Scaling Up: The Validation of Empirically Derived Scheduling Rules on NVIDIA GPUs*
Scaling Up: The Validation of Empirically Derived Scheduling Rules on NVIDIA GPUs*
复制标题
DOI：
--
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Joshua Bakita;Nathan Otterness;James H. Anderson;F. Donelson
中科院分区：
文献类型：
--
作者：
Joshua Bakita;Nathan Otterness;James H. Anderson;F. Donelson
— 使用图形处理单元 (GPU) 增强的嵌入式系统在自动驾驶汽车等安全关键型实时系统中的使用越来越多。目前这些 GPU 的黑匣子和专有性质使得很难确定它们在最坏情况下的行为，从而威胁到自主系统的安全。在这项工作中，我们引入了一种新的自动验证框架来分析 GPU 执行轨迹，并确定从黑盒实验推断出的行为假设是否与现实世界设备的行为一致。我们发现，在之前的工作中观察到的行为在小范围内是一致的，但这些规则并没有扩展到明显较旧的 GPU 上，并且无法应对复杂的 GPU 工作负载。
—Embedded systems augmented with graphics processing units (GPUs) are seeing increased use in safety-critical real-time systems such as autonomous vehicles. The current black-box and proprietary nature of these GPUs has made it difficult to determine their behavior in worst-case scenarios, threatening the safety of autonomous systems. In this work, we introduce a new automated validation framework to analyze GPU execution traces and determine if behavioral assumptions inferred from black-box experiments consistently match behavior of real-world devices. We find that the behaviors observed in prior work are consistent on a small scale, but the rules do not stretch to significantly older GPUs and struggle with complex GPU workloads.