Xipeng Shen: Then and Now / 2011 Early Career


Image: Xipeng Shen Winner of the 2011 Early Career Award
view more

Credit: Department of Energy Office of Science


The US Department of Energy’s (DOE) Early Career Award (ECA) gave me the opportunity to advance my understanding of storage performance in extreme-scale or exascale computing. This understanding is key to increasing the accuracy and scalability of many mission-critical scientific simulations running on modern supercomputers.

Exascale computing systems often integrate thousands of processing units into a single chip. Each of the computing chips requires data for processing. As the number of computing units continues to grow, the total demand for data is growing rapidly. But the speed at which data is transferred from memory to processors is increasing at a much slower rate. This growing gap fundamentally limits the currently achievable performance of exascale computing.

With ECA support I have developed some research directions to fill this gap. These instructions are primarily related to data reorganization and code optimization on graphics processing units (GPU), an important processor type in exascale systems.

One of the techniques deals with irregular memory during program executions. Irregular memory accesses read or write data without a pattern. They are not helpful for storage systems and affect memory access speed. The new technology transforms a program in such a way that its irregular accesses become regular at runtime. There is little overhead, but the data access speed increases greatly. Other techniques allow computer systems to flexibly manage the many parallel contexts on exascale systems to further speed up memory access.

These techniques have laid some important foundations for code optimizations on exascale systems. They have influenced the development and improvement of much modern software in high-performance computing and beyond. They have inspired many studies and received over 3,000 citations. By significantly reducing simulation times, these techniques have helped accelerate scientific research and discovery.


Xipeng Shen is Professor of Computer Science at North Carolina State University.


The Early Career Research Program provides financial support essential for researchers early in their careers, allowing them to define and direct independent research in areas important to DOE missions. The development of outstanding scientists and research leaders is of paramount importance to the Department of Energy Office of Science. By investing in the next generation of researchers, the Office of Science champions lifelong careers in discovery science.

See the Early Career Research Program for more information.


Improving data locality of dynamic simulations for exascale computing

Computer simulation is important for scientific research in many disciplines. Many such programs are complex and transfer a large amount of data in a dynamically changing pattern. Memory performance is key to maximizing computational efficiency in the age of chip multiprocessors (CMP) due to the growing mismatch between slowly expanding memory bandwidth and rapidly increasing data demands from processors.

The importance is underscored by the trend towards exascale computing, where processors are said to contain hundreds or thousands of (heterogeneous) cores each. Unfortunately, today’s computer systems lack support for high levels of memory transfer. This project proposes to improve memory performance of dynamic applications by developing two new techniques specifically tailored to the new characteristics of CMP.

The first technique is asynchronous streamlining, which analyzes an application’s memory reference patterns at runtime and regulates both control flows and memory references on the fly.

The second technique is neighborhood-aware locality optimization, which focuses on the uneven relationships between compute elements.

This research will yield a robust tool for scientific users to improve program locality on multi- and many-core systems, which is not possible with existing tools. In addition, it will contribute to the advancement of computer science and encourage academic research and education in the challenging field of scientific computing.


EZ Zhang, Y Jiang, Z Guo, K Tian, ​​and X Shen, “On-the-Fly Elimination of Dynamic Irregularities for GPU Computing.” Proceedings of the Sixteenth International Conference on Architecture Support for Programming Languages ​​and Operating Systemspages 369-380, (March 2011). [DOI: 10.1145/1950365.1950408]

G Chen, B Wu, D Li, and X Shen, “PORPLE: An extensible optimizer for portable data placement on the GPU.” The 47th Annual IEEE/ACM International Symposium on Microarchitecture(December 2014). [DOI: 10.1109/MICRO.2014.20]

G Chen, X Shen, B Wu, and D Li, “Optimizing Data Placement in GPU Memory: A Portable Approach.” IEEE transactions on computers 66(2017). [DOI: 10.1109/TC.2016.2604372]

DOE explains… provides simple explanations of key words and concepts in basic science. It also describes how these concepts apply to the work of the Department of Energy’s Office of Science as it helps the United States excel in research across the scientific spectrum. For more information about exascale computing and DOE’s research in this area, see “DOE Explains… Exascale Computing.”.”

Additional profiles of Early Career Research Program awardees can be found at /science/listings/early-career-program.

The Office of Science is the largest single funder of basic science research in the United States and works to address some of the most pressing challenges of our time. Visit www.energy.gov/science for more information.

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of the press releases published on EurekAlert! by contributing institutions or for the use of information about the EurekAlert system.


Comments are closed.