Uploaded September 2023 | Updated September 2026, 2 weeks ago
Amidst the PCon23 Hackathon buzz, Erik Drake, the Director of the High Performance Computing Innovation Centre and Scientific Computing group leader at the Lawrence Livermore National Laboratory extended a special invitation to join his session on Friday. Register now to secure your spot!
bit.ly/3L5wGWf
Friday 22 September 2023
10:30am AWST
Online
Video transcript
Hello, everybody, my name is Eric Drager, and I'm from Lawrence Livermore National Laboratory. And on Friday, I'm going to be talking about our experiences preparing our applications. The El Capitan supercomputer. El Capitan is going to be our first AMD GPU-based supercomputer, very similar to Setonix. And we have many applications that were developed for CPU-based machines over decades. And so rewriting those codes for the GPU is not an option. We need to adapt them to run well on these new architectures. And we don't just want to do this once and we want to do this for the long term. So I'm hoping that the experiences I can share will be relevant to all sorts of scientific applications. I hope you'll come here what we've learned what what challenges we still face, and together we can try and work together to to solve these problems more effectively.
Amidst the PCon23 Hackathon buzz, Erik Drake, the Director of the High Performance Computing Innovation Centre and Scientific Computing group leader at the Lawrence Livermore National Laboratory extended a special invitation to join his session on Friday. Register now to secure your spot!
bit.ly/3L5wGWf
Friday 22 September 2023
10:30am AWST
Online
Video transcript
Hello, everybody, my name is Eric Drager, and I'm from Lawrence Livermore National Laboratory. And on Friday, I'm going to be talking about our experiences preparing our applications. The El Capitan supercomputer. El Capitan is going to be our first AMD GPU-based supercomputer, very similar to Setonix. And we have many applications that were developed for CPU-based machines over decades. And so rewriting those codes for the GPU is not an option. We need to adapt them to run well on these new architectures. And we don't just want to do this once and we want to do this for the long term. So I'm hoping that the experiences I can share will be relevant to all sorts of scientific applications. I hope you'll come here what we've learned what what challenges we still face, and together we can try and work together to to solve these problems more effectively.
