A load balancing cluster rendering method based on dynamic task granularity

Abstract

The invention discloses a load balancing cluster rendering method based on dynamic task granularity, which belongs to the technical field of computer cluster rendering, and solves the problems of obvious load imbalance of a coarse-grained inter-frame scheduling strategy when the cluster scale is increased and increase of execution expenditure and reduction of the overall performance of operation of a fine-grained intra-frame scheduling strategy. The method comprises the following steps of firstly, determining the number of frames for inter-frame scheduling according to the number of renderingnodes; then proposing a staggered task division method by utilizing inter-frame correlation to distribute frames, and finally splitting the rest frames into a plurality of blocks with smaller granularity to realize load balancing, so that the resource utilization rate is improved, the workload between rendering nodes is effectively balanced, and the overall performance of the cluster rendering system is improved.

Description

technical field [0001] The invention belongs to the technical field of computer cluster rendering, and relates to a load balancing method based on dynamic task granularity in a cluster rendering management system. Background technique [0002] In recent years, the animation film and television industry has developed rapidly. With the rise of 3D and 4D and the trend of high-definition animation, realistic rendering, one of the core links of animation production, is also facing great challenges. The single-frame image of some complex shots Renderings can take hours or even days to complete. Although the computing power of computers has doubled in the past few decades, the increasingly stringent requirements for the realism and resolution of animation movies have resulted in no decline in rendering time. In order to complete tasks on time, many large animation studios will spend a lot of money to build their own local rendering farms or use cloud rendering farms, which use the...

AI Technical Summary

Problems solved by technology

[0007] In order to overcome the shortcomings of the application of the above technologies on the cluster rendering management platform, the present invention solves the problem that the coarse-grained inter-frame scheduling strategy will cause obvious load imbalance when the cluster size increases, and the fine-grained intra-frame scheduling strategy will increase execution overhead and reduce Problems with the overall performance of the job

Images