All of our events may be recorded and shared via the University of Sheffield Kaltura platform so those who cannot attend may still benefit. We will consider your attendance implict consent to this.
GPU computing offers great benefits to simulation performance, but can certainly be intimidating to get to grips with! Join us for a set of short talks describing projects in which the GPU programming aspect has been made more portable in some way, whether through alternative language bindings, or cross-platform portability.
Each talk will be ~10 minutes, leaving roughly 20 minutes for discussion/questions.
This session will take place on Google Meet and participants can join 15 mins before the start of the session.
We also have a Google Jam Board where you can note down any questions or comments before or during the event.
Freddie Witherden, Assistant Professor at Texas A&M University, developer of PyFR, a CUDA accelerated python CFD library
James Knight, Research Fellow in Computer Science at The University of Sussex, developer of PyNN/GeNN, a GPU enhanced Neuronal Network simulation environment
Large-scale numerical simulations of brain circuit models are important for identifying hypotheses on brain functions and testing their consistency and plausibility. Similarly, spiking neural networks are also gaining traction in machine learning with the promise that neuromorphic hardware will eventually make them much more energy efficient than classical ANNs. In this session, I will present the GeNN (GPU-enhanced Neuronal Networks) framework, which uses GPUs to accelerate computational models of large-scale spiking neuronal networks. GeNN is an open source library that generates code to accelerate the execution of network simulations on NVIDIA GPUs through a flexible and extensible interface, which does not require in-depth technical knowledge from the users.
Tom Deakin, Senior Research Associate at Bristol University, researcher in understanding performance portability of massively parallel simulation codes
The range of computer architectures used in supercomputers today is growing in diversity, and we need to obtain high-performance on CPUs and GPUs from a variety of vendors. This talk highlights ongoing research into performance portability and will discuss different parallel programming models that, with some work, are showing it is possible to write codes that achieve good performance everywhere.
For queries relating to collaborating with the RSE team on projects: rse@sheffield.ac.uk
Information and access to JADE II and Bede.
Join our mailing list so as to be notified when we advertise talks and workshops by subscribing to this Google Group.
Queries regarding free research computing support/guidance should be raised via our Code clinic or directed to the University IT helpdesk.