If you think HPC architectures have changed rapidly in recent years, brace yourself for the future when things will be moving at light speed.
As scientific workloads have grown more variable and demanding, and applications like AI have brought large but fluctuating data demands, tightly coupled systems have been stretched, forcing HPC system architects to respond.
One approach is disaggregated architectures, that decouple processors, memory, accelerators and other components, in dramatic contrast to the highly integrated architectures that dominate much of this market.
But this approach raises significant questions of its own, ranging from how to manage workflows around such systems, to the fundamental technical problems of latency and the need for high bandwidth to speed data between these components.
So, what’s really, really fast? Light of course, and optical I/O is expected to be the foundation of new interconnects that will allow heterogenous connectivity, with tremendous bandwidth, low latency and low power, across a range of new system designs.
These are tantalizing topics, taking in developments right at the edge of computing, and you can catchup with some of the people making this all happen at this webcast on Disaggregated System Architectures for Next Generation HPC and AI Workloads on November 11, from 9 AM to 10 AM PT.
Timothy Prickett Morgan of The Next Platform will be joined by a stellar lineup of speakers drawn from industry and the US’s top supercomputing labs including:
- Dr. Ian Karlin, Principal HPC Strategist at Lawrence Livermore National Laboratory
HPC and AI workloads.
- John Shalf, Head for Computer Science at Lawrence Berkeley National Laboratory
- Dr. Vladimir Stojanovic, Chief Architect at Ayar Labs and Professor of EECS at UC Berkeley
- Dr. Josh Fryman, Senior Principal Engineer at Intel Corp
- Doug Carmean, Architect at Microsoft
Together they will introduce you to what’s already been achieved with existing disaggregated systems, as well as looking forward to future designs, and what will be needed to get us there.
Whether your interest is in solving traditional HPC problems, faster, the introduction of AI into scientific computing, or simply what supercomputing will look like in the post-exascale era, this exclusive broadcast is the place to start. And all you need to do to join this lineup of visionaries is sign up here.
Sign up to our Newsletter
Featuring highlights, analysis, and stories from the week directly from us to your inbox with nothing in between.
Be the first to comment