Productive, Performant, and Parallel Generic Lossy Data Compression with LibPressio

Student: Victoriana Malvoso (Clemson University)
Supervisor: Jon Calhoun (Clemson University)

Abstract: Data compression is vital in scientific applications because it reduces the size of files. This is important because running experiments on the data and storing it requires significantly less time and memory. In an effort to create more efficiency, LibPressio was created as a single interface to utilize multiple different compressors. This interface eliminates the need for generating large amounts of code to be able to switch between compressors. This is useful because time spent editing code could be used running more experiments. The goal of our testing is to evaluate metrics when using LibPressio as compared to not using it to determine performance. We evaluate how much LibPressio is able to reduce the size of the original file, as well as how long it takes to compress the data. We also evaluate the parallel performance of LibPressio. These are assessed in order to determine if the abstraction improves efficiency.

ACM-SRC Semi-Finalist: no

Poster: PDF
Poster Summary: PDF


Back to Poster Archive Listing