This is a news story, published by The Next Platform, that relates primarily to AI news.
For more tech industry trends news, you can click here:
more tech industry trends newsFor more news from The Next Platform, you can click here:
more news from The Next PlatformOtherweb, Inc is a public benefit corporation, dedicated to improving the quality of news people consume. We are non-partisan, junk-free, and ad-free. We use artificial intelligence (AI) to remove junk from your news feed, and allow you to select the best tech news, business news, entertainment news, and much more. If you like tech industry trends news, you might also like this article about
Nvidia investment. We are dedicated to bringing you the highest-quality news, junk-free and ad-free, about your favorite topics. Please come every day to read the latest Ultra Accelerator Link consortium news, B100 GPU accelerators news, tech industry trends news, and other high-quality news about any topic that interests you. We are working hard to create the best news aggregator on the web, and to put you in control of your news feed - whether you choose to read the latest news through our website, our news app, or our daily newsletter - all free!
GPU accelerator roadmapsThe Next Platform
•72% Informative
Xscape Photonics is an optical interconnect startup that is spinning out of research at Columbia University .
Xscape raised $44 million in its Series A venture round, following a $13 million seed round after its founding in 2022 .
The demand for optical intercomms is so high, given the immense bandwidth bottlenecks for accelerator-to-accelerator and accelerator to memory needs, that raising venture funding is not a problem.
Xscape Photonics wants to solve the problem of tapering bandwidth in AI accelerators.
The problem is that the bandwidth between compute elements or memory starts to taper off, and rather quickly.
Low utilization on very expensive devices is a problem for AI training and inference.
The Xscape team has a laser that can drive multiple wavelengths at the same time of a fiber like 128 colors.
The ChromX photonics platform is programmable, so the number of wavelengths provided matches the need of a particular AI training or inference workload.
With this architecture, the HBM memory is not attached to the GPUs, but is glued together in banks that would probably be implemented on physically distinct shelves in a rack or across entire racks.
The idea is to reduce the overall power consumption of the interconnects for accelerators and their memories by 10X.
VR Score
72
Informative language
74
Neutral language
66
Article tone
semi-formal
Language
English
Language complexity
58
Offensive language
not offensive
Hate speech
not hateful
Attention-grabbing headline
detected
Known propaganda techniques
not detected
Time-value
long-living
External references
1
Source diversity
1
Affiliate links
no affiliate links