Hardwired memory-compute locality
The only key hardwired architectural feature is tight memory-compute locality—specifically, affinity between HBM and compute cores—to improve performance, power efficiency, and reduce data movement.
Supporting evidence
Inside OpenAI’s First Chip
Original excerpt
The one thing that is really unique about our architecture, that is kind of hardwired in, is the locality and affinity between memory and compute. That’s the key architectural difference of this device.
Share insightCheck this claim