CSR---SMA: Validating Architectural Simulators Using Non-Linear Dynamics Techniques

该项目致力于通过两种方式改进最先进的架构模拟器：(i)使用户能够确定架构模拟器是否正确地模拟了真实硬件系统的关键方面；（ii）通过帮助模拟器实现者识别其模拟器的弱点。这个问题很关键，因为今天很多建筑研究都使用模拟；如果模拟器不准确，那么模拟器的结果可能会产生误导。先前的工作通过使用聚合指标来验证模拟器（例如，模拟器和实际硬件是否以相似的总周期数执行程序？）。这种验证不考虑模拟器和硬件是否表现出类似的时变行为。这个项目的一个关键创新是它包含了非线性动力学领域的技术。由于两个原因，这些技术非常适合于计算机系统。首先，计算机系统在所有层面上都是非线性的（例如，缓存丢失的成本不是固定的，而是取决于许多因素），因此线性技术的使用将是有限的。其次，计算机系统的许多方面是未知的（例如，硬件制造商可能不会透露硬件的完整规格）或不可测量的（例如，可能无法测量硬件深处特定事件发生的频率）。非线性动力学技术就是专门用来处理这种信息缺失的。

This project strives to improve the state-of-the-art in architecturalsimulators in two ways: (i) by enabling users to determine if an architecturalsimulator correctly models the key aspects of a real hardware system; and (ii)by helping simulator implementors to identify weaknesses in their simulators.This problem is critical because much architectural research today usessimulations; if the simulators are not accurate then the results from thesimulator may be misleading. Prior work validates simulators by usingaggregate metrics (e.g., do the simulator and the real hardware execute aprogram in a similar number of total cycles?). This validation does notconsider whether or not the simulator and the hardware exhibit similartime-varying behavior.A key innovation of this project is that it embraces techniques from the fieldof non-linear dynamics. These techniques are well suited for computer systemsfor two reasons. First, computer systems are non-linear at all levels (e.g.,the cost of a cache miss is not fixed and instead depends on many factors) andthus linear techniques will be of only limited use. Second, many aspects ofcomputer systems are unknown (e.g., the hardware manufacturer may not revealthe full specifications of the hardware) or unmeasurable (e.g., it may not bepossible to measure how often a particular event occurs deep down in thehardware). Non-linear dynamics techniques are specifically designed to dealwith such missing information.