Ernest Y Wu, Takashi Ando, et al.
IEDM 2023
Analog Non-Volatile Memory-based accelerators offer high-throughput and energy-efficient Multiply-Accumulate operations for the large Fully-Connected layers that dominate Transformer-based Large Language Models. We describe architectural, wafer-scale testing, chip-demo, and hardware-aware training efforts towards such accelerators, and quantify the unique raw-throughput and latency benefits of Fully- (rather than Partially-) Weight-Stationary systems.
Ernest Y Wu, Takashi Ando, et al.
IEDM 2023
Irem Boybat-Kara
IEDM 2023
Laura Bégon-Lours, Mattia Halter, et al.
MRS Spring Meeting 2023
Ying Zhou, Gi-Joon Nam, et al.
DAC 2023