top | item 40638263 (no title) alextheparrot | 1 year ago Bit of a detail, but where are you deriving “with hundreds of terabytes of unified GPU memory” from? discuss order hn newest talldayo|1 year ago I was an order of magnitude off, at least in the case of NVL72: https://www.nvidia.com/en-us/data-center/gb200-nvl72/But the point stands, these systems occupy a niche that Apple Silicon is poorly suited to filling. They run normal Linux, they support common APIs, and network to dozens of other machines using Infiniband.
talldayo|1 year ago I was an order of magnitude off, at least in the case of NVL72: https://www.nvidia.com/en-us/data-center/gb200-nvl72/But the point stands, these systems occupy a niche that Apple Silicon is poorly suited to filling. They run normal Linux, they support common APIs, and network to dozens of other machines using Infiniband.
talldayo|1 year ago
But the point stands, these systems occupy a niche that Apple Silicon is poorly suited to filling. They run normal Linux, they support common APIs, and network to dozens of other machines using Infiniband.