We present an evaluation method for estimating the lower bound number of Monte Carlo STA trials required to obtain at least one sample which falls within top-k of its parent population. The sample can be used to ensure that target designs are timing-error free with a predefined probability using the minimum computational cost. The lower bound number is represented as a closed-form formula which is general enough to be applied to other verifications. For validation, Monte Carlo STA was carried out on various benchmark data including ISCAS circuits. The minimum number of Monte Carlo runs determined using the proposed method successfully extracted one or more top-k delay instances.
展开▼
机译:我们提出了一种评估方法,用于估计获得至少一个样本所需的蒙特卡洛 STA 试验的下限数量,该样本属于其亲本群体的前 k %。该样本可用于确保目标设计无时序误差,并具有预定义的概率,使用最小的计算成本。下限数表示为一个封闭式公式,该公式足够通用,可以应用于其他验证。为了进行验证,蒙特卡洛STA在包括ISCAS电路在内的各种基准数据上进行了验证。使用所提出的方法确定的最小蒙特卡罗运行数成功提取了一个或多个top-k %延迟实例。
展开▼