First Technology Transfer

Standard and Advanced Technical Training, Consultancy and Mentoring

OpenCL Programming on the RaspberryPi and the RaspberryPi Cluster

Duration: 5 Days

Much work on machine learning using neural networks has made use of CUDA based code running on NVidia GPUs (Graphics Process Units). OpenCL views a computing system as consisting of a number of compute devices, which might be central processing units (CPUs) or "accelerators" such as graphics processing units (GPUs), attached to a host processor (a CPU), and as such is more general than CUDA.

This workshop, which is not for the faint hearted, but is not diabolically difficult either, will explore the uses of OpenCL on both the GPU processor found on a RaspberrPi 3 board and also explore running OpenCL over a RaspberryPi cluster. The latter will make use of the SnuCL /SnuCL-D suite of tools developed at the centre for manycore programming research centre of Seoul National University.

In addition to providing an introduction to OpenCL programming this workshop will introduce the approaches involved in the use of OpenCL for machine learning using Neural Networks.