AMD already uses a Perceptron for their branch prediction. I would say they're talking about support for ML speed ups in hardware but maybe the plans also inclu

AMD already uses a Perceptron for their branch prediction. I would say they're talking about support for ML speed ups in hardware but maybe the plans also include a more complex complete neural net in hardware for data prediction.