Accelerating Spark-Based Applications with MPI and OpenACC

The amount of data produced in scientific and commercial fields is growing dramatically. Correspondingly, big data technologies, such as Hadoop and Spark, have emerged to tackle the challenges of collecting, processing, and storing such large-scale data. Unfortunately, big data applications usually have performance issues and do not fully exploit a hardware infrastructure. One reason is that applications are developed using high-level programming languages that do not provide low-level system control in terms of performance of highly parallel programming models like message passing interface (MPI). Moreover, big data is considered a barrier of parallel programming models or accelerators (e.g., CUDA and OpenCL). Therefore, the aim of this study is to investigate how the performance of big data applications can be enhanced without sacrificing the power consumption of a hardware infrastructure. A Hybrid Spark MPI OpenACC (HSMO) system is proposed for integrating Spark as a big data programming model, with MPI and OpenACC as parallel programming models. Such integration brings together the advantages of each programming model and provides greater effectiveness. To enhance performance without sacrificing power consumption, the integration approach needs to exploit the hardware infrastructure in an intelligent manner. For achieving this performance enhancement, a mapping technique is proposed that is built based on the application’s virtual topology as well as the physical topology of the undelaying resources. To the best of our knowledge, there is no existing method in big data applications related to utilizing graphics processing units (GPUs), which are now an essential part of high-performance computing (HPC) as a powerful resource for fast computation.

The OpenMP API for High Integrity Systems

OpenMP is traditionally focused on boosting performance in HPC systems. However, other domains are showing an increasing interest in the use of OpenMP by virtue of key aspects introduced in recent versions of the specification: the tasking model, the accelerator model, and other features like the requires and the assumes directives, which allow defining certain contracts. One example is the safety-critical embedded domain, where several efforts have been initiated towards the adoption of OpenMP. However, the OpenMP specification states that "application developers are responsible for correctly using the OpenMP API to produce a conforming program", being not acceptable in high integrity systems, where aspects such as reliability and resiliency have to be ensured at different levels of criticality. In this scope, programming languages like Ada propose a different paradigm by exposing fewer features to the user, and leaving the responsibility of safely exploiting the full underlying architecture to the compiler and the runtime systems, instead. The philosophy behind this kind of model is to move the responsibility of producing correct parallel programs from users to vendors. In this panel, actors from different domains involved in the use of parallel programming models for the development of high-integrity systems share their thoughts about this topic.

AtTune: A Heuristic based Framework for Parallel Applications Autotuning

Several aspects limit the scalability of parallel applications, e.g., off-chip bus saturation and data synchronization. Moreover, the high cost of cooling HPC systems, which can outweigh the cost of developing the system itself, has pushed the parallel application’s execution to another level of requirements, in terms of performance and energy. In this work, we propose AtTune: a heuristic-based framework for tuning the number of processes/threads and CPU frequency to optimize the parallel applications’ execution. AtTune is transparent for the user, independent of the input size, and it optimizes for different parallel programming models. We evaluated our proposed solution considering five well-known kernels implemented in MPI and OpenMP. Experimental results on two real multi-core systems showed that AtTune improves up to 36%, 11%, and 32% the energy efficiency, performance, and Energy-Delay Product, respectively.

Algorithmic Skeletons and Parallel Design Patterns in Mainstream Parallel Programming

This paper discusses the impact of structured parallel programming methodologies in state-of-the-art industrial and research parallel programming frameworks. We first recap the main ideas underpinning structured parallel programming models and then present the concepts of algorithmic skeletons and parallel design patterns. We then discuss how such concepts have permeated the wider parallel programming community. Finally, we give our personal overview—as researchers active for more than two decades in the parallel programming models and frameworks area—of the process that led to the adoption of these concepts in state-of-the-art industrial and research parallel programming frameworks, and the perspectives they open in relation to the exploitation of forthcoming massively-parallel (both general and special-purpose) architectures.