Three papers coauthored by researchers in the Mathematics and Computer Science Division at Argonne National Laboratory were highlighted recently on the U.S. Department of Energy Exascale Computing Project website.
- Rob Ross, a senior computer scientist, was coauthor of “Improving MPI collective I/O for high volume non-contiguous requests with intra-node aggregation.” The paper, published in IEEE Transactions on Parallel and Distributed Systems 31 (11): 2682–2695, presents a new design for collective I/O by adding an extra communication layer that performs request aggregation among processes within the same compute nodes. The method reduces the communication cost and hence maintains the scalability for a large number of processes.
- Sheng Di, a computer scientist, and Franck Cappello, a senior computer scientist, were coauthors of “Significantly improving lossy compression for HPC datasets with second-order prediction and parameter optimization.” The researchers developed second-order prediction methods based on Lorenzo prediction and regression prediction and formulated an algorithm that can select the best-fit predictors and optimized parameter settings at runtime. The paper was published in the Proceedings of the 29th International Symposium on High-Performance Parallel and Distributed Computing, 89–100, 2020.
- Misun Min, a computational scientist, and Paul Fischer, a senior computational scientist, were coauthors of the paper “CEED project examines scalability and performance of key algorithms for HPC applications.” They developed community benchmarks for a set of operations common to many simulation codes used in nuclear reactor modeling, compressible flow, wind energy, and climate modeling. The paper was published in the International Journal of High Performance Computing Applications 34(5): 562–586, 2020.
For further information, see the DOE ECP website: https://www.exascaleproject.org/featured-publication-summaries/.