TY - GEN
T1 - Collective communication costs analysis over Gigabit Ethernet and InfiniBand
AU - Mamadou, Hyacinthe Nzigou
AU - Nanri, Takeshi
AU - Murakami, Kazuaki
PY - 2006
Y1 - 2006
N2 - Users of parallel machines need to have a good grasp for how different communication patterns and styles affect the performance of message-passing applications. MPI Collective communications involve multiple processors, and their performance prediction is a tricky task to perform. In order to evaluate the performance of collective communications, we attempt to extend LogGP and P-LogP standard point-to-point models. Our objective is to compare these models with the empirical data, and identify the most suitable for performance characterization of collective communications. The models proposed are related with the implemented algorithms in MPICH. The experimental results performed on clusters of 16 and 64 processors connected by Infiniband and Gigabit Ethernet networks respectively, show the same trend. For any collective operation, given a number of processors and a range of message sizes, there is at least one model that predicts the performance precisely.
AB - Users of parallel machines need to have a good grasp for how different communication patterns and styles affect the performance of message-passing applications. MPI Collective communications involve multiple processors, and their performance prediction is a tricky task to perform. In order to evaluate the performance of collective communications, we attempt to extend LogGP and P-LogP standard point-to-point models. Our objective is to compare these models with the empirical data, and identify the most suitable for performance characterization of collective communications. The models proposed are related with the implemented algorithms in MPICH. The experimental results performed on clusters of 16 and 64 processors connected by Infiniband and Gigabit Ethernet networks respectively, show the same trend. For any collective operation, given a number of processors and a range of message sizes, there is at least one model that predicts the performance precisely.
UR - http://www.scopus.com/inward/record.url?scp=77049123230&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77049123230&partnerID=8YFLogxK
U2 - 10.1007/11945918_52
DO - 10.1007/11945918_52
M3 - Conference contribution
AN - SCOPUS:77049123230
SN - 354068039X
SN - 9783540680390
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 547
EP - 559
BT - High Performance Computing - HiPC 2006 - 13th International Conference Proceedings
T2 - 13th International Conference on High Performance Computing, HiPC 2006
Y2 - 18 December 2006 through 21 December 2006
ER -