基于垂直扩散的FPGA温度优化布局算法

发布时间：2018-06-23 12:47

本文选题：可编程门阵列 + 布局算法　；参考：《计算机辅助设计与图形学学报》2017年01期

【摘要】：为减小FPGA热梯度的增加对芯片性能和可靠性的影响,提出一种基于垂直扩散的温度优化布局算法.首先,通过实验分析了芯片温度特性与芯片尺寸之间的关系;然后,根据布局后网表计算出过热区域,利用该区域边界及扩散系数构建温度优化布局算法,并引入局部位置调整机制解决逻辑块位置重叠问题.实验结果表明,与传统布局算法的芯片温度相比,在线长和延时平均仅增加3.4%和1.4%的情况下,该算法的峰值温度平均减小7.5%,热梯度平均减小20.3%.
[Abstract]:In order to reduce the effect of FPGA thermal gradient on chip performance and reliability, a temperature optimal layout algorithm based on vertical diffusion is proposed. Firstly, the relationship between chip temperature characteristics and chip size is analyzed by experiments, and then the superheated region is calculated according to the layout grid table, and the temperature optimization layout algorithm is constructed by using the boundary of the region and diffusion coefficient. The local position adjustment mechanism is introduced to solve the problem of logical block location overlap. The experimental results show that compared with the chip temperature of the traditional layout algorithm, the peak temperature of the algorithm decreases by 7.5 and the thermal gradient decreases by 20.3in the case of an average increase of 3.4% and 1.4% in on-line length and delay.
【作者单位】：中国科学院电子学研究所可编程芯片与系统研究室;中国科学院大学;
【基金】：国家自然科学基金(61271149,61204045,61404140)
【分类号】：TN791

【相似文献】