Implementation and performance analysis of non-blocking collective operations for MPI. Hoefler, T., Lumsdaine, A., & Rehm, W. In Proceedings of the 2007 ACM/IEEE Conference on Supercomputing, SC'07, 2007. ![Implementation and performance analysis of non-blocking collective operations for MPI [link]](https://bibbase.org/img/filetypes/link.svg "link")
Website doi abstract bibtex Collective operations and non-blocking point-to-point operations have always been part of MPI. Although non-blocking collective operations are an obvious extension to MPI. there have been no comprehensive studies of this functionality. In this paper we present LibNBC, a portable high-performance library for implementing non-blocking collective MPI communication operations. LibNBC provides non-blocking versions of all MPI collective operations, is layered on top of MPI-1, and is portable to nearly all parallel architectures. To measure the performance characteristics of our implementation, we also present a microbenchmark for measuring both latency and overlap of computation and communication. Experimental results demonstrate that the blocking performance of the collective operations in our library is comparable to that of collective operations in other high-performance MPI implementations. Our library introduces a very low overhead between the application and the underlying MPI and thus, in conjunction with the potential to overlap communication with computation, offers the potential for optimizing real-world applications. (c) 2007 ACM.
@inproceedings{
title = {Implementation and performance analysis of non-blocking collective operations for MPI},
type = {inproceedings},
year = {2007},
keywords = {Collective operations; MPI; Non-blocking collecti,Communication,Computers},
websites = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-56749151145&doi=10.1145%2F1362622.1362692&partnerID=40&md5=5003373beb10b1634ff4b9f06c6f0053},
city = {Reno, NV},
id = {8eecf4da-7166-388b-b01d-6629ac25c05c},
created = {2018-01-09T20:30:38.425Z},
file_attached = {false},
profile_id = {42d295c0-0737-38d6-8b43-508cab6ea85d},
last_modified = {2018-03-12T19:03:18.259Z},
read = {false},
starred = {false},
authored = {true},
confirmed = {true},
hidden = {false},
citation_key = {Hoefler2007},
source_type = {conference},
notes = {cited By 96; Conference of 2007 ACM/IEEE Conference on Supercomputing, SC'07 ; Conference Date: 10 November 2007 Through 16 November 2007; Conference Code:74326},
folder_uuids = {2aba6c14-9027-4f47-8627-0902e1e2342b},
private_publication = {false},
abstract = {Collective operations and non-blocking point-to-point operations have always been part of MPI. Although non-blocking collective operations are an obvious extension to MPI. there have been no comprehensive studies of this functionality. In this paper we present LibNBC, a portable high-performance library for implementing non-blocking collective MPI communication operations. LibNBC provides non-blocking versions of all MPI collective operations, is layered on top of MPI-1, and is portable to nearly all parallel architectures. To measure the performance characteristics of our implementation, we also present a microbenchmark for measuring both latency and overlap of computation and communication. Experimental results demonstrate that the blocking performance of the collective operations in our library is comparable to that of collective operations in other high-performance MPI implementations. Our library introduces a very low overhead between the application and the underlying MPI and thus, in conjunction with the potential to overlap communication with computation, offers the potential for optimizing real-world applications. (c) 2007 ACM.},
bibtype = {inproceedings},
author = {Hoefler, T and Lumsdaine, A and Rehm, W},
doi = {10.1145/1362622.1362692},
booktitle = {Proceedings of the 2007 ACM/IEEE Conference on Supercomputing, SC'07}
}
Downloads: 0
{"_id":"KsBrTPTXxTRee5B3A","bibbaseid":"hoefler-lumsdaine-rehm-implementationandperformanceanalysisofnonblockingcollectiveoperationsformpi-2007","downloads":0,"creationDate":"2018-03-12T19:10:28.100Z","title":"Implementation and performance analysis of non-blocking collective operations for MPI","author_short":["Hoefler, T.","Lumsdaine, A.","Rehm, W."],"year":2007,"bibtype":"inproceedings","biburl":"https://bibbase.org/service/mendeley/42d295c0-0737-38d6-8b43-508cab6ea85d","bibdata":{"title":"Implementation and performance analysis of non-blocking collective operations for MPI","type":"inproceedings","year":"2007","keywords":"Collective operations; MPI; Non-blocking collecti,Communication,Computers","websites":"https://www.scopus.com/inward/record.uri?eid=2-s2.0-56749151145&doi=10.1145%2F1362622.1362692&partnerID=40&md5=5003373beb10b1634ff4b9f06c6f0053","city":"Reno, NV","id":"8eecf4da-7166-388b-b01d-6629ac25c05c","created":"2018-01-09T20:30:38.425Z","file_attached":false,"profile_id":"42d295c0-0737-38d6-8b43-508cab6ea85d","last_modified":"2018-03-12T19:03:18.259Z","read":false,"starred":false,"authored":"true","confirmed":"true","hidden":false,"citation_key":"Hoefler2007","source_type":"conference","notes":"cited