Session

DescriptionNear-memory and in-memory acceleration are emerging as powerful paradigms for addressing computational bottlenecks in data-intensive tasks and have applications beyond ML. This session, a fresh look at the roots, covers acceleration of beyond ML applications using unconventional computing architectures with a focus on in-memroy, near-memory and storage accelerators. More specifically, it includes papers ranging from acceleration of vector similarity search, fully homomorphic encryption, algorithms for mapping and scheduling optimization, and in-cache accelerators with ISA extension.