Search NASASearch

Engineering topics

Craw, James

Publications and source records attributed to Craw, James.

Performance Evaluation Tools for Next Generation Scalable Computing Platforms

The Federal High Performance and Communications (HPCC) Program continue to focus on R&D in a wide range of high performance computing and communications technologies. Using its accomplishments in the past four years as building blocks towards a Global Information Infrastructure (GII), an Implementation Plan that identifies six Strategic Focus Areas for R&D has been proposed. This white paper argues that a new generation of system software and programming tools must be developed to support these focus areas, so that the R&D we invest today can lead to technology pay-off a decade from now. The Global Computing Infrastructure (GCI) in the Year 2000 and Beyond would consists of thousands of powerful computing nodes connected via high-speed networks across the globe. Users will be able to obtain computing in formation services the GCI with the ease of using a plugging a toaster into the electrical outlet on the wall anywhere in the country. Developing and managing the GO requires performance prediction and monitoring capabilities that do not exist. Various accomplishments in this field today must be integrated and expanded to support this vision.

Yan, Jerry C.

Are Event Traces Really That Necessary?

Many performance monitoring tools are currently available to the super-computing community. The performance data gathered and analyzed by these tools fall under two categories: statistics and event traces. Statistical data is much more compact but lack the probative power event traces offer. Event traces, on the other hand, can easily fill up the entire file system during execution such that the instrumented execution have to be terminated. In this paper, we propose an innovative methodology for monitoring and trace representation that offers a middle ground. The user can trace-off trace data size vs. quality incrementally. Specifically, the user will be able to limit the amount of trace collected and, at the same time, carry out some of the analysis event traces offer for the entire execution. With the help of a few CFD examples, we illustrate the use of our technique in performance tuning. We also compare quantitatively, the quality of the traces we collected vs. event traces.

Schmidt, Melisa

Event-Based Study of the Effect of Execution Environments on Parallel Program Performance

In this paper we seek to demonstrate the importance of studying the effect of changes in execution environment parameters, on parallel applications executed on state-of-the-art multiprocessors. A comprehensive methodology for event-based analysis of program behavior is introduced. This methodology is used to study the performance significance of various system parameters such as processor speed, message-buffer size, buffer copy speed, network bandwidth, communication latency, interrupt overheads and other system parameters. With the help cf a few CFD examples, we illustrate the use of our technique in determining suitable parameter values of the execution environment for three applications. We also demonstrate how this approach can be used to predict performance across architectures and illustrate the use of visual and profile-like feedback to expose the effect of system parameters changes on the performance of specific applications module.

Sarukkai, Sekhar R.