摘要(英) |
With the demand of the high memory density and fast memory load/store performance, Skyrmion racetrack memory emerges as a leading contender to replace current DRAM technology in main memory applications. However, the distinctive features of skyrmion racetrack memory, including its ability to shift, generate, and destroy skyrmion elements, diverge significantly from traditional DRAM. Without reevaluating conventional algorithms and data structures to align with the unique properties of skyrmion-based main memory, system performance becomes poor. The swapping operation, fundamental in various algorithms like sorting, searching, graph, and machine learning, poses challenges when directly applied to skyrmion-based main memory. The conventional approach can lead to heightened power consumption from frequent skyrmion element creation or suboptimal system performance due to frequent shifting operations. To address this, our paper proposes a skyrmion-friendly approach, introducing a swappingfriendly racetrack memory architecture to concurrently minimize energy consumption and access latency. Furthermore, we integrate this solution with a decision tree-based machine learning framework, recognizing the pivotal role of the swapping process in frequent execution during decision tree construction. Our experiments yield encouraging results, demonstrating reductions in energy consumption and access latency. |
參考文獻 |
[1] K. Bache and M. Lichman. 2013. UCI Machine Learning Repository. http://archive.ics.uci.edu/ml
[2] P. F. Bessarab, D. Yudin, D. R. Gulevich, P. Wadley, M. Titov, and Oleg A. Tretiakov.2019. Stability and lifetime of antiferromagnetic skyrmions. Phys. Rev. B 99 (Apr2019),140411. Issue
14.https://doi.org/10.1103/PhysRevB.99.140411
[3] Jin-Wei Chang and Tseng-Yi Chen. 2022. When B-Tree Meets Skyrmion Memory:How Skyrmion Memory Affects an Indexing Scheme. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems 41, 11 (2022), 3814–3825.https://doi.org/10.1109/TCAD.2022.3197519
[4] Tianqi Chen and Carlos Guestrin. 2016. XGBoost: A Scalable Tree Boosting System. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (San Francisco, California, USA) (KDD’16). Association for Computing Machinery, New York, NY, USA, 785–794. https://doi.org/10.1145/2939672.2939785
[5] Albert Fert, Nicolas Reyren, and Vincent Cros. 2017. Magnetic skyrmions: advances in physics and potential applications. Nature Reviews Materials 2, 7 (2017), 1–15.
[6] Yanhong Gu, Yilin Wang, Jiaqi Lin, Jonathan Pelliciari, Jiemin Li, Myung-Geun Han, Marcus Schmidt, Gabriel Kotliar, Claudio Mazzoli, Mark PM Dean, et al. 2022. Site-specific electronic and magnetic excitations of the skyrmion material
Cu2OSeO3. Communications Physics 5, 1 (2022), 156.
[7] Yun-Shan Hsieh, Po-Chun Huang, Ping-Xiang Chen, Yuan-Hao Chang, Wang
Kang, Ming-Chang Yang, and Wei-Kuan Shih. 2020. Shift Limited Sort: Optimizing Sorting Performance on Skyrmion Memory-Based Systems. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems 39, 11 (2020), 4115–4128. https://doi.org/10.1109/TCAD.2020.3012880
[8] Wang Kang, Bi Wu, Xing Chen, Daoqian Zhu, Zhaohao Wang, Xichao Zhang, Yan Zhou, Youguang Zhang, and Weisheng Zhao. 2019. A Comparative Cross-Layer Study on Racetrack Memories: Domain Wall vs Skyrmion. J. Emerg. Technol.Comput. Syst. 16, 1, Article 2 (oct 2019), 17 pages.https://doi.org/10.1145/3333336
[9] Yong-Cheng Liaw, Shuo-Han Chen, Yuan-Hao Chang, and Yu-Pei Liang. 2023.Sky-NN: Enabling Efficient Neural Network Data Processing with Skyrmion Racetrack Memory. In 2023 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED). 1–6.https://doi.org/10.1109/ISLPED58423.2023.10244351
[10] Scott M Lundberg, Gabriel Erion, Hugh Chen, Alex DeGrave, Jordan M Prutkin,Bala Nair, Ronit Katz, Jonathan Himmelfarb, Nisha Bansal, and Su-In Lee. 2020.
From local explanations to global understanding with explainable AI for trees.Nature machine intelligence 2, 1 (2020), 56–67.
[11] Sougata Mallick, Sujit Panigrahy, Gajanan Pradhan, and Stanislas Rohart. 2022.Current-Induced Nucleation and Motion of Skyrmions in Zero Magnetic Field.Phys. Rev. Appl. 18 (Dec 2022), 064072. Issue 6. https://doi.org/10.1103/
PhysRevApplied.18.064072
[12] Zhenyu Sun, Wenqing Wu, and Hai (Helen) Li. 2013. Cross-Layer Racetrack Memory Design for Ultra High Density and Low Power Consumption. In Proceedings of the 50th Annual Design Automation Conference (Austin, Texas) (DAC ’13).
Association for Computing Machinery, New York, NY, USA, Article 53, 6 pages.https://doi.org/10.1145/2463209.2488799
[13] Marvin N. Wright and Andreas Ziegler. 2017. ranger: A Fast Implementation of Random Forests for High Dimensional Data in C++ and R. Journal of Statistical Software 77, 1 (2017), 1–17. https://doi.org/10.18637/jss.v077.i01
[14] Ya-Hui Yang, Shuo-Han Chen, and Yuan-Hao Chang. 2022. Evolving Skyrmion Racetrack Memory as Energy-Efficient Last-Level Cache Devices (ISLPED ’22).Association for Computing Machinery, New York, NY, USA, Article 8, 6 pages.https://doi.org/10.1145/3531437.3539709
[15] Ruifu Zhang, Chunli Tang, Xiaozhen Sun, Mengyuan Li, Wencan Jin, Peng Li,Xiaomin Cheng, and X. Sharon Hu. 2023. Sky-TCAM: Low-Power SkyrmionBased Ternary Content Addressable Memory. IEEE Transactions on Electron Devices 70, 7 (2023), 3517–3522.https://doi.org/10.1109/TED.2023.3274506
[16] Daoqian Zhu, Wang Kang, Sai Li, Yangqi Huang, Xichao Zhang, Yan Zhou, and Weisheng Zhao. 2018. Skyrmion Racetrack Memory With Random Information Update/Deletion/Insertion. IEEE Transactions on Electron Devices 65, 1 (2018), 87–95. https://doi.org/10.1109/TED.2017.2769672 |