Exposing I/O concurrency with informed prefetching

Author

Patterson, R. Hugo ; Gibson, Garth A.

Author_Institution

Dept. of Electr. & Comput. Eng., Carnegie Mellon Univ., Pittsburgh, PA, USA

fYear

1994

fDate

28-30 Sep 1994

Firstpage

7

Lastpage

16

Abstract

Informed prefetching provides a simple mechanism for I/O-intensive, cache-ineffective applications to efficiently exploit highly-parallel I/O subsystems such as disk arrays. This mechanism, dynamic disclosure of future accesses, yields substantial benefits over sequential readahead mechanisms found in current file systems for non-sequential workloads. This paper reports the performance of the Transparent Informed Prefetching system (TIP), a minimal prototype implemented in a Mach 3.0 system with up to four disks. We measured reductions by factors of up to 1.9 and 3.7 in the execution time of two example applications: multi-file text search and scientific data visualization

Keywords

buffer storage; concurrency control; data visualisation; input-output programs; operating systems (computers); parallel processing; I/O concurrency; I/O-intensive cache-ineffective applications; Mach 3.0 system; Transparent Informed Prefetching system; disk arrays; dynamic disclosure; execution time; file systems; highly-parallel I/O subsystems; informed prefetching; input output concurrency; multi-file text search; nonsequential workloads; scientific data visualization; sequential readahead mechanisms; Application software; Computer science; Concurrent computing; Delay; File systems; Prefetching; Prototypes; Testing; Throughput; Time measurement;

fLanguage

English

Publisher

ieee

Conference_Titel

Parallel and Distributed Information Systems, 1994., Proceedings of the Third International Conference on

Conference_Location

Austin, TX

Print_ISBN

0-8186-6400-2

Type

conf

DOI

10.1109/PDIS.1994.331737

Filename

331737