Title: Modeling Cache Sharing for MPI Programs on Multicore Machines

Authors: Bin Bao, Chen Ding
Speaker: Bin Bao

Abstract:

The emergence of chip multiprocessors has enriched the computation
resource available for large-scale scientific computing where Message
Passing Interface (MPI) is the dominant programming model. Meanwhile,
multi-core architecture raises new challenges to MPI
applications. Memory system contention is a major problem and has been
studied extensively. Recent advances in cache sharing models provide
an opportunity to dissect the cause of the memory system contention,
in particular, to separate the effect of MPI program parallelization
from the effect of MPI task interaction.

In this talk, we present two scaling models: one for MPI task
partitioning, and the other for MPI cache sharing. We show that
programs have different patterns in task partitioning but the patterns
are consistent for the same program. The pattern can be exposed
through training analysis, task normalization, and linear regression
analysis. We show the effectiveness of our models using several MPI
benchmarks.