| 摘要 |
The present invention discloses an apparatus for controlling a central processing unit (CPU) utilization, which includes a first acquisition unit, a second acquisition unit, a selection unit, and a control unit, and a method for the same. The first acquisition unit acquires the utilization of each core based on Bayesian learning and Gaussian process regression. The second acquisition unit acquires the average standby time and the blocking rate of each core corresponding to the acquired utilization. The selection unit selects the target processing of a source core, and a destination core based on the average standby time and the blocking rate, which are acquired. The control unit performs a control operation to extract the target processing of the selected source ore so that the selected source core becomes inactivated, and to migrate the extracted target processing to the selected destination core. Accordingly, the utilization of each core in a large-scale system can be controlled based on Bayesian learning and Gaussian process regression. |
