Workshops Archive

Comparing Distributed-Memory Programming Frameworks with Radix Sort

Workshop: The 8th Annual Parallel Applications Workshop: Alternatives To MPI+X (PAW-ATM 2025)

Authors: Michael P. Ferguson (Hewlett Packard Enterprise (HPE)), Ryan D. Friese (Pacific Northwest National Laboratory (PNNL)), and Shreyas Khandekar and Matt Drozt (Hewlett Packard Enterprise (HPE))

Abstract: Distributed-memory parallel processing addresses computational problems requiring significantly more memory or computational resources than can be found on one node. Software written for distributed-memory parallel processing typically uses a distributed-memory parallel programming framework to enhance productivity, scalability, and portability across supercomputers and cluster systems.

These frameworks vary in their capabilities and support for managing communication and synchronization overhead to achieve scalability. This paper employs a communication-intensive distributed radix sort algorithm to examine and compare the performance, scalability, usability, and productivity differences between five distributed-memory parallel programming frameworks: Chapel, MPI, OpenSHMEM, Conveyors, and Lamellar. The Chapel implementation has the fewest source lines of code (113) and is the most performant on 128 nodes of an HPE Cray Supercomputing EX (achieving about 17 billion elements sorted per second). The source code is available at https://github.com/mppf/distributed-lsb, and we welcome contributions, including optimizations to the implementations and results from runs on different systems.

Back to The 8th Annual Parallel Applications Workshop: Alternatives To MPI+X (PAW-ATM 2025) Archive Listing Back to Full Workshop Archive Listing