GPU · Graphics Processing Unit
A massively parallel processor that became the workhorse of AI training and inference.
Current numbers
~18x (GB200, 400G) / ~9x (GB300, 800G) / ~7.5x (Vera Rubin, 1.6T)scale-up versus back-end scale-out peak bandwidth per GPU: roughly 8-18x per direction by generation
2.3 kW TDP support per Rubin GPU (Pegatron)Pegatron's published TDP for each Rubin GPU in its RA4803-72N3 rack; an OEM rating, not measured draw
186 GBGB200 usable HBM per GPU