A recent student project, run by BEng Aerospace Engineering students James Harvey (Year 1) and George Johnson (Year 3), has looked to build a server initially consisting of 21 raspberry-pi credit-card-sized computers. The project was supervised by Dr Xiaodong Xu, Dr Matt O’Donnell and Dr Andre Jesus, and supported by the Engineering, Modelling and Simulation Group (EMSG)
The Raspberry Pi is a low cost, credit-card sized computer that plugs into a computer monitor or TV and uses a standard keyboard and mouse. It is a capable little device that enables people of all ages to explore computing, and to learn how to program in languages like Scratch and Python.
To solve current global challenges complex simulations, that require many hours of computer processing time, are often needed – for example, predicting climate behaviour, simulating proteins, and optimising aero-structural performance. High-Performance computer clusters are typically used to reduce the processing time of these simulations from weeks to hours, but this is expensive and not readily accessible. Instead, by combining the computing power of several low-cost Raspberry Pi’s we can emulate the capabilities of much more expensive devices in a modular scalable manner.
The aim of this project was to build the server in a modular manner so that it has the capability of working with more raspberry-pi’s in the future for higher processing power and effectiveness. Skills such as 3D printing, programming and computer networking were required in this project.
The computer cluster has now been built and hosted in the new Engineering Building in UWE Bristol. It features 3D printed prototype parts to hold each individual Raspberry Pi, a PC which was installed with Ubuntu Server 18.04 to network boot the Raspberry Pi clients. In the next step, the students will further boost the cluster’s computational power by doubling the number of Raspberry Pi’s to 42. A variety of simulation tasks will be run to support the final year students’ projects and research projects in UWE Bristol, by using Matlab, Python, OpenFOAM and ANSYS etc.
The Engineering, Modelling and Simulation Group (EMSG) aims to deliver high quality research, postgraduate training and consultancy services in engineering modelling and simulation. We are a multi-disciplinary research group providing a unified platform to bring academics, researchers and research students together. By combining the expertise of each individual, we are able to tackle sophisticated yet practically important engineering problems through individual efforts and teamwork.