Managing adaptivity in parallel systems

The management of non-functional features (performance, security, power management, etc.) is traditionally a difficult, error prone task for programmers of parallel applications. To take care of these non-functional features, autonomic managers running policies represented as rules using sensors and actuators to monitor and transform a running parallel application may be used. We discuss an approach aimed at providing formal tool support to the integration of independently developed autonomic managers taking care of different non-functional concerns within the same parallel application. Our approach builds on the Behavioural Skeleton experience (autonomic management of non-functional features in structured parallel applications) and on previous results on conflict detection and resolution in rule-based systems. © 2013 Springer-Verlag Berlin Heidelberg.

Targeting heterogeneous architectures via macro data flow

We propose a data flow based run time system as an efficient tool for supporting execution of parallel code on heterogeneous architectures hosting both multicore CPUs and GPUs. We discuss how the proposed run time system may be the target of both structured parallel applications developed using algorithmic skeletons/parallel design patterns and also more "domain specific" programming models. Experimental results demonstrating the feasibility of the approach are presented. © 2012 World Scientific Publishing Company.

Strategies for Energy-Efficient Resource Management of Hybrid Programming Models

Many scientific applications are programmed using hybrid programming models that use both message passing and shared memory, due to the increasing prevalence of large-scale systems with multicore, multisocket nodes. Previous work has shown that energy efficiency can be improved using software-controlled execution schemes that consider both the programming model and the power-aware execution capabilities of the system. However, such approaches have focused on identifying optimal resource utilization for one programming model, either shared memory or message passing, in isolation. The potential solution space, thus the challenge, increases substantially when optimizing hybrid models since the possible resource configurations increase exponentially. Nonetheless, with the accelerating adoption of hybrid programming models, we increasingly need improved energy efficiency in hybrid parallel applications on large-scale systems. In this work, we present new software-controlled execution schemes that consider the effects of dynamic concurrency throttling (DCT) and dynamic voltage and frequency scaling (DVFS) in the context of hybrid programming models. Specifically, we present predictive models and novel algorithms based on statistical analysis that anticipate application power and time requirements under different concurrency and frequency configurations. We apply our models and methods to the NPB MZ benchmarks and selected applications from the ASC Sequoia codes. Overall, we achieve substantial energy savings (8.74 percent on average and up to 13.8 percent) with some performance gain (up to 7.5 percent) or negligible performance loss.

Behavioural skeletons in GCM: autonomic management of grid components

Autonomic management can be used to improve the QoS provided by parallel/distributed applications. We discuss behavioural skeletons introduced in earlier work: rather than relying on programmer ability to design “from scratch” efficient autonomic policies, we encapsulate general autonomic controller features into algorithmic skeletons. Then we leave to the programmer the duty of specifying the parameters needed to specialise the skeletons to the needs of the particular application at hand. This results in the programmer having the ability to fast prototype and tune distributed/parallel applications with non-trivial autonomic management capabilities. We discuss how behavioural skeletons have been implemented in the framework of GCM(the Grid ComponentModel developed within the CoreGRID NoE and currently being implemented within the GridCOMP STREP project). We present results evaluating the overhead introduced by autonomic management activities as well as the overall behaviour of the skeletons. We also present results achieved with a long running application subject to autonomic management and dynamically adapting to changing features of the target architecture.
Overall the results demonstrate both the feasibility of implementing autonomic control via behavioural skeletons and the effectiveness of our sample behavioural skeletons in managing the “functional replication” pattern(s).

An efficient unbounded lock-free queue for multi-core systems

The use of efficient synchronization mechanisms is crucial for implementing fine grained parallel programs on modern shared cache multi-core architectures. In this paper we study this problem by considering Single-Producer/Single- Consumer (SPSC) coordination using unbounded queues. A novel unbounded SPSC algorithm capable of reducing the row synchronization latency and speeding up Producer-Consumer coordination is presented. The algorithm has been extensively tested on a shared-cache multi-core platform and a sketch proof of correctness is presented. The queues proposed have been used as basic building blocks to implement the FastFlow parallel framework, which has been demonstrated to offer very good performance for fine-grain parallel applications. © 2012 Springer-Verlag.

