All terms
Hardware & Systems
Cluster
A group of networked machines that work together as a single computing system.
Definition
A cluster is a collection of connected servers that coordinate to act as one larger system, pooling their processors, memory, and accelerators. In AI, clusters of GPU-equipped compute nodes are linked by high-speed networks so that training and inference can be spread across many machines at once. The size of a cluster and the quality of its interconnect set the ceiling on how large a model it can train or serve.