Abstract
Collective communication performance is critical in a number of MPI
applications. In this paper we focus on two widely used primitives,
broadcast and reduce, and present experimental results obtained on a cluster
of PC connected by Myrinet.
We compare the performance of the MPICH primitives with our
implementation based on alpha-trees. Our tests show that the
MPICH implementation can be improved.
Anno
2005
Autori IAC
Tipo pubblicazione
Altri Autori
Bernaschi M., Crea S., Iannello G.