TY - GEN
T1 - Performance expectations and guidelines for MPI derived datatypes
AU - Gropp, William
AU - Hoefler, Torsten
AU - Thakur, Rajeev
AU - Träff, Jesper Larsson
N1 - Funding Information:
Acknowledgments. This work was supported in part by the Office of Advanced Scientific Computing Research, Office of Science, U.S. Department of Energy, under contract DE-AC02-06CH11357 and DE-FG02-08ER25835, and by the Blue Waters sustained-petascale computing project, which is supported by the National Science Foundation (award number OCI 07-25070) and the state of Illinois.
PY - 2011
Y1 - 2011
N2 - MPI's derived datatypes provide a powerful mechanism for concisely describing arbitrary, noncontiguous layouts of user data for use in MPI communication. This paper formulates self-consistent performance guidelines for derived datatypes. Such guidelines make performance expectations for derived datatypes explicit and suggest relevant optimizations to MPI implementers. We also identify self-consistent guidelines that are too strict to enforce, because they entail NP-hard optimization problems. Enforced self-consistent guidelines assure the user that certain manual datatype optimizations cannot lead to performance improvements, which in turn contributes to performance portability between MPI implementations that behave in accordance with the guidelines. We present results of tests with several MPI implementations, which indicate that many of them violate the guidelines.
AB - MPI's derived datatypes provide a powerful mechanism for concisely describing arbitrary, noncontiguous layouts of user data for use in MPI communication. This paper formulates self-consistent performance guidelines for derived datatypes. Such guidelines make performance expectations for derived datatypes explicit and suggest relevant optimizations to MPI implementers. We also identify self-consistent guidelines that are too strict to enforce, because they entail NP-hard optimization problems. Enforced self-consistent guidelines assure the user that certain manual datatype optimizations cannot lead to performance improvements, which in turn contributes to performance portability between MPI implementations that behave in accordance with the guidelines. We present results of tests with several MPI implementations, which indicate that many of them violate the guidelines.
UR - http://www.scopus.com/inward/record.url?scp=80052984639&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=80052984639&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-24449-0_18
DO - 10.1007/978-3-642-24449-0_18
M3 - Conference contribution
AN - SCOPUS:80052984639
SN - 9783642244483
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 150
EP - 159
BT - Recent Advances in the Message Passing Interface - 18th European MPI Users' Group Meeting, EuroMPI 2011, Proceedings
T2 - 18th European Message Passing Interface Users' Group Meeting, EuroMPI 2011
Y2 - 18 September 2011 through 21 September 2011
ER -