A Compact, Dynamic, and Distributed GPU Data Structure Library

Compact, dynamic, and distributed data structures that exploit the massive parallelism of GPUs.

GPUs offer enormous parallelism but are notoriously difficult to use for dynamic, pointer-based data structures. This project builds a library of compact and dynamic GPU data structures, together with the memory-management substrate they need.

We built Gallatin (McCoy & Pandey, 2024), a general-purpose GPU memory manager, and WarpSpeed (Mccoy & Pandey, 2026), a high-performance library of concurrent GPU hash tables. On top of these we have designed high-performance filters for GPUs (McCoy et al., 2023) and distributed-memory $k$-mer counting on GPUs (Nisa et al., 2021).

References

2026

  1. ALENEX 2026
    WarpSpeed: A High-Performance Library for Concurrent GPU Hash Tables
    Hunter Mccoy and Prashant Pandey
    In SIAM Symposium on Algorithm Engineering and Experiments, 2026
    Best Artifact Award PDF Code

2024

  1. PPOPP 2024
    Gallatin: A General-Purpose GPU Memory Manager
    Hunter McCoy and Prashant Pandey
    In Proceedings of the 29th ACM SIGPLAN Annual Symposium on Principles and Practice of Parallel Programming, PPoPP 2024, Edinburgh, United Kingdom, March 2-6, 2024, 2024
    DOI PDF Code

2023

  1. PPOPP 2023
    High-Performance Filters for GPUs
    Hunter McCoy, Steven A. Hofmeyr, Katherine A. Yelick, and 1 more author
    In Proceedings of the 28th ACM SIGPLAN Annual Symposium on Principles and Practice of Parallel Programming, PPoPP 2023, Montreal, QC, Canada, 25 February 2023 - 1 March 2023, 2023
    DOI PDF Code

2021

  1. IPDPS 2021
    Distributed-Memory k-mer Counting on GPUs
    Israt Nisa, Prashant Pandey, Marquita Ellis, and 3 more authors
    In 35th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2021, Portland, OR, USA, May 17-21, 2021, 2021
    DOI Code