Hacker News new | past | comments | ask | show | jobs | submit
teaearlgraycold | on: AMD acquires Taalas to boost inference performance by etching models in silicon
Is that analogue or are they baking floating points into the silicon?
AlotOfReading | parent
It's entirely possible they're using something like block floating point, where most of the hardware is simply fixed point. AMD's NPU does this, for example.