Inaugural HPC/GPU Meetup

Location is the Gustav Tuck Lecture Theatre, Second Floor Wilkins Building, UCL, London WC1E 6BT. Map is here:

http://www.ucl.ac.uk/maps/gustave-tuck-lt

Two great speakers:

1. OpenCL: The shape of GPU programming to come? Malcolm Sherrington

If you are looking to get into GPU programming, you are faced with the following choice: OpenCL or CUDA? In this talk Malcolm outlines the major differences between the two approaches and the factors you may wish to take into consideration when making the choice.

The second half of the talk discusses the work of the Khronos Group, focusing on OpenCL and demonstrating how easy it is to get started with GPU programming without the need of specialist hardware or complex programming paradigms.

Malcolm has been working on parallel processing and asynchronous computing projects for over 25 years. He is currently writing the PACKT book "Mastering Julia",https://www.packtpub.com/application-development/mastering-julia

2.Deep Learning with GPU's - Alison Lowndes

Alison Lowndes who is a Deep Learning Solutions Architect at NVIDIA will be giving us the lowdown on GPU use for various deep learning techniques and use cases plus the latest news on DIGITS, their open-source Deep Learning platform. Convolutional and recurrent neural nets and the cuDNN library will all be covered along with the various Nvidia hardware on which the DL algorithms are run. Further information can be found here: https://developer.nvidia.com/digits

Alison is a 2015 Graduate in Computer Science and Artificial Intelligence from the University of Leeds, combining technical and theoretical knowledge of deep learning, specifically with GPU technology. She undertook an empirical study of deep learning, supported by NVIDIA and Facebook, running convolutional neural nets on GPU's for image and feature recognition. The end goal was to teach machines how to recognize signs of cancer in high-resolution tissue images. She also studied astrophysics in a past lifetime.