CPU Load Reduction by Re-Computation Check Based on Input Parameters
Original Publication Date: 2004-Jan-25
Included in the Prior Art Database: 2004-Jan-25
Related PeopleOther Related People:
Complex computations consume a lot of CPU (central processing unit) load. In case of a periodical recurrence the overall CPU load can be reduced, if computations are only performed if mandatory. A computation is classified as mandatory, if the result is different to the last recurrence. The result of a computation can only differ from the result of the last recurrence, if the input parameters have been changed. Therefore at the beginning of each complex computation it has to be checked, whether a (re-)computation is really mandatory. This check is based on the related input parameters. In case of a scalar input, the actual input parameter is compared with the stored one of the last recurrence. In case of several input data a checksum or CRC (cyclic redundancy code) can be helpful to determine a change. Those input checks assume integer input. For some computation an enhanced check algorithm might be helpful. For example, a defined interval has to be given. This interval is based on the value of the parameter at last recurrence. If the new value of the input parameter is within the defined interval, the input is treated as unchanged.