The DEEP Project An alternative approach to heterogeneous cluster‐computing in the many‐core era

N. Eicker, Thomas Lippert, Thomas Moschny, Estela Suarez

Concurrency and Computation Practice and Experience · 2015 · 34 citations · 12 references

DOIFull text

Open access

Concepts

TL;DR

HPC has shifted from homogeneous cluster architectures to heterogeneous systems that employ accelerator or co‑processor devices. The DEEP project seeks to link a standard HPC cluster with a Booster of many‑core processors and to provide a seamless programming environment for applications. Its architecture connects the cluster to a Booster of many‑core processors, and its programming environment merges MPI offloading with an OmpSs task‑based abstraction layer. © 2015 John Wiley & Sons, Ltd.

Abstract

Summary Homogeneous cluster architectures, which used to dominate high‐performance computing (HPC), are challenged today by heterogeneous approaches utilizing accelerator or co‐processor devices. The DEEP (Dynamical Exascale Entry Platform) project is implementing a novel architecture for HPC, in which a standard HPC cluster is directly connected to a so‐called ‘Booster’: a cluster of many‐core processors. By these means heterogeneity is organized differently as in today's standard approach, where accelerators are added to each node of the cluster. In order to adapt application codes to this Cluster‐Booster architecture as seamless as possible, DEEP has developed a complete programming environment. It integrates the offloading functionality given by the Message Passing Interface standard with an abstraction layer based on the task‐based OmpSs programming paradigm. This paper presents the DEEP project with an emphasis on the DEEP programming environment. Copyright © 2015 John Wiley & Sons, Ltd.

References

12