ABSTRACT

This chapter presents an overview of the Argonne Leadership Computing Facility (ALCF), including details on recent and future upgrades. It describes the ALCF’s newest supercomputer that officially entered production mode in July 2017. The chapter provides the ALCF’s IBM Blue Gene/Q Mira and some of the major scientific accomplishments it has enabled over the past four years. It also presents an overview of the ALCF job failure analysis system and the Cobalt resource scheduler – two important system software features unique to the facility. Theta is part of the Cray line of supercomputers and continues the ALCF’s architectural direction of highly scalable, homogeneous many-core systems. Cray XC systems like Theta run the Cray Linux Environment. Cobalt utilizes a component daemon architecture that allows rapid deployment to new supercomputing architectures without compromising support for existing architectures Mira, the Cooley cluster, and Theta XC40.