Molecular dynamics calculation chip design method and system based on neural network

The invention relates to the field of artificial intelligence and integrated circuit design, and discloses a molecular dynamics calculation chip design method and system based on a neural network. The method comprises the following steps: firstly, generating a system feature data set required for training a deep neural network; then training a deep neural network model to precisely fit inter-atomic stress; the trained model is deployed in a chip of a non-von Noiemann architecture, so that the precision of predicting the atomic stress can be ensured, and the molecular dynamics calculation is accelerated; then, a molecular dynamics calculation system is built; molecular dynamics calculation is carried out, the atomic trajectory of each step is stored, other macroscopic properties of the system are calculated by using atomic trajectory information, and the method has the characteristics of high calculation precision, high calculation speed and the like. 本发明涉及人工智能和集成电路设计领域，公开了一种基于神经网络的分子动力学计算芯片设计方法及系统。本发明首先生成训练深度