Ampere claimed that an 80-core Ampere Altra CPU enables a 28 percent cost savings compared to Nvidia’s A10 GPU for producing one million tokens running at roughly 80 tokens per second with Meta ...

The Ampere server could either be eight GPUs working together for training, or it could be 56 GPUs made for inference,' Nvidia CEO Jensen Huang says of the chipmaker's game changing A100 GPU.

According to the test, Nvidia's Ampere GPUs - which constitute the RTX 30 series of graphics cards - maybe be far more easily bottlenecked by less powerful CPUs in comparison to AMD's Radeon cards.