中文摘要：

处理器发展已进入多核时代．现有并行仿真内核常常以多进程方式使用多核资源，存在较大的同步和通信开销，无法深入发掘多核处理器潜能．基于层次化并行仿真内核（HPSK）模型，重点对时间管理服务和事件管理服务进行优化，支持多线程架构下进行高效能仿真：（1）基于混合时间推进模式，提出最小发送时戳（EETS）计算协议，可根据仿真应用特点灵活配置为异步EETS算法以支持高效的全局同步，并证明了计算协议的正确性；（2）基于并行仿真事件交互的特点，提出无锁创建、异步提交和指针通信的事件管理算法，最小化线程之间的锁开销并减少了内存的消耗．实验结果表明，采用上述优化服务的HPSK能够在各种条件下获得很好的加速效果．

英文摘要：

The development of CPU has made its way into the era of multicore. The current parallel simulation kernel utilizes multicore resource by a multiprocess, which leads to inefficiency in synchronization and communication. This study has optimized two services based on hierarchical parallel simulation kernel （HPSK） model to support high performance simulation in multithread paradigm. First, the paper proposes a protocol of EETS computation based on hybrid time management, which can be configured flexibly as asynchronous EETS algorithm according to application＇s characteristics. Second, the study proposes an event management algorithm based on characteristics of events interaction, which can create events lock-free, commits events asynchronously, and transfers events based on pointers, to eliminate the overhead of locks and to reduce the usage of memory. Experimental results from phold show that the optimized HPSK works well on different conditions.