The Ultra Accelerator Link (UALink) Consortium, founded last summer, brings together some of the biggest names in semiconductors and data centres to develop standards to enhance connecting multiple GPUs.
The trio become the newest members of the group’s Board of Directors and will help to drive the development of UALink technology ahead of Version 1.0 of its standard, which is set to launch during the first quarter of 2025.
Subscribe today for free
“UALink shows great promise in addressing connectivity challenges and creating new opportunities for expanding AI capabilities and demands,” said Becky Loop, director of platform architecture at Apple. "Apple has a long history of pioneering and collaborating on innovations that drive our industry forward, and we’re excited to join the UALink Board of Directors.”
UALink is used to connect multiple GPUs to create larger, powerful clusters of compute used to power workloads like AI training.
The soon-to-be-released Version 1.0 of the UALink standard will enable firms to achieve speeds of up to 200 gigabytes per second (Gbps) per lane scale-up for up to 1024 accelerators in a single cluster.
The creation of UALink standard aims to provide cluster operators with the ability to get more computing power out of their hardware and scale capacity without disrupting ongoing workloads.
Also involved in the group include AMD, AWS, Cisco, Google, Intel, Meta, and Microsoft, with members also participating in the Ultra Ethernet Consortium (UEC), an industry group cooperation around Ethernet-based networking.
One company that won’t be joining the group is Nvidia, as UALink would compete with its proprietary NVLink interconnect.
Qiang Liu, VP of Alibaba Cloud and general manager of Alibaba Cloud Server Infrastructure, said: “The UALink consortium, as a leader in the interconnect field of AI accelerators, has brought together key members from the AI infrastructure industry to work together to define interconnect protocol which is natively designed for AI accelerators, driving innovation in AI infrastructure.
“This will strongly promote the innovation of AI infrastructure and improve the execution efficiency of AI workloads, contributing to the establishment of an open and innovative industry ecosystem.”
RELATED STORIES
AI surge driving global data centre market growth: study
IBM researchers unveil optics tech to supercharge AI training speeds
