Automatic parameter assessment of logp-based communication models in MPI environments

Abstract

Communications in modern interconnection networks become a complex issue to deliver the highest performance to the user. In MPI environments, this complexity leads to different communication protocols for different message sizes. The characterization of these different behaviors is important and useful for software developers and network designers. This paper presents an automatic method to obtain a characterization of the communication behavior of a specific MPI environment using LogP-based models. This procedure automatically detects the message sizes where the communication behavior changes due to the influence of different architectural features and protocols. Thereby, the range of message sizes is split in different intervals, in a transparent way to the user. Then, using LogP-based models, each interval is characterized by its own set of parameters. The procedure is based on the parameterized roundtrip-time low-overhead microbenchmark. The detection of different communication behaviors and the assessment of the parameters of the LogP-based models were implemented in an easy to use statistical environment. Real MPI environments were characterized using our proposal.

Modeling the performance of parallel applications using model selection techniques

“…Therefore, it is possible to directly compare these theoretical models with the AIC-based models obtained by TIA. The value of the LogGP parameters of the cluster was assessed by the NGMPI [16] module of TIA. This module implements an extension of the low-overhead measurement method described in [25], which produces an accurate assessment of all LogGP parameters without network flooding or unnecessary contention.…”

“…In particular, the main objective of this framework is to obtain analytical expressions of the performance of measurable metrics as a function of the execution parameters. Anyway, the framework structure is flexible enough to carry out other types of analysis that partially use the TIA capabilities, such as the parameter assessment of LogP‐based communication models .…”