Scalable black-box prediction models for multi-dimensional adaptation on NUMA multi-cores

This paper presents a scalable, statistical ‘black-box’ model for predicting the performance of parallel programs on multi-core non-uniform memory access (NUMA) systems. We derive a model with low overhead, by reducing data collection and model training time. The model can accurately predict the behaviour of parallel applications in response to changes in their concurrency, thread layout on NUMA nodes, and core voltage and frequency. We present a framework that applies the model to achieve significant energy and energy-delay-square (ED2) savings (9% and 25%, respectively) along with performance improvement (10% mean) on an actual 16-core NUMA system running realistic application workloads. Our prediction model proves substantially more accurate than previous efforts.

Autonomic management of multiple non-functional concerns in behavioural skeletons

We introduce and address the problem of concurrent autonomic management of different non-functional concerns in parallel applications build as a hierarchical composition of behavioural skeletons. We first define the problems arising when multiple concerns are dealt with by independent managers, then we propose a methodology supporting coordinated management, and finally we discuss how autonomic management of multiple concerns may be implemented in a typical use case. Being based on the behavioural skeleton concept proposed in the CoreGRID GCM, it is anticipated that the methodology will be readily integrated into the current reference implementation of GCM based on Java Pro Active and running on top of major grid middleware systems.

Using Computational Significance and Resilience in System Software Stacks.:Keynote Talk

This talk explores how the runtime system and operating system can leverage metrics that express the significance and resilience of application components in order to reduce the energy footprint of parallel applications. We will explore in particular how software can tolerate and indeed exploit higher error rates in future processors and memory technologies that may operate outside their safe margins.

Series-L/parallel-tuned comparison with shunt-C/series-tuned class-E power amplifier

An analysis of the operation of a series-L/parallel-tuned class-E amplifier and its equivalence to the classic shunt-C/series-tuned class-E amplifier are presented. The first reported closed form design equations for the series-L/parallel-tuned topology operating under ideal switching conditions are given. Furthermore, a design procedure is introduced that allows the effect that nonzero switch resistance has on amplifier performance efficiency to be accounted for. The technique developed allows optimal circuit components to be found for a given device series resistance. For a relatively high value of switching device ON series resistance of 4O, drain efficiency of around 66% for the series-L/parallel-tuned topology, and 73% for the shunt-C/series-tuned topology appear to be the theoretical limits. At lower switching device series resistance levels, the efficiency performance of each type are similar, but the series-L/parallel-tuned topology offers some advantages in terms of its potential for MMIC realisation. Theoretical analysis is confirmed by numerical simulation for a 500mW (27dBm), 10% bandwidth, 5 V series-L/parallel-tuned, then, shunt-C/series-tuned class E power amplifier, operating at 2.5 GHz, and excellent agreement between theory and simulation results is achieved. The theoretical work presented in the paper should facilitate the design of high-efficiency switched amplifiers at frequencies commensurate with the needs of modern mobile wireless applications in the microwave frequency range, where intrinsically low-output-capacitance MMIC switching devices such as pHEMTs are to be used.

Semi-formal Models to Support Program Development: Autonomic Management within Component Based Parallel and Distributed Programming

Functional and non-functional concerns require different programming effort, different techniques and different methodologies when attempting to program efficient parallel/distributed applications. In this work we present a "programmer oriented" methodology based on formal tools that permits reasoning about parallel/distributed program development and refinement. The proposed methodology is semi-formal in that it does not require the exploitation of highly formal tools and techniques, while providing a palatable and effective support to programmers developing parallel/distributed applications, in particular when handling non-functional concerns.