Nvidia Rubin CPX forms one half of new, "disaggregated" AI inference architecture — approach splits work between compute- and bandwidth-optimized chips for best performance
by from Latest from Tom's Hardware on (#6ZY18)
Nvidia's Rubin CPX GPU will be Nvidia's complement to the upcoming Vera Rubin AI rack, a wholly inference-focused GPU to work in tandem with high-end Rubin products. The cut-down GPU is paired with onboard GDDR7 memory rather than HBM and will be available as an optional upgrade from the base Vera Rubin.