南京大学学报(自然科学版) ›› 2019, Vol. 55 ›› Issue (5): 733–739.doi: 10.13232/j.cnki.jnju.2019.05.004
一种负载均衡的LSTM硬件加速器设计
查羿,潘红兵(
)
- 南京大学电子科学与工程学院，南京，210093
A load balanced LSTM hardware accelerator design
Yi Zha,Hongbing Pan()
- School of Electronic Science and Engineering，Nanjing University, Nanjing, 210093, China
摘要：
神经网络在嵌入式端的应用日益广泛，为满足嵌入式端低功耗，低延迟等特点，通常的解决方案是针对长短记忆序列LSTM模型（Long?Short Term Memory）进行压缩，并定制专用的硬件加速器.当LSTM模型经过剪枝等压缩操作后，其网络模型将变得稀疏且不规则，会给PE（Process Element）运算单元带来负载不均衡的问题.通过排序的方法，将权重矩阵按一定的规则重新分发给各个PE单元，并在此基础上针对稀疏化的模型定制专用的硬件单元.在赛灵思zynq系列XCZU9EG?2FFVB1156E开发板上进行实验，实验结果显示，当PE单元多消耗0.314%硬件资源的情况下，其运算速度取得了2%的提升.
中图分类号:
- TN492
| 1 | Graves A ， Mohamed A R ， Hinton G . Speech recognition with deep recurrent neural networks∥Proceedings of the 2013 IEEE International Conference on Acoustics，Speech and Signal Processing. Vancouver，Canada：IEEE，2013：6645-6649. |
| 2 | Rybalkin V ， Wehn N ， Yousefi M R ，et al . Hardware architecture of bidirectional long short?term memory neural network for optical character recognition∥Design，Automation & Test in Europe Conference & Exhibition (DATE)，2017.Lausanne，Switzerland：IEEE，2017. |
| 3 | Cho K ， Van Merrienboer B ， Gulcehre C ，et al . Learning phrase representations using RNN encoder?decoder for statistical machine translation. arXiv：1406.1078，2014. |
| 4 | Cho K , Van Merrienboerning B , Gulcehre C ，et al .Learning phrase representations using RNN encoder?decoder for statistical machine translation. arXiv：1406.1078，2014. |
| 5 | Ye J M ， Wang L N ， Li G X ，et al . Learning compact recurrent neural networks with block?term tensor decomposition. arXiv：1712.05134，2018. |
| 6 | Hochreiter S ， Schmidhuber J . Long short?term memory. Neural Computation，1997，9(8)：1735-1780. |
| 7 | Guan Y J ， Yuan Z H ， Sun G Y ，et al . FPGA?based accelerator for long short?term memory recurrent neural networks∥2017 22nd Asia and South Pacific Design Automation Conference (ASP?DAC). Chiba，Japan：IEEE，2017：629-634. |
| 8 | Ouyang P ， Yin S Y ， Wei S J . A fast and power efficient architecture to parallelize LSTM based RNN for cognitive intelligence applications∥2017 54th ACM/EDAC/IEEE Design Automation Conference (DAC). Austin，TX，USA：IEEE，2017：1-63. |
| 9 | Han S ， Kang J L ， Mao H Z ，et al . ESE：Efficient speech recognition engine with sparse LSTM on FPGA∥Proceedings of the 2017 ACM/SIGDA International Symposium on Field?Programmable Gate Arrays. Monterey，CA，USA：ACM，2017. |
| 10 | Han S ， Liu X Y ， Mao H Z ，et al . EIE：Efficient inference engine on compressed deep neural network∥2016 ACM/IEEE 43rd Annual Interna?tional Symposium on Computer Architecture (ISCA). Seoul，South Korea：IEEE，2016：243-254. |
| 11 | Wang Z S ， Lin J ， Wang Z F . Accelerating recurrent neural networks： A memory?efficient approach. IEEE Transactions on Very Large Scale Integration (VLSI) Systems，2017，25(10)：2763-2775. |
| 12 | Wang S ， Li Z ， Ding C W ，et al . C?LSTM： Enabling efficient LSTM using structured compression techniques on FPGAs∥Proceedings of the 2018 ACM/SIGDA International Sympo?sium on Field?Programmable Gate Arrays. Monterey，CA，USA：ACM，2017. |
| 13 | Conti F ， Cavigelli L ， Paulin G ，et al . CHIPMUNK：A systolically scalable 0.9 mm2,3.08 Gop/s/mW @1.2mW accelerator for near?sensor recurrent neural network inference. arXiv： 1711.05734，2018. |
| 14 | Liu B ， Dong W ， Xu T T ，et al . E?ERA：An energy?efficient reconfigurable architecture for RNNs using dynamically adaptive approximate computing. IEICE Electronics Express，2017，14(15)：20170637. |
| 15 | Shin D ， Lee J ， Lee J ，et al . 14.2 DNPU：An 8.1TOPS/W reconfigurable CNN?RNN processor for general?purpose deep neural networks∥2017 IEEE International Solid?State Circuits Conference (ISSCC). San Francisco ，CA，USA：IEEE，2017. |
| 16 | Park J ， Kung J ， Yi W ，et al . Maximizing system performance by balancing computation loads in LSTM accelerators∥2018 Design，Automation & Test in Europe Conference & Exhibition (DATE). Dresden，Germany：IEEE，2018. |
| No related articles found! |
|