GSST: Parallel string decompression at 191 GB/s on GPU

Robin Vonk, Joost Hoozemans, Zaid Al-Ars

Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientificpeer-review

Abstract

Most of the commonly used compression standards make use of some form of the LZ algorithm. Decompressing this type of data is not a good match for the Single-Instruction, Multiple Thread (SIMT) model of computation used by GPUs, resulting in low throughput and poor utilization of the GPU parallel compute capabilities. In this paper, we introduce GSST, a GPU-optimized version of the FSST compression algorithm, which targets string compression. The optimizations proposed in this paper make the algorithm particularly suitable for GPUs, which allows it to achieve a significantly better tradeoff for decompression throughput vs compression ratio as compared to the state of the art. Our results show that the new algorithm pushes the Pareto curve closer towards the ideal region, completely dominating LZ-based compressors in the nvCOMP library (LZ4, Snappy, GDeflate). GSST provides a compression ratio of 2.74x and achieves a throughput of 191 GB/s on an A100 GPU.
Original languageEnglish
Title of host publicationCHEOPS '25: Proceedings of the 5th Workshop on Challenges and Opportunities of Efficient and Performant Storage Systems
Place of PublicationNew York, NY
PublisherAssociation for Computing Machinery, Inc
Pages8-14
Number of pages7
ISBN (Electronic)979-8-4007-1529-7
DOIs
Publication statusPublished - 2025
Event5th Workshop on Challenges and Opportunities of Efficient and Performant Storage Systems, CHEOPS 2025 - Rotterdam, Netherlands
Duration: 31 Mar 202531 Mar 2025
https://cheops-workshop.github.io/2025.html

Conference

Conference5th Workshop on Challenges and Opportunities of Efficient and Performant Storage Systems, CHEOPS 2025
Abbreviated titleCHEOPS 2025
Country/TerritoryNetherlands
CityRotterdam
Period31/03/2531/03/25
Internet address

Bibliographical note

Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Fingerprint

Dive into the research topics of 'GSST: Parallel string decompression at 191 GB/s on GPU'. Together they form a unique fingerprint.

Cite this