Theme : GPU Computing
John Murray is owner of Fusion Data Science and a research fellow in the geographic data science lab at the University of Liverpool. He's spent more than 40 years working at the cutting edge of data analysis. Before his current roles, John was technical director of NDL, a consumer survey company that's now part of Acxiom. John started his career at Reader's Digest, where he pioneered the use of data analytics to predict customer behaviour
Moore’s Law is Dead. Limitations in the ability to shrink transistors further have stalled the exponential growth of CPU power in the last few years. To satisfy the demands of modern computing requirements, data scientists are increasing turning to parallelisation to analyse large and complex datasets.
The massively parallel computing capability of Graphics Processing Units (GPUs), originally designed for gaming, is increasingly being leveraged to perform computationally intensive tasks such as mathematical modelling, statistical analysis, graph analysis and deep learning on large datasets.
John will present an introduction to GPU computing, followed by a live demonstration of NVIDIA’s new open source RAPIDS accelerated data science platform, which implements Pandas like dataframes on GPUs, achieving speed ups of more than 1000 times that of a top end CPU on large datasets.
RAPIDS also simplifies the process of implementing GPU computing by removing the need to learn CUDA through a Python interface. John will also show how you can easily port your existing functions to work with RAPIDS.
Talk 1 - 7:15pm - Session 1
Break for drinks & networking
Talk 2 - 8:00pm - Session 2
There will be time for Q&A as well as networking at the bar downstairs.