GPU memory errors are a critical threat to deep learning (DL) frameworks, leading to crashes or even security issues. We introduce GPU-Fuzz, a fuzzer locating these issues efficiently by modeling operator parameters as formal constraints. GPU-Fuzz utilizes a constraint solver to generate test cases that systematically probe error-prone boundary conditions in GPU kernels. Applied to PyTorch, TensorFlow, and PaddlePaddle, we uncovered 13 unknown bugs, demonstrating the effectiveness of GPU-Fuzz in finding memory errors.

翻译：GPU内存错误是深度学习框架面临的严重威胁，可能导致系统崩溃甚至引发安全问题。本文提出GPU-Fuzz——一种通过将算子参数建模为形式化约束来高效定位此类问题的模糊测试工具。该工具利用约束求解器生成测试用例，系统性地探测GPU内核中易出错的边界条件。在PyTorch、TensorFlow和PaddlePaddle框架上的实验表明，GPU-Fuzz成功检测出13个未知漏洞，验证了其在内存错误发现方面的有效性。