A survey of MPI usage in the US exascale computing project

Abstract

The Exascale Computing Project (ECP) is currently the primary effort in the United States focused on developing "exascale" levels of computing capabilities, including hardware, software and applications. In order to obtain a more thorough understanding of how the software projects under the ECP are using, and planning to use the Message Passing Interface (MPI), and help guide the work of our own project within the ECP, we created a survey. Of the 97 ECP projects active at the time the survey was distributed, we received 77 responses, 56 of which reported that their projects were using MPI. This paper reports the results of that survey for the benefit of the broader community of MPI developers.

Application health monitoring for extreme‐scale resiliency using cooperative fault management

“…Trident uses a probabilistic model to predict silent data corruption effects, in contrast to fault injection–based methods. A recent survey found the majority of exascale computing projects use MPI . Therefore, work to add resilience capabilities to MPI will be useful to help support resilience in next‐generation applications.…”

“…A recent survey found the majority of exascale computing projects use MPI. 18 Therefore, work to add resilience capabilities to MPI [15][16][17] will be useful to help support resilience in next-generation applications.…”

“…However, the more intrusive such strategies become in application code bases, the greater the resistance to their adoption, especially for the large body of existing applications. In our experience, application‐level checkpoint/restart is the most commonly employed fault tolerance technique at present, and a recent survey of exascale application teams suggests that it will continue to be for at least the next generation of large‐scale supercomputers, despite dire predictions for their MTTIs.…”

Hardware topology management in MPI applications through hierarchical communicators

“…Once again, the lowest level is chosen and in this second step, the new communicators correspond to the L1 caches. Practically, a new call to MPI Comm split type with a newcomm [1] communicator as input will produce the following four communicators in node 0:…”

“…This is known as the data locality issue. Indeed, a recent survey on MPI usage for future exascale systems [1] outlines that a significant fraction of applications (18 out of 48) exhibit sensitivity to the hardware topology.…”