热效应布局下的缓冲器插入时序优化方法

王新胜; 韩良; 喻明艳

doi:10.13700/j.bh.1001-5965.2014.0809

热效应布局下的缓冲器插入时序优化方法

doi: 10.13700/j.bh.1001-5965.2014.0809

1.
哈尔滨工业大学信息与电气工程学院, 威海 264209
2. 浙江大学宁波理工学院, 宁波 315100

基金项目: 国家自然科学基金(61201307)

详细信息

作者简介:
王新胜(1978-),男,山东威海人,讲师,xswang@hit.edu.cn

通讯作者:
韩良(1969-),男,吉林榆树人,副教授,hanliang@hit.edu.cn,主要研究方向为大规模集成电路设计.

中图分类号: TN405.97;TN402
计量
- 文章访问数: 874
- HTML全文浏览量: 155
- PDF下载量: 476
- 被引次数: 0
出版历程
- 收稿日期: 2014-12-23
- 修回日期: 2015-02-12
- 网络出版日期: 2015-10-20

Thermal aware floor planning timing optimal method for buffer insertion

1.
Shool of Information and Electrical Engineering, Harbin Institute of Technology, Weihai 264209, China
2. Ningbo Institute of Technology, Zhejiang University, Ningbo 315100, China

摘要

摘要: 随着集成电路的集成度越来越高,芯片的发热量越来越大且其内部温度呈不均匀分布,这会影响关键路径的传播延时,进而影响基于缓冲器插入的关键路径性能.提出了一种考虑芯片热效应布局优化的缓冲器插入时序优化方法,在版图设计的早期估计芯片的热分布和温度分布并且把其应用到版图布局优化和RC延时模型中.同时利用模拟退火算法基于热分布调整并优化布局,最后在最优布局下利用提出的缓冲器插入模型和快速插入算法进行时序优化.仿真结果表明相对于不考虑温度效应布局优化的缓冲器插入方法,缓冲器插入延时优化方法能有效降低最坏延时和缓冲器插入数目,最坏延时比传统方法降低9%~18%,比文献已经提出的最好方法降低5%~7%,缓冲器插入数比其少10~20个.
- 大规模集成电路 /
- 布局规划 /
- 缓冲器插入 /
- 互连线 /
- 模拟退火算法
Abstract: With the integration degree of integrated circuit (IC) is increasingly high, the heat on a chip is also growing, which leads to an uneven temperature distribution intra-die and affects the propagation delay of the critical path thereby affecting the performance of buffer insertion path. A buffer insertion timing optimization method which considered the heat distribution condition optimization floor-planning was proposed. It estimates the temperature and heat distribution of chip in the early stages of layout design and is applied to layout optimization floor-planning. The thermal aware floor planning based on simulated annealing algorithm was used to adjust and optimize planning, and then we made an optimization for timing by proposed buffer insertion model and fast buffer insertion algorithm. Simulation results show that the use of the proposed buffer insertion delay optimization method can effectively reduce the worst delay and the number of buffer insertion, worst delay is 9%-18% lower than traditional methods, 5%-7% lower than the best method shown in reference, and the insertion buffer numbers are 10 to 20 less than its.
- very large scale integrated circuit /
- floor planning /
- buffer insertion /
- interconnect /
- simulated annealing algorithm

HTML全文

参考文献(17)

[1]	Tsai C H,Kang S M.Cell-level placement for improving substrate thermal distribution[J].IEEE Transactions on Computer-Aided Design for Integrated Circuits and Systems,2000,19(2):253-266.
[2]	Ajami A H,Banerjee K,Pedram M.Modeling and analysis of non-uniform substrate temperature effects on global ULSI interconnects[J].IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems,2005,24(6):849-861.
[3]	Ajami A H,Banerjee K,Pedram M.Analysis of substrate thermal gradient effects on optimal buffer insertion[C]∥IEEE ACM International Conference on ICCAD.Piscataway,NJ:IEEE Press,2001:44-48.
[4]	Huang W,Ghosh S,Velusamy S,et al.HotSpot:A compact thermal modeling methodology for early-stage VLSI design[J].IEEE Transactions on Very Large Scale Integration Systems,2006,14(5):501-513.
[5]	Sankaranarayanan K,Velusamy S,Stan M R,et al.A case for thermal-aware floorplanning at the microarchitectural level[J].Journal of Instruction-Level Parallelism,2005(8):1-16.
[6]	Saxena P,Menezes N,Cocchini P,et al.Repeater scaling and its impact on CAD[J].IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems,2004,23(4):451-463.
[7]	Kim M,Ahn B G,Kim J,et al.Thermal aware timing budget for buffer insertion in early stage of physical design[C]∥IEEE International Symposium on Circuit and Systems.Piscataway,NJ:IEEE Press,2012:357-360.
[8]	Kirkpatrick S.Optimization by simulated annealing:Quantitative studies[J].Journal of Statistical Physics,1983,34(5-6):975-986.
[9]	Kevin S.HotSpot[EB/OL].2014[2015-03-22].http:∥lava.cs.virginia.edu/hotspot.
[10]	Kevin S.HotSpot[EB/OL].2014[2015-03-22].http:∥lava.cs.virginia.edu/HotSpot/HotSpotHowTo.htm.
[11]	van Ginneken L P P P.Buffer placement in distributed RC-tree networks for minmal elmore delay[C]∥IEEE International Symposium on Circuits and Systems.Piscataway,NJ:IEEE Press,1990:865-868.
[12]	Gupta R,Tutuianu B,Pileggi L T.The Elmore delay as a bound for RC trees with generalized input signals[J].IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems,1997,16(1):95-104.
[13]	Abou A I,Nowak B,Chu C.Fitted Elmore delay:A simple and accurate interconnect delay model[J].IEEE Transactions on Very Large Scale Integration System.Piscataway,NJ:IEEE Press,2004,12(7):691-696.
[14]	Athikulwongse K,Zhao X,Lim S K.Buffered clock tree sizing for skew minimization under power and thermal budgets[C]∥Proceedings of the 2010 Asia and South Pacific Design Automation Conference.Piscataway,NJ:IEEE Press,2010:474-479.
[15]	Skadron K,Stan M,Velusamy S,et al.A case for thermal aware floor planning at the microarchitectural level[J].Journal of Instruction-level Parallelism,2005,7:1-16.
[16]	Patrick H M,Ameya R A.GSRC bookshelf benchmarks[EB/OL].2001(2007-06-01)[2015-03-22].http:∥vlsicad.cs.binghamton.edu/benchgsrc.html.
[17]	Yu C.Predictive technology model[EB/OL].Arizona:[s.n.],2005(2012-06-01)[2015-03-22].http:∥ptm.asu.edu/.