TUSCALOOSA, Ala. – A University of Alabama assistant professor of computer science was awarded three Best Paper Awards at the 2018 Institute of Electrical and Electronics Engineers International Conference on Cluster Computing including winning the Overall Best Paper Award.
Dr. Dingwen Tao co-authored two papers submitted to the 2018 Cluster Conference held in Belfast, United Kingdom. Both submissions received best paper in their respective tracks. One paper was awarded first in the area of Data, Storage and Visualization, and the second paper won both in the area of Applications, Algorithms and Libraries as well as the conference Overall Best Paper Award.
“I felt very excited when I heard the general chair announced our awards,” Tao said. “This is the first time we won the best paper award in a top-tier international conference.”
The IEEE Cluster Conference is a major international forum for presenting recent accomplishments and technological developments in the field of cluster computing. A total of 154 papers were submitted and 44 of those were accepted into the conference.
The overall winning paper is titled “PaSTRI: Error-Bounded Lossy Compression for Two Electron Integrals in Quantum Chemistry.” This paper developed a fast and effective data compression algorithm for two-electron repulsion integrals in a parallel quantum chemistry simulation.
“Previously, some domain scientists were concerned about the feasibility of using lossy compression for scientific data in real-world HPC applications,” Tao said. “This award means that our community has been gradually understanding and accepting that our research in lossy compression can help domain scientists scale mission-critical applications on extreme-scale systems to achieve significant advances in science, engineering, technology, medicine and industry.”
The second paper is titled “An Efficient Transformation Scheme for Lossy Data Compression with Point-wise Relative Error Bound.” This paper proposed a transformation scheme that can transfer pointwise relative-error-bounded problems to an absolute-error-bounded compression issue.
Tao says the next step for his teams’ research is to continue developing more powerful, efficient and effective lossy compression algorithms and software to serve broader scientific applications.