TY - GEN
T1 - MPICH-GQ
T2 - 2000 ACM/IEEE Conference on Supercomputing, SC 2000
AU - Roy, Alain
AU - Foster, Ian
AU - Gropp, William
AU - Karonis, Nicholas
AU - Sander, Volker
AU - Toonen, Brian
N1 - We gratefully acknowledge assistance given by Linda Winkler and Becca Nitzan with the testbed used in these experiments and by Andy Adamson who wrote the UDP traffic generator. John Bresnahan and Joe Link have been instrumental in implementing our QoS extensions to MPI. Numerous discussions with our colleagues Keith Jackson, Gary Hoo, Bill Johnston, Carl Kesselman, and Steven Tuecke have helped shape our approach to quality of service. We also thank Cisco Systems for an equipment donation that allowed the creation of the GARNET testbed. This work was supported in part by the Mathematical, Information, and Computational Sciences Division subprogram of the Office of Advanced Scientific Computing Research, U.S. Department of Energy, under Contract W-31-109-Eng-38; by the Defense Advanced Research Projects Agency under contract N66001-96-C-8523; by the National Science Foundation; and by the NASA Information Power Grid program.
PY - 2000
Y1 - 2000
N2 - Parallel programmers typically assume that all resources required for a program's execution are dedicated to that purpose. However, in local and wide area networks, contention for shared networks, CPUs, and I/O systems can result in significant variations in availability, with consequent adverse effects on overall performance. We describe a new message-passing architecture, MPICH-GQ, that uses quality of service (QoS) mechanisms to manage contention and hence improve performance of message passing interface (MPI) applications. MPICH-GQ combines new QoS specification, traffic shaping, QoS reservation, and QoS implementation techniques to deliver QoS capabilities to the high-bandwidth bursty flows, complex structures, and reliable protocols used in high-performance applications-characteristics very different from the low-bandwidth, constant bit-rate media flows and unreliable protocols for which QoS mechanisms were designed. Results obtained on a differentiated services testbed demonstrate our ability to maintain application performance in the face of heavy network contention.
AB - Parallel programmers typically assume that all resources required for a program's execution are dedicated to that purpose. However, in local and wide area networks, contention for shared networks, CPUs, and I/O systems can result in significant variations in availability, with consequent adverse effects on overall performance. We describe a new message-passing architecture, MPICH-GQ, that uses quality of service (QoS) mechanisms to manage contention and hence improve performance of message passing interface (MPI) applications. MPICH-GQ combines new QoS specification, traffic shaping, QoS reservation, and QoS implementation techniques to deliver QoS capabilities to the high-bandwidth bursty flows, complex structures, and reliable protocols used in high-performance applications-characteristics very different from the low-bandwidth, constant bit-rate media flows and unreliable protocols for which QoS mechanisms were designed. Results obtained on a differentiated services testbed demonstrate our ability to maintain application performance in the face of heavy network contention.
KW - Differentiated services
KW - MPI
KW - Quality of service
KW - TCP
UR - https://www.scopus.com/pages/publications/84941289558
UR - https://www.scopus.com/pages/publications/84941289558#tab=citedBy
U2 - 10.1109/SC.2000.10017
DO - 10.1109/SC.2000.10017
M3 - Conference contribution
AN - SCOPUS:84941289558
T3 - Proceedings of the International Conference on Supercomputing
BT - SC 2000 - Proceedings of the 2000 ACM/IEEE Conference on Supercomputing
PB - Association for Computing Machinery
Y2 - 4 November 2000 through 10 November 2000
ER -