USENIX ATC '23 - Arbitor: A Numerically Accurate Hardware Emulation Tool for DNN Accelerators

USENIX ATC '23 - Arbitor: A Numerically Accurate Hardware Emulation Tool for DNN Accelerators

Chenhao Jiang, University of Toronto and Vector Institute, Anand Jayarajan, University of Toronto and Vector Institute, Hao Lu, University of Toronto, Gennady Pekhimenko, University of Toronto and Vector Institute

Recently there has been considerable attention on designing and developing hardware accelerators for deep neural network (DNN) training workloads. However, designing DNN accelerators is often challenging as many commonly used hardware optimization strategies can potentially impact the …

emulation hardware tool toronto university university of toronto usenix