Towards modeling and integrated design automation of supercomputing clusters (MIDAS)

Abstract

The choice of supercomputers, by the user community, should not merely be based on the benchmarks. With the advent of supercomputers delivering petaops performance, computationally intensive applications such as brain modeling and energy requirement prediction have become predominant. Conventional benchmarks do not reflect the functional characteristics of these applications. Thus, the gap between the performance projected and the actual performance delivered when the application is ported onto the cluster is wide. The primary cause for this disparity in performance is due to the user community's lack of knowledge about the system and improper design choices. With the supercomputer market being seller biased this problem is worsened. Design and development of very large systems like the clusters involve massive efforts and financial investment. Creating a design automation environment for the clusters is bound to bring down the man years and hence the production cost. Currently there are adhoc approaches which lack an integrated design space to cater to the wider needs of the user community. Thus, an integrated design framework that takes all the components of the supercomputing cluster and the relationship across the design spaces is essential. In this paper we propose a novel methodology called the Modeling and Integrated Design Automation of Supercomputing Cluster (MIDAS). MIDAS is the birth of a new design philosophy which requires lot of research and tuning, providing scope for evolution and optimization. Keywords Design automation · Supercomputer model Need for an integrated design automation approachSupercomputing clusters are not mass produced till date unlike general purpose and application specific programmable processors. With the fast and far widening areas of applications, the need will arise for not just powerful but more clusters, in the field of life sciences, energy, environmental, space and particle sciences well supported by technology. Then the question is how to bring down the production cost of such clusters and possibly make it a buyer's market. Design and development of very large systems like the clusters involve massive efforts and financial investment. Creating a design automation environment for the clusters is bound to bring down the man years and hence the production cost. Currently there are ad hoc approaches which lack an integrated design space to cater the wider needs of the user community.The goal of this design automation process is not only to speed up the design and development process but also build clusters that genuinely cater to the requirements of the user and the application, hence providing long term costeffectiveness. To create an integrated design automation for developing the clusters one needs to build a highly com-1 3

SADDLE: A Modular Design Automation Framework for Cluster Supercomputers and Data Centres

“…Nagarajan Venkateswaran et al [5] addressed the problem of automated design again in 2009. Their methodology, "Modeling and integrated design automation of supercomputers (MIDAS)", is aimed at Supercomputers-on-a-Chip (SCOC), but can be generalised to wider areas as well.…”