构建标准化执行轨迹，推动AI系统软硬件协同设计与性能评测
MLCommons Chakra: Advancing Performance Benchmarking and Co-design using Standardized Execution Traces
- 用图结构统一表示分布式AI任务的计算、通信、内存等行为
- 已在真实生产集群采集轨迹，支持多场景分析与验证
- 被多家科技巨头采用，适合系统优化与架构设计研究者
人工智能创新速度迅猛，亟需一种敏捷方法来观测、复现和优化生产环境中分布式机器学习工作负载的行为，并支持未来系统的软硬件协同设计。本文提出Chakra——一个开放可移植的性能基准与协同设计生态系统。其核心是基于图结构的分布式AI/ML工作负载标准化表示，称为Chakra执行轨迹（ET），可刻画计算、内存、通信等关键操作，以及数据与控制依赖、时间信息和资源约束。Chakra还配套提供采集、分析、生成和应用这些轨迹的工具链，支持各类模拟器、仿真器和回放工具。我们对生产级AI集群采集的Chakra ET进行了分析，并通过实际案例展示了其价值。该系统已被MLCommons采纳，产业界包括NVIDIA、AMD、Meta、Keysight、HPE、Scala等多家企业积极参与贡献与应用。
原文摘要 · Abstract (English)
The fast pace of artificial intelligence~(AI) innovation demands an agile methodology for observation, reproduction and optimization of distributed machine learning~(ML) workload behavior in production AI systems and enables efficient software-hardware~(SW-HW) co-design for future systems. We present Chakra, an open and portable ecosystem for performance benchmarking and co-design. The core component of Chakra is an open and interoperable graph-based representation of distributed AI/ML workloads, called Chakra execution trace~(ET). These ETs represent key operations, such as compute, memory, and communication, data and control dependencies, timing, and resource constraints. Additionally, Chakra includes a complementary set of tools and capabilities to enable the collection, analysis, generation, and adoption of Chakra ETs by a broad range of simulators, emulators, and replay tools. We present analysis of Chakra ETs collected on production AI clusters and demonstrate value via real-world case studies. Chakra has been adopted by MLCommons and has active contributions and engagement across the industry, including but not limited to NVIDIA, AMD, Meta, Keysight, HPE, and Scala, to name a few.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。