Study of combinatorial and mathematical programming methods for high level synthesis

2016 Fiscal Year Final Research Report

• Project/Area Number: 26420323
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Electron device/Electronic equipment
• Research Institution: Waseda University
• Principal Investigator: Yoshimura Takeshi 早稲田大学, 理工学術院(情報生産システム研究科・センター), 教授 (80367177)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: 高位レベル合成 / 低消費電力化 / TSV / スケジューリング / 最適化

Outline of Final Research Achievements

The research on high level synthesis of system LSI were conducted. First, in the scheduling problem, a method based on a combination of mathematical programming and graph theory was proposed for the dynamic power optimization problem. For leakage power optimization　problem, a method with the modifications of the above method and additional post-processing were proposed. In both problems, optimal solutions were obtained in most cases. In the port assignment problem, a method to avoid local optimal solutions considering sub-solution space and a method to reduce processing time were proposed. Optimal solutions were obtained for all the evaluation data. In the TSV assignment problem for three-dimensional LSI, a method to reduce the scale of the problem without sacrificing the solution quality was proposed based on a hierarchical design method. The CPU time was reduced to about 1/38 of the conventional methods.

Free Research Field

工学