A TASK DISTRIBUTION METHOD USING PARALLEL PROCESSING ALGORITHM

摘要

The present invention relates to a task distribution method using a parallel processing algorithm and, more specifically, to a task distribution method using a parallel processing algorithm, comprising the steps of: (1) configuring an execution scenario for n tasks to be processed in a heterogeneous computing platform environment as an implementation graph; (2) obtaining an execution time measured by performing each of the n tasks respectively on a device CPU and a device GPU; and (3) determining a device CPU or a device GPU suitable for each of the n tasks as a corresponding processor. According to the present invention, optimal scheduling can be provided through task distribution.