Reducing the burden of parallel loop schedulers for many-core processors
Reducing the burden of parallel loop schedulers for many-core processors
复制标题
减轻多核处理器并行循环调度程序的负担
DOI：
10.1145/3178487.3178517
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Arif M
中科院分区：
文献类型：
--
作者：
Arif M
本文提出了一种低开销的半屏障模式来调度细粒度并行循环，并考虑了它在Intel OpenMP和Cilkplus调度器中的集成。实验评估表明，该技术的调度开销比Intel OpenMP低43%，比Cilk低12.1x。我们在48个线程上观察到22%的加速比，峰值加速比为2.8倍。
This work proposes a low-overhead half-barrier pattern to schedule fine-grain parallel loops and considers its integration in the Intel OpenMP and Cilkplus schedulers. Experimental evaluation demonstrates that the scheduling overhead of our techniques is 43% lower than Intel OpenMP and 12.1x lower than Cilk. We observe 22% speedup on 48 threads, with a peak of 2.8x speedup.