Self-adaptive computer load control

摘要

A self-adaptive computer load control apparatus and method are disclosed for allocating servicing time by a Central Processing Unit (CPU) to several Peripheral Processing Units (PE) under processing overload conditions. Each PE contains registers for storing the value of an expected interval of time between consecutive service periods, for storing the time of commencement of the last service period, and for storing the time of commencement of the present service period. Each PE also includes a comparator for comparing the difference between the times of commencement of the last and present service periods with the value of the expected interval between consecutive service periods. A control word generator is included in each PE for specifying the number of processing steps to be performed by the CPU during the present service period for the respective PE's. The number of processing steps specified has a first magnitude when the comparator determines that the difference between the commencement times is less than the expected time. The number of processing steps specified is made less by the control word generator when the difference in commencement times is greater than the expected time between consecutive service periods.