工业场景AI及算力研究

Research on AI and Computing Power in Industrial Scenarios

近年来AI飞速发展,深度学习、大语言模型等的出现加速了AI的广泛应用。在工业领域,AI也开始广泛应用在物料识别、设备诊断、控制量预测、任务调度等场景,其对算力的要求不断提升。GPU是AI训练和推理的关键设备,投入较大;在当前西方发达国家对我国高端芯片进行封锁的情况下,如何选择合适的AI算力对于确保计算性能、控制投资至关重要。结合AI自身特点和工业场景,对AI算力的选择提出了相关建议。

With the rapid development of AI in recent years,the emergence of deep learning and large language models has accelerated the wide application of AI.In industry,AI is widely used in material identification,equipment diagnosis,control quantity prediction,task scheduling and other scenarios,and thus its requirements for computing power is demanding.GPU is the key equipment for AI training and reasoning,and the investment is heavy.In the current situation that Western developed countries are blocking China’s high-end chips,how to choose the right AI computing power is crucial to ensure computing performance and control budget.Based on the characteristics of AI and industrial scenarios,this paper puts forward some suggestions on the selection of AI computing power.