arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2009.04061cs.PFcs.DC
收藏

GPA:一种基于指令采样的GPU性能顾问

GPA: A GPU Performance Advisor Based on Instruction Sampling

Keren Zhou, Xiaozhu Meng, Ryuichi Sai, John Mellor-Crummey

摘要:

由于GPU架构和编程模型的复杂性,开发高效的GPU内核可能很困难。现有的性能工具只在内核级别提供粗粒度的建议(如果有的话)。在本文中,我们介绍了GPA,NVIDIA GPU的性能顾问,它在层次结构中提出了潜在的代码优化机会,包括单独的行,循环和函数。为了减轻用户解释性能计数器和分析瓶颈的负担,GPA使用数据流分析来近似地将测量到的指令停顿归因于其根本原因,并使用有关程序结构和GPU的信息来将低效率模式与优化建议相匹配。为了量化每个建议的潜在好处,我们开发了基于PC采样的性能模型来估计其加速。我们的基准测试和应用程序的实验表明,GPA提供了一个有见地的报告,以指导性能优化。使用GPA,我们在Volta V100 GPU上获得了从1.01 $\times$到3.53 $\times$的加速比,几何平均值为1.22 $\times$。

英文摘要:

Developing efficient GPU kernels can be difficult because of the complexity of GPU architectures and programming models. Existing performance tools only provide coarse-grained suggestions at the kernel level, if any. In this paper, we describe GPA, a performance advisor for NVIDIA GPUs that suggests potential code optimization opportunities at a hierarchy of levels, including individual lines, loops, and functions. To relieve users of the burden of interpreting performance counters and analyzing bottlenecks, GPA uses data flow analysis to approximately attribute measured instruction stalls to their root causes and uses information about a program's structure and the GPU to match inefficiency patterns with suggestions for optimization. To quantify each suggestion's potential benefits, we developed PC sampling-based performance models to estimate its speedup. Our experiments with benchmarks and applications show that GPA provides an insightful report to guide performance optimization. Using GPA, we obtained speedups on a Volta V100 GPU ranging from 1.01$\times$ to 3.53$\times$, with a geometric mean of 1.22$\times$.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