1 paper
Junrui Pan, Weili An, Cesar Avalos Baddouh +9
The rapid evolution of machine learning workloads has fundamentally transformed GPU hardware, driving architectures toward Multi-Chip Module (MCM) topologies, asynchronous executio…