一種用于加速FPGA設(shè)計空間探索的電路特性驅(qū)動半監(jiān)督建模方法
doi: 10.11999/JEIT150162
-
2.
(中國科學院電子學研究所 北京 100190) ②(中國科學院大學 北京 100049)
基金項目:
國家自然科學基金(61271149)
Circuit Characteristics-driven Semi-supervised Modelling Approach for Accelerating FPGA Design Space Exploration
-
2.
(Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China)
Funds:
The National Natural Science Foundation of China (61271149)
-
摘要: 該文提出一種電路特性驅(qū)動的半監(jiān)督建模方法來探索FPGA架構(gòu)設(shè)計空間。通過加入電路特性作為輸入來構(gòu)建一個通用的FPGA性能模型,該方法能夠精確預(yù)測指定電路在特定FPGA架構(gòu)上實現(xiàn)的性能。實驗結(jié)果顯示該方法在預(yù)測電路在FPGA上實現(xiàn)的面積時,平均相對誤差達到6.25%;預(yù)測延時時,平均相對誤差可達4.23%,具有與半監(jiān)督模型樹(Semi-supervised Model Tree, SMT)方法可比的預(yù)測精度。同時,該文方法加速了FPGA性能建模過程,與SMT方法比較,在6核Intel服務(wù)器平臺Intel Xeon E7-4807上,探索具有百萬架構(gòu)的FPGA設(shè)計空間時,該文方法可將時間成本由500 h降低為250 h。
-
關(guān)鍵詞:
- FPGA設(shè)計空間 /
- 半監(jiān)督建模方法 /
- 電路特性
Abstract: A circuit characteristics-driven semi-supervised modelling approach is proposed for FPGA architecture design space exploration. By including circuit characteristics as input, the proposed approach can estimate the performance of specific circuit on certain architecture accurately. Experimental results illustrate that the approach estimates the area with Mean Relative Error (MRE) up to 6.25%, and delay up to 4.23%, which is comparable to the Semi-supervised Model Tree (SMT) approach. Meanwhile, the proposed approach speedups the modelling process. Compared to the SMT approach, the proposed approach reduces the time cost from 500 h to 250 h when exploring a design space with millions of architectures inside on Intel Xeon E7-4807 platform. -
Luu J, Goeders J, Wainberg M, et al.. VTR 7.0: next generation architecture and CAD system for FPGAs[J]. ACM Transactions on Reconfigurable Technology and Systems, 2014, 7(2): 6:1-6:30. Marquardt A, Betz V, and Rose J. Speed and area tradeoffs in cluster-based FPGA architectures[J]. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, 2000, 8(1): 84-93. Ahmed E and Rose J. The effect of LUT and cluster size on deep-submicron FPGA performance and density[J]. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, 2004, 12(3): 288-298. Jiang Zheng-hong, Lin C Y, Yang Li-qun, et al.. Exploring architecture parameters for dual-output LUT based FPGAs[C]. Proceedings of the 24th International Conference on Field Programmable Logic and Applications, Munich, Germany, 2014: 1-6. Betz V, Rose J, and Marquardt A. Architecture and CAD for Deep-Submicron FPGAs [M]. Netherlands, Kluwer Academic Publishers, 1999: 15-20. Gao Hai-xia, Yang Yin-tang, Ma Xiao-hua, et al.. Analysis of the effect of LUT size on FPGA area and delay using theoretical derivations[C]. Proceedings of the 6th International Symposium on Quality of Electronic Design, San Jose, CA, USA, 2005: 370-374. Smith A, Wilton S, and Das J. Wirelength modeling for homogeneous and heterogeneous FPGA architectural development[C]. Proceedings of the 17th ACM/SIGDA international symposium on Field programmable gate arrays, Monterey, CA, USA, 2009: 181-190. Das J, Lam A, Wilton S, et al.. An analytical model relating FPGA architecture to logic density and depth[J]. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, 2011, 19(12): 2229-2242. Smith A, Constantinides G, and Cheung P. FPGA architecture optimization using geometric programming[J]. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2010, 29(8): 1163-1176. Pandurangaiah Y, Reddy V, and Chakravarthy K. Area-delay estimation by concurrent optimization of FPGA architecture parameters using geometric programming[J]. International Journal of Computer Applications, 2013, 82(18): 4-11. Mehri H and Alizadeh B. An analytical dynamic and leakage power model for FPGAs[C]. Proceedings of the 22nd Iranian Conference on Electrical Engineering (ICEE), Tehran, 2014: 300-305. Leow Yoon-kah. Post-routing analytical models for homogeneous FPGA architectures[D]. [Ph. D. dissertation], The University of Arizona, 2013. Yang Li-qun, Yang Hai-gang, Li Wei, et al.. A semi-supervised modeling approach for performance characterization of FPGA architectures[C]. Proceedings of the 24th International Conference on Field Programmable Logic and Applications, Munich, Germany, 2014: 1-6. Wang Y and Witten I H. Inducing model trees for continuous classes[C]. Proceedings of the Ninth European Conference on Machine Learning, Prague, Czech Republic, 1997: 128-137. Brayton R and Mishchenko A. ABC: an academic industrial-strength verification tool[J]. Computer Aided Verification, 2010, 6174: 24-40. Zgheib G, Yang Li-qun, Huang Zhi-hong, et al.. Revisiting and-inverter cones[C]. Proceedings of the 2014 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays, Monterey, CA, USA, 2014: 45-54. -
計量
- 文章訪問數(shù): 1260
- HTML全文瀏覽量: 114
- PDF下載量: 540
- 被引次數(shù): 0