D³: Differential Testing of Distributed Deep Learning With Model Generation

Deep Learning (DL) techniques have been widely deployed in many application domains. The growth of DL models’ size and complexity demands distributed training of DL models. Since DL training is complex, software implementing distributed DL training is error-prone. Thus, it is crucial to test distributed deep learning software to improve its reliability and quality. To address this issue, we propose a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">differential</i> testing technique—D<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup>, which leverages a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">distributed</i> equivalence rule that we create to test distributed <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">deep</i> learning software. The rationale is that the same model trained with the same model input under different distributed settings should produce equivalent prediction output within certain thresholds. The different output indicates potential bugs in the distributed deep learning software. D<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> automatically generates a diverse set of distributed settings, DL models, and model input to test distributed deep learning software. Our evaluation on two of the most popular DL libraries, i.e., PyTorch and TensorFlow, shows that D<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> detects 21 bugs, including 12 previously unknown bugs.