Disk-directed I/O for MIMD Multiprocessors. Kotz, D. Disk-directed I/O for MIMD Multiprocessors, pages 513-535. IEEE Computer Society Press and John Wiley & Sons, 2001. abstract bibtex Many scientific applications that run on today's multiprocessors, such as weather forecasting and seismic analysis, are bottlenecked by their file-I/O needs. Even if the multiprocessor is configured with sufficient I/O hardware, the file-system software often fails to provide the available bandwidth to the application. Although libraries and enhanced file-system interfaces can make a significant improvement, we believe that fundamental changes are needed in the file-server software. We propose a new technique, disk-directed I/O, to allow the disk servers to determine the flow of data for maximum performance. Our simulations show that tremendous performance gains are possible both for simple reads and writes and for an out-of-core application. Indeed, our disk-directed I/O technique provided consistent high performance that was largely independent of data distribution, obtained up to 93% of peak disk bandwidth, and was as much as 18 times faster than the traditional technique.
@inBook{
title = {Disk-directed I/O for MIMD Multiprocessors},
type = {inBook},
year = {2001},
keywords = {dartmouth-cs,file-system,parallel-computing,parallel-io},
pages = {513-535},
publisher = {IEEE Computer Society Press and John Wiley & Sons},
chapter = {35},
editors = {[object Object],[object Object],[object Object]},
id = {6692dbeb-1c46-3e33-90ed-dfd983cd8406},
created = {2018-07-12T21:32:30.856Z},
file_attached = {false},
profile_id = {f954d000-ce94-3da6-bd26-b983145a920f},
group_id = {b0b145a3-980e-3ad7-a16f-c93918c606ed},
last_modified = {2018-07-12T21:32:30.856Z},
read = {false},
starred = {false},
authored = {false},
confirmed = {true},
hidden = {false},
citation_key = {kotz:bdiskdir},
source_type = {incollection},
notes = {In jin:io-book, reprinted from kotz:jdiskdir.},
private_publication = {false},
abstract = {Many scientific applications that run on today's multiprocessors, such as weather forecasting and seismic analysis, are bottlenecked by their file-I/O needs. Even if the multiprocessor is configured with sufficient I/O hardware, the file-system software often fails to provide the available bandwidth to the application. Although libraries and enhanced file-system interfaces can make a significant improvement, we believe that fundamental changes are needed in the file-server software. We propose a new technique, disk-directed I/O, to allow the disk servers to determine the flow of data for maximum performance. Our simulations show that tremendous performance gains are possible both for simple reads and writes and for an out-of-core application. Indeed, our disk-directed I/O technique provided consistent high performance that was largely independent of data distribution, obtained up to 93% of peak disk bandwidth, and was as much as 18 times faster than the traditional technique.},
bibtype = {inBook},
author = {Kotz, David},
book = {High Performance Mass Storage and Parallel I/O: Technologies and Applications}
}
Downloads: 0
{"_id":"rScDePLhgFdyQruAn","bibbaseid":"kotz-diskdirectedioformimdmultiprocessors-2001","downloads":0,"creationDate":"2019-02-15T15:15:01.781Z","title":"Disk-directed I/O for MIMD Multiprocessors","author_short":["Kotz, D."],"year":2001,"bibtype":"inBook","biburl":null,"bibdata":{"title":"Disk-directed I/O for MIMD Multiprocessors","type":"inBook","year":"2001","keywords":"dartmouth-cs,file-system,parallel-computing,parallel-io","pages":"513-535","publisher":"IEEE Computer Society Press and John Wiley & Sons","chapter":"35","editors":"[object Object],[object Object],[object Object]","id":"6692dbeb-1c46-3e33-90ed-dfd983cd8406","created":"2018-07-12T21:32:30.856Z","file_attached":false,"profile_id":"f954d000-ce94-3da6-bd26-b983145a920f","group_id":"b0b145a3-980e-3ad7-a16f-c93918c606ed","last_modified":"2018-07-12T21:32:30.856Z","read":false,"starred":false,"authored":false,"confirmed":"true","hidden":false,"citation_key":"kotz:bdiskdir","source_type":"incollection","notes":"In jin:io-book, reprinted from kotz:jdiskdir.","private_publication":false,"abstract":"Many scientific applications that run on today's multiprocessors, such as weather forecasting and seismic analysis, are bottlenecked by their file-I/O needs. Even if the multiprocessor is configured with sufficient I/O hardware, the file-system software often fails to provide the available bandwidth to the application. Although libraries and enhanced file-system interfaces can make a significant improvement, we believe that fundamental changes are needed in the file-server software. We propose a new technique, disk-directed I/O, to allow the disk servers to determine the flow of data for maximum performance. Our simulations show that tremendous performance gains are possible both for simple reads and writes and for an out-of-core application. Indeed, our disk-directed I/O technique provided consistent high performance that was largely independent of data distribution, obtained up to 93% of peak disk bandwidth, and was as much as 18 times faster than the traditional technique.","bibtype":"inBook","author":"Kotz, David","book":"High Performance Mass Storage and Parallel I/O: Technologies and Applications","bibtex":"@inBook{\n title = {Disk-directed I/O for MIMD Multiprocessors},\n type = {inBook},\n year = {2001},\n keywords = {dartmouth-cs,file-system,parallel-computing,parallel-io},\n pages = {513-535},\n publisher = {IEEE Computer Society Press and John Wiley & Sons},\n chapter = {35},\n editors = {[object Object],[object Object],[object Object]},\n id = {6692dbeb-1c46-3e33-90ed-dfd983cd8406},\n created = {2018-07-12T21:32:30.856Z},\n file_attached = {false},\n profile_id = {f954d000-ce94-3da6-bd26-b983145a920f},\n group_id = {b0b145a3-980e-3ad7-a16f-c93918c606ed},\n last_modified = {2018-07-12T21:32:30.856Z},\n read = {false},\n starred = {false},\n authored = {false},\n confirmed = {true},\n hidden = {false},\n citation_key = {kotz:bdiskdir},\n source_type = {incollection},\n notes = {In jin:io-book, reprinted from kotz:jdiskdir.},\n private_publication = {false},\n abstract = {Many scientific applications that run on today's multiprocessors, such as weather forecasting and seismic analysis, are bottlenecked by their file-I/O needs. Even if the multiprocessor is configured with sufficient I/O hardware, the file-system software often fails to provide the available bandwidth to the application. Although libraries and enhanced file-system interfaces can make a significant improvement, we believe that fundamental changes are needed in the file-server software. We propose a new technique, disk-directed I/O, to allow the disk servers to determine the flow of data for maximum performance. Our simulations show that tremendous performance gains are possible both for simple reads and writes and for an out-of-core application. Indeed, our disk-directed I/O technique provided consistent high performance that was largely independent of data distribution, obtained up to 93% of peak disk bandwidth, and was as much as 18 times faster than the traditional technique.},\n bibtype = {inBook},\n author = {Kotz, David},\n book = {High Performance Mass Storage and Parallel I/O: Technologies and Applications}\n}","author_short":["Kotz, D."],"bibbaseid":"kotz-diskdirectedioformimdmultiprocessors-2001","role":"author","urls":{},"keyword":["dartmouth-cs","file-system","parallel-computing","parallel-io"],"downloads":0},"search_terms":["disk","directed","mimd","multiprocessors","kotz"],"keywords":["dartmouth-cs","file-system","parallel-computing","parallel-io","parallel computing","file system","parallel computing"],"authorIDs":["5c66d9b9e4b1db10000001ba","5c69efe6ac591010000000e0"]}