Dynamic server computing resource allocation method

The invention discloses a dynamic server computing resource allocation method and belongs to the technical field of computing resource allocation. A server forms a partition by multiple computing resource nodes, wherein the computing resource nodes refer to one master computing resource node and multiple slave computing resource nodes. The method includes that under the condition that a server system is not shut down, the slave computing resource nodes are dynamically allocated, namely dynamic adding and removing operations are performed; instructions of the dynamic allocation process of the entire slave computing resource nodes are initiated by a management system of the server, dynamic allocation is dynamically logically performed on relevant portions after a South Bridge chip, a BIOS (basic input/output system) and an operation system (OS) receive the instructions, and finally, the slave computing resource nodes are dynamically allocated from hardware; the management system of the server comprises a system management controller and a substrate management controller. Under the condition that the server system is not shut down, adding and removing operations of the slave computing resource nodes can be performed dynamically.