July 29, 2021 — ACM’s Special Interest Group on High Performance Computing (SIGHPC) is pleased to announce that Dr. Rohit Zambre of the University of California, Irvine has won the 2021 SIGHPC Doctoral Dissertation Award. Dr. Zambre’s dissertation is titled “Exascalable Communication for Modern Supercomputing.” This award is given each year for the best doctoral dissertation completed in high performance computing (HPC) in the previous year, and includes a $2,000 cash prize, a plaque, and recognition at the International Supercomputing Conference (SC’XY) in November. Nominations were evaluated on technical depth, the significance of the research contribution, the potential impact on theory and practice, and overall quality of work.
This year’s award is presented for outstanding contributions to the design, development, and measurement of a new MPI+threads library for scalable communication of multithreaded applications on current supercomputers. Dr. Zambre’s dissertation, titled “Exascalable Communication for Modern Supercomputing,” analyzes the decade-long problem of slow multithreaded communication in supercomputing applications, and eliminates the communication bottleneck by bridging the two ends of the HPC stack – MPI library developers and domain experts – that typically do not talk to each other directly. Dr. Zambre collaborated with MPI library developers at Argonne National Laboratory to make high-speed multithreaded communication a reality (ICS’20), and he collaborated with developers of applications targeted to run on the upcoming exascale machines to achieve significant performance improvements by utilizing the new MPI library (TPDS’21).
Dr. Zambre received his PhD in Computer Engineering from UC Irvine in 2020. He has now joined AMD as an HPC Architecture Researcher. He authored the first communication-oriented study on an Arm-based server (ICPP’19) and received the Best Poster Presentation award for his low-level multithreaded communication study at ICPADS’18.
“I am grateful to SIGHPC for the recognition of my efforts,” said Dr. Zambre. “I hope supercomputing applications are able to benefit from the case studies in the dissertation and utilize the new high-speed multithreaded performance capabilities of MPI libraries.”
“We are delighted to grant this year’s Dissertation Award to Rohit. His work exemplifies the kind of HPC research that SIGHPC believes makes our community more effective, and helps to raise the standards of the profession,” said John West, SIGHPC Chair. “The applicant pool was made up of outstanding recent PhDs, whom we hope will continue to make exciting contributions to high performance computing techniques and technologies in the future.”
In addition, SIGHPC selected two submissions for Honorable Mention in the 2021 Dissertation Award.
Dr. AJ Lauer (Creighton University) was selected for contributions to knowledge of the workplace climate among racial and gender minorities in the field of high performance computing. Her thesis, titled, Cyberinfrastructure Professionals’ Withdrawal Cognitions: An Exploration of Race, Gender, and Occupational Climate in High Performance Computing, explored what elements of occupational climate should be addressed in order to have a positive impact on the retention and promotion of individuals from underrepresented groups in HPC.
Dr. Yang You (University of California Berkeley) was selected for developing LARS (Layer-wise Adaptive Rate Scaling) and LAMB (Layer-wise Adaptive Moments for Batch training) to accelerate machine learning on HPC platforms. His thesis, entitled, “Fast and Accurate Machine Learning on Distributed Systems and Supercomputers” focuses on improving the speed and accuracy of Machine Learning training to optimize the use of parallel programming on supercomputers. He is currently a Presidential Young Professor in Computer Science at the National University of Singapore.
About this Award:
This award is open to students studying anywhere in the world who have completed a PhD dissertation with HPC as a central research theme. In this case, HPC refers to the study or application of computational capabilities delivering much higher performance or larger scales than could be accomplished with a desktop or simple server system in order to solve large problems. Nominations for the next Doctoral Dissertation Award will open in August 2021.
Click here to learn more.