arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2109.14878cs.DCcs.AIcs.LG
收藏

加速全连接神经网络片上光网络(ONoC)

Accelerating Fully Connected Neural Network on Optical Network-on-Chip (ONoC)

University of Otago(奥塔哥大学)

Fei Dai, Yawen Chen, Haibo Zhang, Zhiyi Huang

摘要:

全连接神经网络(FCNN)是一类广泛应用于计算机科学和工程的人工神经网络,而在现有的众核系统中,训练过程可能需要很长时间来处理大型数据集。片上光网络(Optical Network-on-Chip,ONoC)是一种新兴的芯片级光互连技术,具有很大的潜力来加速低传输延迟、低功耗和高吞吐量的FCNN的训练。然而,现有的基于片上网络(ENoC)的方法不能适应ONoC,因为ONoC的独特属性。在本文中,我们提出了一种细粒度并行计算模型,用于加速ONoC上的FCNN训练,并推导出每个执行阶段的最佳核心数量,目标是最小化完成一个FCNN训练时期的总时间。为了给每个执行阶段分配最优的核数,我们提出了三种映射策略,并在热点级别、内存需求和状态转换方面比较了它们的优缺点。仿真结果表明,神经网络基准测试中最佳核数的平均预测误差在2.3%以内。我们进一步进行了大量的模拟,结果表明,与分配固定数量的核心或分配尽可能多的核心的传统并行计算方法相比,使用我们提出的方案,FCNN训练时间平均可以减少22.28%和4.91%。仿真结果表明,在批量为64和128的情况下,与ENoC相比,ONoC平均训练时间分别减少了21.02%和12.95%,平均能耗分别节省了47.85%和39.27%.

英文摘要:

Fully Connected Neural Network (FCNN) is a class of Artificial Neural Networks widely used in computer science and engineering, whereas the training process can take a long time with large datasets in existing many-core systems. Optical Network-on-Chip (ONoC), an emerging chip-scale optical interconnection technology, has great potential to accelerate the training of FCNN with low transmission delay, low power consumption, and high throughput. However, existing methods based on Electrical Network-on-Chip (ENoC) cannot fit in ONoC because of the unique properties of ONoC. In this paper, we propose a fine-grained parallel computing model for accelerating FCNN training on ONoC and derive the optimal number of cores for each execution stage with the objective of minimizing the total amount of time to complete one epoch of FCNN training. To allocate the optimal number of cores for each execution stage, we present three mapping strategies and compare their advantages and disadvantages in terms of hotspot level, memory requirement, and state transitions. Simulation results show that the average prediction error for the optimal number of cores in NN benchmarks is within 2.3%. We further carry out extensive simulations which demonstrate that FCNN training time can be reduced by 22.28% and 4.91% on average using our proposed scheme, compared with traditional parallel computing methods that either allocate a fixed number of cores or allocate as many cores as possible, respectively. Compared with ENoC, simulation results show that under batch sizes of 64 and 128, on average ONoC can achieve 21.02% and 12.95% on reducing training time with 47.85% and 39.27% on saving energy, respectively.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