Computing node processor resource optimization method, computing node and server cluster

A computing node, comprising:

• a memory configured to store an expected computing power in the computer node in terms of a number of standard float point instructions per second by determining a quantity of processors, a number of cores per processor, and a number of threads per processor core in the computing node;
• acquiring an actual processing load of floating point instructions per second for the computer node;
• the processors configured to acquire a relative computing power of the processors in the computing node by seeking a ratio of the actual processing load to the expected computing power in the computing node and to determine a relative computing resource of the computing node based on the relative computing power;
• and
• a communication bus configured to connect the processor with the memory.

A server cluster, comprising a plurality of computing nodes, wherein:

• at each computing node, determining a relative computing resource of the computing node includes:
• acquiring an expected computing power of a computing node in terms of a number of standard float point instructions per second by determining a quantity of processors, a number of cores per processor, and a number of threads per processor core in the computing node;
• acquiring an actual processing load of float points instructions per second for the computer node;
• acquiring a relative computing power of the processor in the computing node by seeking a ratio of the actual processing load to the expected computing power in the computing node;
• and
• based on the relative computing power, determining the relative computing resource of the computing node.