What is Ultra Accelerator Link (UALink™) and how will it be used to build massive clusters of GPUs?
Kurtis Bowman, Chairman from UALink Consortium explains:
– UALink enables connecting up to 1024 GPUs together, dramatically increasing memory and compute capacity for large AI models with billions to trillions of parameters
– The technology delivers ultra-low latency (100-150 nanoseconds) through single-switch architecture and power-efficient memory sharing between GPUs at up to 800 gigabits per second
– As an open standard supported by major tech companies like AMD, Apple, Google, and Meta, UALink complements Ultra Ethernet Consortium (UEC) to enable both scale-up within pods and scale-out across pods for massive AI training deployments
Want to be involved our video series? Contact info@nextgeninfra.io
