Noncontiguous I/O through PVFS

TL;DR

This paper introduces list I/O, a native noncontiguous I/O method implemented in PVFS, which significantly improves performance for scientific applications with noncontiguous data access patterns.

Contribution

The paper presents list I/O, a novel native noncontiguous I/O approach that outperforms existing methods in PVFS, enhancing high-performance scientific computing.

Findings

List I/O outperforms current noncontiguous I/O methods in most situations.

List I/O substantially improves performance of real-world scientific applications.

Experimental results demonstrate significant performance gains with list I/O.

Abstract

With the tremendous advances in processor and memory technology, I/O has risen to become the bottleneck in high-performance computing for many applications. The development of parallel file systems has helped to ease the performance gap, but I/O still remains an area needing significant performance improvement. Research has found that noncontiguous I/O access patterns in scientific applications combined with current file system methods to perform these accesses lead to unacceptable performance for large data sets. To enhance performance of noncontiguous I/O we have created list I/O, a native version of noncontiguous I/O. We have used the Parallel Virtual File System (PVFS) to implement our ideas. Our research and experimentation shows that list I/O outperforms current noncontiguous I/O access methods in most I/O situations and can substantially enhance the performance of real-world scientific applications.