← Back to context Comment by numpad0 4 days ago Maybe latency. IIRC flash is a lot laggier than DRAMs and SRAMs. 1 comment numpad0 Reply DoctorOetker 3 days ago The random access memory models is not really representative of ML workloads (both training and inference), where multiplying large tensors result in predictable memory access patterns.
DoctorOetker 3 days ago The random access memory models is not really representative of ML workloads (both training and inference), where multiplying large tensors result in predictable memory access patterns.
The random access memory models is not really representative of ML workloads (both training and inference), where multiplying large tensors result in predictable memory access patterns.