Dec. 18, 1997 to Dec. 20, 1997
Martha Torres , Escola Politecnica da Universidade de Sao Paulo
Sergio Takeo Kofuji , Escola Politecnica da Universidade de Sao Paulo
The aim of this work is to measure the barrier synchronization influence on the overall performance of several application programs. In order to do that, we use the MPICH implementation version 1.1 of the MPI library. Moreover, we choose two barrier synchronization solutions: the MPICH software solution, i.e. the MPI_Barrier routine, and the TTL_PAPERS hardware solution. The last one was implemented as a collective operation in MPICH we called MPI_Papers. Finally, we compare the execution time of MPI programs using the MPI_Barrier and MPI_Papers routines. Our testing environment consists of a cluster of 8 Dual Pentium Pro processors interconnected via both 10Mb/s ethernet network and a TTL_PAPERS prototype. We use two kinds of MPI programs: test programs that perform few messages into the network and local calculations, and scientific calculation programs that carry out date exchange and local calculations.
Barrier synchronization, performance evaluation, MPI, NOWs, parallel programming
Martha Torres, Sergio Takeo Kofuji, "The Barrier Synchronization Impact on the MPI-Programs Performance Using a Cluster of Workstations", ISPAN, 1997, Parallel Architectures, Algorithms, and Networks, International Symposium on, Parallel Architectures, Algorithms, and Networks, International Symposium on 1997, pp. 429, doi:10.1109/ISPAN.1997.645132