Lin, M.-Y.M.-Y.LinCheng, H.-Y.H.-Y.ChengLin, W.-T.W.-T.LinYang, T.-H.T.-H.YangTseng, I.-C.I.-C.TsengYang, C.-L.C.-L.YangHu, H.-W.H.-W.HuChang, H.-S.H.-S.ChangLi, H.-P.H.-P.LiCHIA-LIN YANG2020-05-042020-05-042018https://scholars.lib.ntu.edu.tw/handle/123456789/487719Memristor-based deep learning accelerators provide a promising solution to improve the energy efficiency of neuromorphic computing systems. However, the electrical properties and crossbar structure of memristors make these accelerators error-prone. To enable reliable memristor-based accelerators, a simulation platform is needed to precisely analyze the impact of non-ideal circuit and device properties on the inference accuracy. In this paper, we propose a flexible simulation framework, DL-RSIM, to tackle this challenge. DL-RSIM simulates the error rates of every sum-of-products computation in the memristor-based accelerator and injects the errors in the targeted TensorFlow-based neural network model. A rich set of reliability impact factors are explored by DL-RSIM, and it can be incorporated with any deep learning neural network implemented by TensorFlow. Using three representative convolutional neural networks as case studies, we show that DL-RSIM can guide chip designers to choose a reliability-friendly design option and develop reliability optimization techniques. © 2018 ACM.[SDGs]SDG7Acceleration; Computer aided design; Energy efficiency; Errors; Green computing; Memristors; Neural networks; Reliability; RRAM; Convolutional neural network; Cross-bar structures; Flexible simulation; Learning neural networks; Neural network model; Neuromorphic computing; Reliability optimization; Simulation framework; Deep learningconference paper10.1145/3240765.3240800https://www.scopus.com/inward/record.uri?eid=2-s2.0-85058160814&doi=10.1145%2f3240765.3240800&partnerID=40&md5=14a0105cb1e5abbb0b5831902d7fd955