4 ms·
The future is inference. Many inference stacks already support AMD although the kernels are less optimized. This will of course change over time, but if AMD can
by qeternity 2y ago
The future is inference. Many inference stacks already support AMD although the kernels are less optimized. This will of course change over time, but if AMD can crack the inference demand, it will put NVDA under huge pressure.