E Wu, X Zhang, D Berman, I Cho… - Proceedings of the 2019 …, 2019 - dl.acm.org
… , maximizing both operating clock rates and compute efficiency is … neuralnetworkaccelerator that operates at 92.9% of the peak FPGA clock rate. First, we map neural-network operators …
… flexible interconnection, the DNN accelerator can support different … accelerator design. This paper systematically investigates the interconnection networks in modern DNN accelerator …
W Liu, J Lin, Z Wang - … Transactions on Circuits and Systems I …, 2020 - ieeexplore.ieee.org
… This paper presents an energy-efficient precision-scalable CNN accelerator which reduces the computational complexity of processing CNN models. In this accelerator, a complexity…
… to improve the efficiency of computing systems by trading … efficient automated framework for mapping the NN weights to the accuracy levels of the approximate reconfigurable accelerator…
E Park, D Kim, S Yoo - 2018 ACM/IEEE 45th Annual …, 2018 - ieeexplore.ieee.org
… accelerator called OLAccel that implements 4-bit computations on very deep neuralnetworks… OLAccel differently handles outlier activations and weights for computational efficiency. • …
X Xie, J Lin, Z Wang, J Wei - … on Circuits and Systems I: Regular …, 2021 - ieeexplore.ieee.org
… power efficiency with the same sparsity. Compared to prior dense accelerators, this accelerator … To resolve this issue, an efficient sparsity-aware accelerator for BLOs is developed in this …
… efficiency compared to general-purpose platforms [5, 7, 12, 18, 36]. Ideally, we would like to continue scaling the performance and efficiency of NN accelerators … for NN accelerators. To …
… and processing elements connected through a dedicated on-chip network. Compared with the state-of-the-art accelerators which only exploit the input sparsity, SparseNN can achieve a …
… In this paper, a GRU-RNN accelerator architecture called the DeltaRNN (DRNN) is proposed. This implementation is based on the Delta Network (DN) algorithm that skips dispensable …