The invention provides a parallel task scheduling
algorithm based on the significance proportion of time and
energy consumption to realize tradeoff optimization of the time and the
energy consumption which are used by a task under a heterogeneous parallel
system when double restrictions of the time and the
energy consumption are met. The
algorithm considers that the task in the actual condition is affected by various factors, the
execution time is not fixed and determined and is approximate to normal distribution, so that when the task scheduling order is determined, the method for utilizing a mean value of the
execution time in the prior art is not adopted, but the mean value and the variance of the
execution time are simultaneously considered to utilize an approximate weight. During a task allocation stage, two performance indexes, namely the time and the energy consumption, are simultaneously considered, and the task is allocated to the proper processor according to the significance proportion of the time and the energy consumption, so that the tradeoff optimization of the time and the energy consumption according to the
specific gravity is realized, and thus the
system obtains higher weighting performance. The
algorithm is combined with a DVFS (Dynamic
Voltage and
Frequency Scaling) technology and is also suitable for an isomorphic parallel
system and a computer system without
embedment of the DVFS technology.