This paper deals with the weighted minimum-variance self-tuning regulation of stochastic time-varying systems, which can be described by linear input-output mathematical models. We consider the input-output ARMAX mathematical models with unknown time-varying parameters. The recursive extended least squares RELS algorithm, which can be applied to the stochastic time-varying systems, is presented...

This quarter’s newsletter features the performance area, which has a broad focus including autotuning, performance modeling, end-to-end performance measurement, and tool integration. We aim to extend and integrate mature, robust performance measurement technologies and develop a comprehensive performance data management framework that can be used by other areas of the SUPER project. End-to-end ...

Three interacting tools to assist distributed memory programmers in developing, optimising and understanding application performance have been developed. These tools perform automatic code generation from an initial workload speciication, performance prediction using memory hierarchy simulation, and performance visualisation for distributed memory message passing applications. Their combination...

It is well known that the transient behaviors of the traditional adaptive control may be very poor in general, and that the adaptive control designed based on switching between multiple models is an intuitively appealing and practically feasible approach to improve the transient performances. This paper proves that for a typical class of linear systems disturbed by white noises, the multiple mo...

I/O performance is an increasingly important factor in the productivity of large-scale HPC systems such as Hopper, a 153,216 core Cray XE6 system operated by the National Energy Research Scientific Computing Center. The scientific workload diversity of such systems presents a challenge for I/O performance tuning, however. Applications vary in terms of data volume, I/O strategy, and access metho...

Assessment of performance in tracking set-point changes for single loop systems is presented. In contrast to the works that used stochastic performances, the current assessment uses a deterministic performance measure, i.e. the integration of absolute tracking errors (abbr. IAE). A benchmark system that has an open loop transfer function (abbr. OLTF) comprised of one LHP zero, one integrator, a...

Small organisations can now have access to high raw processing power using networks of workstations (NOW) as parallel computing platforms. Distributed Virtual Shared Memory (DVSM) packages have been developed to facilitate the programming of such systems. However, because of the high interprocess latencies in a NOW, the performance of a DVSM application is more susceptible to the partitioning o...

Performance Extrapolation is the process of evaluating the performance of a parallel program in a target execution environment using performance information obtained for the same program in a different execution environment. Performance extrapolation techniques are suited for rapid performance tuning of parallel programs, particularly when the target environment is unavailable. This paper descr...