3 ms·
Ignoring latency but not throughput, How does this compare in terms of Cost ( cards Acquisition cost and Power needed) with Nvidia GPU for inference?
by supercharger9 3y ago
Ignoring latency but not throughput, How does this compare in terms of Cost ( cards Acquisition cost and Power needed) with Nvidia GPU for inference?
- matanyal 3y agoWe intend to be very competitive on cost, power, hardware, TCO, whatever it is. Custom-built silicon+hardware has the advantage in this space.