Performance management of subsystems in a server by effective usage of resources

摘要

An approach is provided in which a subsystem cooling manager detects an increased workload indicator corresponding to a computer subsystem's forthcoming workload requirement. The forthcoming workload requirement corresponds to future computing resources required by the subsystem to support one or more software programs executing on the computer system. The subsystem cooling manager determines that the forthcoming workload requirement exceeds a utilization threshold and in turn, directs one or more cooling systems towards the corresponding subsystem according.

主权项

1. A method comprising:
computing, by one or more processors, a queue utilization percentage in response to detecting that one or more instructions are loaded into an instruction queue corresponding to a first subsystem in a computer system, wherein the queue utilization percentage identifies a utilization of capacity of the instruction queue prior to the first subsystem commencing execution of the one or more instructions, and wherein the instruction queue comprises at least the one or more instructions awaiting execution from the first subsystem; determining, by one or more of the processors, that the queue utilization percentage exceeds a utilization threshold for a predetermined amount of time; and directing, by the one or more processors, one or more cooling systems from a second subsystem included in the computer system to the first subsystem in response to the determination.