Analysis of Overall Assignment and Sorting of Tasks in Heterogeneous Computing Systems Based on Mathematical Programming Algorithms

The problem of assignment and sequencing of tasks is a very complex problem, which is related to whether the computer system can effectively exert the overall efficiency. Solving this problem can make the lowest cost and obtain the greatest benefit. However, the current algorithms for coordinating job assignment and sorting are not completely suitable for heterogeneous computing systems. In order to rationally arrange the problem of computer assignment and sorting, this paper proposes a mathematical programming algorithm to effectively solve the inadaptability of assignment and sorting to heterogeneous computing systems. This paper adopts the control variable method and the comparative analysis method, selects the mathematical programming algorithm and the genetic algorithm, the simulated annealing algorithm these two algorithms, selects the relevant performance indicators, designs the experiments to perform calculations and collects the data. Through the comparison of different algorithms in heterogeneous computing systems, it can be seen that in terms of performance, the average response time and node utilization of the three algorithms are not much different, but the availability of the mathematical programming algorithm is significantly higher than that of the other two. When the rate is 1.0, it still has an availability of 0.59. With the increase in the number of tasks and CPU utilization, the advantages of the mathematical programming algorithm are gradually becoming obvious. Although the receiving capabilities of the three algorithms are decreasing with the increase of these two indicators, when the number of tasks reaches 140, the mathematical programming algorithm can receive tasks remains at 78%, indicating that the algorithm is stable. By applying heterogeneous computing systems on different platforms, GPU and FPGA each have their own advantages. The purpose of coordinating assignments and sequencing is to better allocate resources in the future and maximize benefits. Through the study of mathematical programming algorithms, the time required to execute programs in heterogeneous computing systems can be better reduced, thereby improving the overall system effectiveness.