrum. Today at the International Supercomputing Conference, the company's Raj....
Jun 18, 2013 |
The world's largest supercomputers, like Tianhe-2, are great at traditional, compute-intensive HPC workloads, such as simulating atomic decay or modeling tornados. But data-intensive applications--such as mining big data sets for connections--is a different sort of workload, and runs best on a different sort of computer.
Jun 18, 2013 |
Researchers are finding innovative uses for Gordon, the 285 teraflop supercomputer housed at the San Diego Supercomputer Center (SDSC) that has a unique Flash-based storage system. Since going online, researchers have put the incredibly fast I/O to use on a wide variety of workloads, ranging from chemistry to political science.
Jun 17, 2013 |
The advent of low-power mobile processors and cloud delivery models is changing the economics of computing. But just as an economy car is good at different things than a full size truck, an HPC workload still has certain computing demands that neither the fastest smartphone nor the most elastic cloud cluster can fulfill.
Jun 14, 2013 |
For all the progress we've made in IT over the last 50 years, there's one area of life that has steadfastly eluded the grasp of computers: understanding human language. Now, researchers at the Texas Advanced Computing Center (TACC) are utilizing a Hadoop cluster on its Longhorn supercomputer to move the state of the art of language processing a little bit further.
Jun 13, 2013 |
Titan, the Cray XK7 at the Oak Ridge National Lab that debuted last fall as the fastest supercomputer in the world with 17.59 petaflops of sustained computing power, will rely on its previous LINPACK test for the upcoming edition of the Top 500 list.
05/10/2013 | Cleversafe, Cray, DDN, NetApp, & Panasas | From Wall Street to Hollywood, drug discovery to homeland security, companies and organizations of all sizes and stripes are coming face to face with the challenges – and opportunities – afforded by Big Data. Before anyone can utilize these extraordinary data repositories, however, they must first harness and manage their data stores, and do so utilizing technologies that underscore affordability, security, and scalability.
04/15/2013 | Bull | “50% of HPC users say their largest jobs scale to 120 cores or less.” How about yours? Are your codes ready to take advantage of today’s and tomorrow’s ultra-parallel HPC systems? Download this White Paper by Analysts Intersect360 Research to see what Bull and Intel’s Center for Excellence in Parallel Programming can do for your codes.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?
Join our webinar to learn how IT managers can migrate to a more resilient, flexible and scalable solution that grows with the data center. Mellanox VMS is future-proof, efficient and brings significant CAPEX and OPEX savings. The VMS is available today.