Litcius/Paper detail

RePIM: Joint Exploitation of Activation and Weight Repetitions for In-ReRAM DNN Acceleration

Chen‐Yang Tsai, Chin-Fu Nien, Tz-Ching Yu, Hung-Yu Yeh, Hsiang-Yun Cheng

202122 citationsDOI

Abstract

Eliminating redundant computations is a common approach to improve the performance of ReRAM-based DNN accelerators. While existing practical ReRAM-based accelerators eliminate part of the redundant computations by exploiting sparsity in inputs and weights or utilizing weight patterns of DNN models, they fail to identify all the redundancy, resulting in many unnecessary computations. Thus, we propose a practical design, RePIM, that is the first to jointly exploit the repetition of both inputs and weights. Our evaluation shows that RePIM is effective in eliminating unnecessary computations, achieving an average of $ 15.24\times$ speedup and 96.07% energy savings over the state-of-the-art practical ReRAM-based accelerator.

Topics & Concepts

Resistive random-access memoryComputer scienceComputationRedundancy (engineering)SpeedupExploitAccelerationParallel computingEfficient energy useHardware accelerationComputer engineeringComputer architectureComputer hardwareAlgorithmField-programmable gate arrayElectrical engineeringVoltageEngineeringOperating systemPhysicsComputer securityClassical mechanicsAdvanced Memory and Neural ComputingFerroelectric and Negative Capacitance DevicesParallel Computing and Optimization Techniques
RePIM: Joint Exploitation of Activation and Weight Repetitions for In-ReRAM DNN Acceleration | Litcius