By 96; Conference of 2007 ACM/IEEE Conference on Supercomputing, SC'07 ; Conference Date: 10 November 2007 Through 16 November 2007; Conference Code:74326","folder_uuids":"2aba6c14-9027-4f47-8627-0902e1e2342b","private_publication":false,"abstract":"Collective operations and non-blocking point-to-point operations have always been part of MPI. Although non-blocking collective operations are an obvious extension to MPI. there have been no comprehensive studies of this functionality. In this paper we present LibNBC, a portable high-performance library for implementing non-blocking collective MPI communication operations. LibNBC provides non-blocking versions of all MPI collective operations, is layered on top of MPI-1, and is portable to nearly all parallel architectures. To measure the performance characteristics of our implementation, we also present a microbenchmark for measuring both latency and overlap of computation and communication. Experimental results demonstrate that the blocking performance of the collective operations in our library is comparable to that of collective operations in other high-performance MPI implementations. Our library introduces a very low overhead between the application and the underlying MPI and thus, in conjunction with the potential to overlap communication with computation, offers the potential for optimizing real-world applications. (c) 2007 ACM.","bibtype":"inproceedings","author":"Hoefler, T and Lumsdaine, A and Rehm, W","doi":"10.1145/1362622.1362692","booktitle":"Proceedings of the 2007 ACM/IEEE Conference on Supercomputing, SC'07","bibtex":"@inproceedings{\n title = {Implementation and performance analysis of non-blocking collective operations for MPI},\n type = {inproceedings},\n year = {2007},\n keywords = {Collective operations; MPI; Non-blocking collecti,Communication,Computers},\n websites = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-56749151145&doi=10.1145%2F1362622.1362692&partnerID=40&md5=5003373beb10b1634ff4b9f06c6f0053},\n city = {Reno, NV},\n id = {8eecf4da-7166-388b-b01d-6629ac25c05c},\n created = {2018-01-09T20:30:38.425Z},\n file_attached = {false},\n profile_id = {42d295c0-0737-38d6-8b43-508cab6ea85d},\n last_modified = {2018-03-12T19:03:18.259Z},\n read = {false},\n starred = {false},\n authored = {true},\n confirmed = {true},\n hidden = {false},\n citation_key = {Hoefler2007},\n source_type = {conference},\n notes = {cited By 96; Conference of 2007 ACM/IEEE Conference on Supercomputing, SC'07 ; Conference Date: 10 November 2007 Through 16 November 2007; Conference Code:74326},\n folder_uuids = {2aba6c14-9027-4f47-8627-0902e1e2342b},\n private_publication = {false},\n abstract = {Collective operations and non-blocking point-to-point operations have always been part of MPI. Although non-blocking collective operations are an obvious extension to MPI. there have been no comprehensive studies of this functionality. In this paper we present LibNBC, a portable high-performance library for implementing non-blocking collective MPI communication operations. LibNBC provides non-blocking versions of all MPI collective operations, is layered on top of MPI-1, and is portable to nearly all parallel architectures. To measure the performance characteristics of our implementation, we also present a microbenchmark for measuring both latency and overlap of computation and communication. Experimental results demonstrate that the blocking performance of the collective operations in our library is comparable to that of collective operations in other high-performance MPI implementations. Our library introduces a very low overhead between the application and the underlying MPI and thus, in conjunction with the potential to overlap communication with computation, offers the potential for optimizing real-world applications. (c) 2007 ACM.},\n bibtype = {inproceedings},\n author = {Hoefler, T and Lumsdaine, A and Rehm, W},\n doi = {10.1145/1362622.1362692},\n booktitle = {Proceedings of the 2007 ACM/IEEE Conference on Supercomputing, SC'07}\n}","author_short":["Hoefler, T.","Lumsdaine, A.","Rehm, W."],"urls":{"Website":"https://www.scopus.com/inward/record.uri?eid=2-s2.0-56749151145&doi=10.1145%2F1362622.1362692&partnerID=40&md5=5003373beb10b1634ff4b9f06c6f0053"},"biburl":"https://bibbase.org/service/mendeley/42d295c0-0737-38d6-8b43-508cab6ea85d","bibbaseid":"hoefler-lumsdaine-rehm-implementationandperformanceanalysisofnonblockingcollectiveoperationsformpi-2007","role":"author","keyword":["Collective operations; MPI; Non-blocking collecti","Communication","Computers"],"metadata":{"authorlinks":{}},"downloads":0},"search_terms":["implementation","performance","analysis","non","blocking","collective","operations","mpi","hoefler","lumsdaine","rehm"],"keywords":["collective operations; mpi; non-blocking collecti","communication","computers"],"authorIDs":[],"dataSources":["zgahneP4uAjKbudrQ","ya2CyA73rpZseyrZ8"]}