Research on task-level multi-thread scheduling for coarse-grained multi-core systems

Nowadays Multi-core system is the main direction of processor development. How to reasonably perform task scheduling and ensure that all processing cores are in effective working state is an important direction of multi-core system research. The key difficulty of multi-core task scheduling is to explore the parallelism of tasks. To solve this problem, this paper draws on the instruction-level multi-threading idea and combines the coarse-grained characteristics of tasks in multi-core systems to propose a new coarse-grained multi-threaded multi-core architecture. A multi-threaded fetching strategy, resource allocation strategy and thread switching mechanism are established, and the multi-threaded scheduler circuit design of this structure is completed. A coarse-grained multi-core computing platform is built around this scheduler, and implemented on the FPGA. The results show that the design scheme improves the task parallelism of the multi-core computing platform by about 34.29% on average compared with single-threaded.