论文部分内容阅读
在多核处理器系统中，多个计算任务映射到多核处理器内核的方式对于系统吞吐率至关重要。针对此问题提出一种新的多应用任务到多核的映射算法，该算法在应用到来之前预测应用的相关性能，并采用分支限界法提前为未来应用预留合适的内核几何位置。当应用真正到来时，根据预留的区域完成映射。实验结果表明，该算法相比其他传统算法，在多任务通信量的减少和多核系统的吞吐率等方面都收到了良好效果。
In multi-core processor systems, the way multiple compute tasks map to multi-core processor cores is critical to system throughput. A new multi-application task-to-multi-core mapping algorithm is proposed to solve this problem. The algorithm predicts the application performance before it arrives, and uses the branch and bound method to reserve the appropriate kernel geometry position for future application in advance. When the app arrives, the mapping is done according to the reserved area. Experimental results show that the proposed algorithm achieves better results than other traditional algorithms in reducing the amount of multitasking traffic and throughput of multi-core systems.