Workshop on GPU computing on September 20

Over the last decade, GPUs (graphical processing units) have emerged as powerful and affordable alternatives to CPUs (central processing units) and CPU clusters. GPUs are used for a wide range of data intensive scientific computing tasks that require, or benefit from, massively parallel processing (e.g.: training neural networks, some physic simulations, 3D movies etc.). GPUs also play an important role in many applications relevant to the Institut Pasteur, for example bioinformatics analyses, microscopy image processing, CryoEM structure determination and molecular dynamics simulations.

Over the past years, the Institut Pasteur has been investing in GPU computing infrastructures for accelerating scientific computing. In addition to the computing resources provided by the DSI (Direction of Computer Systems), 28 professional-grade GPUs of professional quality available on Tars, a GPU lab ("Inception GPU lab") was recently set up at the Institut Pasteur and funded by the INCEPTION program (ANR -16-CONV-0005). The GPU lab provides computing resources complementary to those offered by the DSI. It consists of 3 servers hosting a total of 24 consumer-grade GPUs, and has been successfully tested for selected applications, including molecular dynamics and deep learning.

In order to help researchers at the Institut Pasteur to use these GPU resources, a GPU computing workshop will be organized. The workshop features an overview of the GPU infrastructure, presentation on applications in deep learning and molecular dynamics, and a hands-on training session. Join us on September 20, 2018, at 1.30pm, at the François Jacob Center (28-01-01A room).

Space is limited, so please register before September 12, 2018, by sending an email to gpulab@pasteur.fr

Please bring your own laptop for the hands-on training session.