DocumentCode
3414906
Title
MTIO. A multi-threaded parallel I/O system
Author
More, Sachin ; Choudhary, Alok ; Foster, Ian ; Xu, Ming Q.
Author_Institution
Dept. of Electr. & Comput. Eng., Northwestern Univ., Evanston, IL, USA
fYear
1997
fDate
1-5 Apr 1997
Firstpage
368
Lastpage
373
Abstract
Presents the design and evaluation of MTIO (Multi-Threaded Input/Output), a multi-threaded runtime library for parallel I/O. We extend the multi-threading concept to separate the computation and I/O tasks into two separate threads of control. Multi-threading in our design permits (a) asynchronous I/O even if the underlying file system does not support asynchronous I/O; (b) copy avoidance from the I/O thread to the compute thread by sharing address space; and (c) a capability to perform collective I/O asynchronously without blocking the compute threads. Further, this paper presents techniques for collective I/O which maximize load balance and concurrency while reducing communication overhead in an integrated fashion. Performance results on an IBM SP2 for various data distributions and access patterns are presented. The results show that there is a tradeoff between the amount of concurrency in I/O and the buffer size designated for I/O, and that there is an optimal buffer size beyond which the benefits of larger requests diminish due to large communication overheads
Keywords
concurrency control; input-output programs; parallel programming; resource allocation; shared memory systems; software libraries; I/O thread; IBM SP2; MTIO; access patterns; asynchronous I/O; collective I/O; communication overhead; computation tasks; compute thread; concurrency; copy avoidance; data distributions; file system; large requests; load balance; multi-threaded runtime library; optimal buffer size; parallel I/O system; performance; shared address space; Computer science; Concurrent computing; Control systems; Design engineering; Design optimization; File systems; Mathematics; Physics computing; Runtime library; Yarn;
fLanguage
English
Publisher
ieee
Conference_Titel
Parallel Processing Symposium, 1997. Proceedings., 11th International
Conference_Location
Genva
ISSN
1063-7133
Print_ISBN
0-8186-7793-7
Type
conf
DOI
10.1109/IPPS.1997.580928
Filename
580928
Link To Document