News

34 Member Companies to Showcase Cutting-Edge Solutions, OIF Experts to Speak in Three Market Focus Sessions

Data and computation are fundamental to the success of AI, which is why hyperscale data center operators are racing to build ever larger accelerator-based compute clusters. The market jolt caused by LLMs and ChatGPT has further accelerated the race. Scaling clusters requires innovation across many domains: the accelerator chips, their linkages within a node, the fabrics linking the nodes, whether InfiniBand or Ethernet-based, and their topology. The input-output (I/O) growth of each new generation of accelerator and the growth of bandwidth that needs to escape a compute node means using optical I/O is a matter of time. Each hyperscale operator has its preference; some back co-packaged optics (CPO), while others favor pluggable optics. But the use of optics continues to grow in such clusters. For example, Google uses its optical circuit switches as part of its TPU v4 accelerator-based AI compute clusters to reduce cost and improve system reliability and flexibility. And Broadcom has already developed CPO-based Tomahawk Ethernet switches; its latest Jericho3-AI chip developed to scale AI compute clusters can now also be CPO-based. This webinar will discuss the challenges of scaling AI clusters as well as the emerging solutions for improving optical I/O density, bandwidth and power efficiency.

Lessengers announces its immersion cooling proof AI/ML datacenter optics solutions at ISC High Performance 2024

LESSENGERS Inc., a provider of innovative optical components based on its patented “direct optical wiring” (DOW) technology, today announced a comprehensive portfolio of 800G optics designed for AI/ML workloads in hyperscale data centers. Lessengers is adding partially retimed capability to its 800G transceiver product portfolio, which in one product entails an integrated DSP chip on the transmitter side.

Lessengers will showcase its latest advances in 800G AI/ML optics solutions at OFC 2024 and announce new product launch.

The optical I/O movement has many facets and promising technologies, and each has a significant impact in addressing certain system architectural challenges. DOW can be seen as a useful tool to help them all achieve their maximum potential when it comes to the higher bandwidth, lower latency and improved energy efficiency we will need for near-term 800G and 1.6T speed connections in the data center, maximized scale-out architectures and the longer-term zettascale computing platforms.

After a successful live demonstration of polymer-wired 800G QSFP-DD and 400G QSFP112 active optical cables...

Coverage of our 800G transceiver announcement at ECOC 2023

Coverage of our announcement at ECOC 2023 in Lightwave

Ramping up production to meet the demand for optical connectivity in AI clusters in hyperscale, HPC and supercomputing environments

An incisive analysis of how AI workloads are impacting the data center