RePIM: Joint Exploitation of Activation and Weight Repetitions for In-ReRAM DNN Acceleration
Authors:
- 5 December 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 589-594
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.
Keywords
Funding Information
- Ministry of Science and Technology
This publication has 10 references indexed in Scilit:
- DNN+NeuroSim: An End-to-End Benchmarking Framework for Compute-in-Memory Accelerators with Versatile Device Technologies2019
- Sparse ReRAM engine2019
- 24.1 A 1Mb Multibit ReRAM Computing-In-Memory Macro with 14.6ns Parallel MAC Computing Time for CNN Based AI Edge Processors2019
- Learning the sparsity for ReRAM2019
- DL-RSIM2018
- SNrram: An Efficient Sparse Neural Network Computation Architecture Based on Resistive Random-Access Memory2018
- A 65nm 1Mb nonvolatile computing-in-memory ReRAM macro with sub-16ns multiply-and-accumulate for binary DNN AI edge processors2018
- Compute Caches2017
- A case for intelligent RAMIEEE Micro, 1997
- The JPEG still picture compression standardIEEE Transactions on Consumer Electronics, 1992