The IEEE Computer Society Publications Board recently announced the winner of the 2025 Best Paper Award for IEEE Transactions on Parallel and Distributed Systems (TPDS). The paper “Leveraging Graph Analysis to Pinpoint Root Causes of Scalability Issues for Parallel Applications,” published by the Institute of High Performance Computing in Tsinghua University’s Department of Computer Science and Technology, received the award. The first author of this paper is Yuyang Jin, an assistant researcher in the department, and the corresponding author is Professor Jidong Zhai.
On large-scale supercomputers, the scalability of parallel applications is often constrained by load imbalance, resource contention, inter-process communication, and other factors. Existing performance-analysis approaches each have their advantages and disadvantages: sampling-based analysis incurs little overhead but cannot identify root causes, while tracing-based analysis provides comprehensive information but carries prohibitively high overhead. To get the best of both worlds, the paper proposes a lightweight, graph-based method for identifying scalability bottlenecks in parallel programs. By combining the advantages of static analysis and lightweight runtime sampling, the method constructs a Parallel Performance Graph and uses a backtracking algorithm to automatically locate the root causes of scalability issues. ScalAna, a performance‑analysis tool built upon this technique, improves the system’s ability to pinpoint scalability bottlenecks on supercomputers. This technique has already been applied to detecting performance bottlenecks in large-scale parallel applications at Chinese national supercomputing centers.
IEEE Transactions on Parallel and Distributed Systems (TPDS) is a top journal in the field of high-performance computing. It covers parallel and distributed computing algorithms, computing applications, performance analysis, and related fields, and has broad international academic influence. Of the 193 papers published in IEEE TPDS in 2025, only one received the journal’s Best Paper Award, marking a competitive award ratio of merely 0.52%.
Editors: Li Han, John Paul Grima
