Aiming to become the global leader in chip-scale photonic solutions by deploying Optical Interposer technology to enable the seamless integration of electronics and photonics for a broad range of vertical market applications

Free
Message: Applications at the bottleneck

https://www.nextplatform.com/2024/05/15/one-cerebras-wafer-beats-an-exascale-super-at-molecular-dynamics/

Interesting article that describes the reason current super compute systems are not sufficient for the types of problems we need to solve right now. 

"Here is the crux of the problem, and it has to do with the weak scaling of modern massively parallel supercomputers versus the strong scaling of individual compute engines, and what it really comes down to is having high bandwidth between all compute elements and their local memory. "

One way is to put all the components on the same wafer like Cerebras but that's incredibly complex for an iterative design process and more difficult to manufacture with one defect affecting other components and potentially the entire system. Is this type of design going to win out against photonics when it comes to bandwidth? Physics says not likely...

The applicability of the platform continues to grow as near term applications of the technology grow into POETs products very quickly.

DOE would be nice partner given their need for solutions in regards to data bandwidth and system scaling. 

Share
New Message
Please login to post a reply