Artificial intelligence model training requires powerful computing power support, AI platform infrastructure is expensive, platform construction and maintenance is time-consuming and labour-intensive, convenient, readily available, cost-effective model training and debugging platform has become one of the key factors in the development of artificial intelligence.
High performance computing (HPC)
Support GPU fine-grained scheduling: support single-core single card, multi-core multi-card, single node multi-card, multi-node multi-card and other granularity of resource scheduling.