基于热例程的动态二进制翻译优化

依据对系统级程序行为特性的观察,提出了一种基于热例程的动态二进制翻译优化方法。该方法以频繁执行的例程作为优化单位,通过块内和块间优化算法消除动态二进制翻译引入的冗余。相比基于踪迹的优化方法,该方法具有优化单位发现开销更小、代码区域更大、无重复翻译等优点,更适用于系统虚拟机中操作系统代码的优化。在跨平台系统虚拟机监控器ARCH-BRIDGE上的测试表明,通过对内核代码实施该优化方法,SPEC CPUINT 2006程序的效率提升了3.5%~14.4%,相比基于踪迹的优化,性能最大提升了5.1%。

Hot-routine Based Optimization of Dynamic Binary Translation

According to observation of the behavior of system level program,the paper provided a hot-routine based optimization method of dynamic binary translation,which takes frequently executed routines as optimization unit,uses intra-block and inter-block optimization algorithm to remove redundancies introduced by dynamic binary translation.Compared with the trace based optimization,this method has the advantages of less optimization unit discovery overhead,bigger code region,no duplicated translation,and is more suitable for the optimization of OS code in the virtual machine.Evaluation on the cross-platform virtual machine monitor ARCH-BRIDGE demonstrates that,by applying the optimization method to kernel code,performance of SPEC CPUINT 2006 programs gets a speedup of 3.5%~14.4%,and is 5.1% faster than the trace based optimization at most.