Accelerating Computational Fluid Dynamics

Greetings all! Join us for a 40 min conversation with IBM Research around Computational Fluid Dynamics, cutting-edge research, GPU-acceleration and AI applications. This will be an online event run over webex so please RSVP here first and then follow the link to register in advance :)

Computational fluid dynamics (CFD) represents a major class of codes currently running on the world’s largest supercomputers.
Areas of application range from sculpting the optimal form of vehicles to informing safety-critical design of nuclear reactors. However simulations are highly complex and iterative, potentially taking weeks or months to complete, coming up against the limits of Moore’s law.

In this session we’re delighted to be joined by Samuel Antao from IBM Research to share his work around use of GPUs to accelerate open-source CFD code. More specifically we’ll discuss overcoming challenges in accelerating code when working with unstructured sparse data sets, scaling and reducing latencies. We will conclude with some applications of the technology involving the use of AI within CFD for enhanced design and 3D feature detection.