arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2411.02797cs.PFcs.AI
收藏

DeepContext:一种用于深度学习工作负载性能剖析与分析的情境感知、跨平台、跨框架工具

DeepContext: A Context-aware, Cross-platform, and Cross-framework Tool for Performance Profiling and Analysis of Deep Learning Workloads

North Carolina State University(北卡罗来纳州立大学);George Mason University(乔治梅森大学)

Qidong Zhao, Hao Wu, Yuming Hao, Zilingfeng Ye, Jiajia Li, Xu Liu, Keren Zhou

AI总结:

本文提出DeepContext,一种跨平台、跨框架的深度学习性能剖析工具,通过关联程序上下文和自动化分析,帮助用户快速识别并优化性能瓶颈。

摘要:

有效的性能剖析与分析对于优化深度学习模型的训练和推理至关重要,尤其是在异构计算环境日益复杂的背景下。然而,现有工具往往缺乏为CPU和GPU之间复杂交互提供全面程序上下文信息和性能优化见解的能力。本文介绍了DeepContext,一种新型剖析器,它将高级Python代码、深度学习框架、用C/C++编写的底层库以及GPU上执行的设备代码之间的程序上下文关联起来。DeepContext集成了对主流深度学习框架(如PyTorch和JAX)的粗粒度和细粒度性能指标的测量,并兼容Nvidia和AMD的GPU,以及包括x86和ARM在内的各种CPU架构。此外,DeepContext集成了一种新颖的图形用户界面(GUI),使用户能够快速识别热点,以及一种创新的自动化性能分析器,该分析器基于性能指标和程序上下文向用户建议潜在的优化方案。通过详细的使用案例,我们展示了DeepContext如何帮助用户识别和分析性能问题,以实现深度学习工作负载的快速有效优化。我们相信DeepContext是用户在多个计算环境中优化复杂深度学习工作流程的宝贵工具。

英文摘要:

Effective performance profiling and analysis are essential for optimizing training and inference of deep learning models, especially given the growing complexity of heterogeneous computing environments. However, existing tools often lack the capability to provide comprehensive program context information and performance optimization insights for sophisticated interactions between CPUs and GPUs. This paper introduces DeepContext, a novel profiler that links program contexts across high-level Python code, deep learning frameworks, underlying libraries written in C/C++, as well as device code executed on GPUs. DeepContext incorporates measurements of both coarse- and fine-grained performance metrics for major deep learning frameworks, such as PyTorch and JAX, and is compatible with GPUs from both Nvidia and AMD, as well as various CPU architectures, including x86 and ARM. In addition, DeepContext integrates a novel GUI that allows users to quickly identify hotpots and an innovative automated performance analyzer that suggests users with potential optimizations based on performance metrics and program context. Through detailed use cases, we demonstrate how DeepContext can help users identify and analyze performance issues to enable quick and effective optimization of deep learning workloads. We believe Deep Context is a valuable tool for users seeking to optimize complex deep learning workflows across multiple compute environments.

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。

↑