Reconfigurable LSI Systems for Statistical Genetic Algorithms

本研究的目的是开发一个可重构的LSI系统和LSI CAD（计算机辅助设计）工具的统计遗传算法。提出了低功耗高级综合中的线程划分算法、具有异构数据路径的专用处理器的协同综合算法、SIMD处理器核的指令集和功能单元综合、基于FPGA的可重构自适应FEC、基于线程划分的高级功耗优化、基于多扫描设计的混合字典测试数据压缩、通过用于测试数据压缩和扫描功率降低的游程编码的选择性扫描链重新配置，用于可靠无线通信的可重新配置的自适应FEC系统，基于线程划分的高级能量优化的实验评估，基于单体型和双体型配置的全基因组病例对照关联研究中多重比较的新校正，基于IP的SoC设计中的处理器核心综合系统，SIMD处理器综合中的子操作并行优化及其实验评估，一种SIMD指令集和具有SIMD操作分解的功能单元综合算法，一种基于IP的SoC设计中具有可配置处理器核的接口电路综合方法，FCSCAN：用于降低测试成本的有效的基于多扫描的测试压缩技术，嵌入基于字的蒙哥马利乘法器的快速椭圆曲线密码系统LSI，目前，用于统计遗传算法的可重构大规模集成电路系统还没有开发出来，但本研究已经获得了足够的开发技术。

The aim of this research is to develop a reconfigurable LSI system and LSI CAD (Computer-Aided Design) tools for statistical genetic algorithms. We have proposed a thread partitioning algorithm in low power high-level synthesis, a cosynthesis algorithm for applicaton specific processors with heterogeneous datapaths, instruction set and functional unit synthesis for SIMD processor cores, FPGA-based reconfigurable adaptive FEC, high-level power optimization based on thread partitioning, a hybrid dictionary test data compression for multiscan-based designs, a selective scan chain reconfiguration through run-length coding for test data compression and scan power reduction, a reconfigurable adaptive FEC system for reliable wireless communications, experimental evaluation of high-level energy optimization based on thread partitioning, a new correction for multiple comparisons in genome-wide case-control association studies based on haplotypes and diplotype configurations, a processor core synthesis system in IP-based SoC design, sub-operation parallelism optimization in SIMD processor synthesis and its experimental evaluations, A SIMD instruction set and functional unit synthesis algorithm with SIMD operation decomposition, an interface-circuit synthesis method with configurable processor core in IP-based SoC designs, FCSCAN : an efficient multiscan-based test compression technique for test cost reduction, a fast elliptic curve cryptosystem LSI embedding word-based Montgomery multiplier, etc.Reconfigurable LSI systems for statistical genetic algorithms have not been developed yet, but enough technics to develop them areobtained in this research.