1st Workshop on Reconfigurable Computing for HPC and HPDA - ReC4P
London, UK, August 31, 2015

Co-located with the 25th International Conference on Field-programmable Logic and Applications (FPL -

Accelerators implemented on reconfigurable logic appear a promising platform to increase efficiency of high performance computing (HPC) systems. Through specialization, hybrid systems composed of general-purpose processors and reconfigurable devices, such as FPGAs, configured to accelerate specific kernel of an applications, can achieve speed-ups of orders of magnitude, with limited power consumption. Scientific simulations, which employ regular computation, have high arithmetic intensity, and present easily partitionable datasets, can easily benefit from application specific accelerators.

On the other hand, a new class of emerging applications is appearing. This class includes the big field of data analytics applications (e.g. analysis of social networks, security, transportation and communications networks, finance databases, bio-informatics and healthcare, open government, the semantic web, data mining, language understanding, pattern recognition). All these applications have different characteristics and behaviors from conventional HPC applications (typically, scientific simulations). However, they still require High Performance Computing systems to provide results in a timely fashion. Furthermore, they are all experiencing an exponential growth in the availability of data (i.e., big data challenges). For this reason, they have started to be identified as High Performance Data Analytics applications (HPDA).

There still are significant gaps between the complexity and variety of HPC and HPDA applications, and the potentials benefits provided by custom accelerators. 

Objective of this workshop is to identify the breakthroughs required to make reconfigurable logic applicable to HPC and HPDA applications. Intended audience of the workshop includes researchers from industry, academia and government for the area of reconfigurable computing.

Topics of the ReC4P Workshop include, but are not limited to Architecture, Applications and Tools in the area of Reconfigurable Computing for High Performance Computing and High Performance Data Analytics.

ReC4P organizers are looking for talks that can rise interest in the audience. Submissions should be a SINGLE PAGE abstract including title and authors. We would ask prospective presenters to email the abstract of their presentation to the organizers.

If you have any further questions, please feel free to contact us.