Search

As modern supercomputers evolve to exascale, their I/O subsystems are becoming increasingly complex, making optimization of I/O for scientific applications a daunting task. Although I/O profiling tools facilitate the process of optimizing application I/O performance, legacy profiling toolslack flexibility in their level of detail and ability to correlate traces with other sources of data. Additionally, a lack of robust trace analysis tools makes it difficult to derive actionable insights from large-scale I/O traces.Darshan is an HPC I/O characterization tool that records statistics in a lightweight manner that makes it appropriate for full-time production deployment. However, Darshan’s default characterization mechanism records information at a fixed granularity. We augment Darshan by proposing Darshan eXtended Tracing (DXT) for more detailed profiling of I/O software stacks. DXT enables users and administrators to vary the level of fidelity captured by Darshan at run time without modifying or recompiling applications. This capability facilitates systematic analysis on the I/O behavior of applications and can provide useful application kernel I/O traces to help advance parallel I/O research. We have demonstrated the power of DXT by obtaining a wide range of useful statistics for multiple case studies, and we further show that DXT isable to do so same with negligible overhead.

Modern scientific applications utilize numerous software and hardware layers to efficiently access data. This approach poses a challenge for I/O optimization because of the need to instrument and correlate information across those layers. The Darshan characterization tool seeks to address this challenge by providing efficient, transparent, and compact runtime instrumentation of many common I/O interfaces. It also includes command-line tools to generate actionable insights and summary reports. However, the extreme diversity of today's scientific applications means that not all applications are well served by one-size-fits-all analysis tools. In this work we present PyDarshan, a Python-based library that enables agile analysis of I/O performance data. PyDarshan caters to both novice and advanced users by offering ready-to-use HTML reports as well as a rich collection of APIs to facilitate custom analyses. We present the design of PyDarshan and demonstrate its effectiveness in four diverse real-world analysis use cases.

As modern supercomputers evolve to exascale, their I/O subsystems are becoming increasingly complex, making optimization of I/O for scientific applications a daunting task. Although I/O profiling tools facilitate the process of optimizing application I/O performance, legacy profiling toolslack flexibility in their level of detail and ability to correlate traces with other sources of data. Additionally, a lack of robust trace analysis tools makes it difficult to derive actionable insights from large-scale I/O traces.Darshan is an HPC I/O characterization tool that records statistics in a lightweight manner that makes it appropriate for full-time production deployment. However, Darshan’s default characterization mechanism records information at a fixed granularity. We augment Darshan by proposing Darshan eXtended Tracing (DXT) for more detailed profiling of I/O software stacks. DXT enables users and administrators to vary the level of fidelity captured by Darshan at run time without modifying or recompiling applications. This capability facilitates systematic analysis on the I/O behavior of applications and can provide useful application kernel I/O traces to help advance parallel I/O research. We have demonstrated the power of DXT by obtaining a wide range of useful statistics for multiple case studies, and we further show that DXT isable to do so same with negligible overhead.

Modern HEP workflows must manage increasingly large and complex data collections. HPC facilities may be employed to help meet these workflows’ growing data processing needs. However, a better understanding of the I/O patterns and underlying bottlenecks of these workflows is necessary to meet the performance expectations of HPC systems. Darshan is a lightweight I/O characterization tool that captures concise views of HPC application I/O behavior. It intercepts application I/O calls at runtime, records file access statistics for each process, and generates log files detailing application I/O access patterns. Typical HEP workflows include event generation, detector simulation, event reconstruction, and subsequent analysis stages. A study of the I/O behavior of the ATLAS simulation and filtering stage, and the CMS simulation workflow using Darshan is presented, including insights into the I/O operations and data access size.

The utilization and performance of storage, compute, and network resources within HPC data centers have been studied extensively, but much less work has gone toward characterizing how these resources are used in conjunction to solve larger scientific challenges. To address this gap, we present our work in characterizing workloads and workflows at a data-center-wide level by examining all data transfers that occurred between storage, compute, and the external network at the National Energy Research Scientific Computing Center over a three-month period in 2019. Using a simple abstract representation of data transfers, we analyze over 100 million transfer logs from Darshan, HPSS user interfaces, and Globus to quantify the load on data paths between compute, storage, and the wide-area network based on transfer direction, user, transfer tool, source, destination, and time. We show that parallel I/O from user jobs, while undeniably important, is only one of several major I/O workloads that occurs throughout the execution of scientific workflows. We also show that this approach can be used to connect anomalous data traffic to specific users and file access patterns, and we construct time-resolved user transfer traces to demonstrate that one can systematically identify coupled data motion for individual workflows.

I/O performance is a critical aspect of data-intensive scientific computing. We seek to advance the state of the practice in understanding and diagnosing I/O performance issues through investigation of a comprehensive I/O performance data set that captures a full year of production storage activity at two leadership-scale computing facilities. We demonstrate techniques to identify regions of interest, perform focused investigations of both long-term trends and transient anomalies, and uncover the contributing factors that lead to performance fluctuation. We find that a year in the life of a parallel file system is comprised of distinct regions of long-term performance variation in addition to short-term performance transients. We demonstrate how systematic identification of these performance regions, combined with comprehensive analysis, allows us to isolate the factors contributing to different performance maladies at different time scales. From this, we present specific lessons learned and important considerations for HPC storage practitioners.

At present, I/O performance analysis requires different tools to characterize individual components of the I/O subsystem, and institutional I/O expertise is relied upon to translate these disparate data into an integrated view of application performance. This process is labor-intensive and not sustainable as the storage hierarchy deepens and system complexity increases. To address this growing disparity, we have developed the Total Knowledge of I/O (TOKIO) framework to combine the insights from existing component-level monitoring tools and provide a holistic view of performance across the entire I/O stack. A reference implementation of TOKIO, pytokio, is presented here. Using monitoring tools included with Cray XC and ClusterStor systems alongside commonly deployed community-supported tools, we demonstrate how pytokio provides a lightweight foundation for holistic I/O performance analyses on two Cray XC systems deployed at different HPC centers. We present results from integrated analyses that allow users to quantify the degree of I/O contention that affected their jobs and probabilistically identify unhealthy storage devices that impacted their performance.We also apply pytokio to inspect the utilization of NERSC’s DataWarp burst buffer and demonstrate how pytokio can be used to identify users and applications who may stand to benefit most from migrating their workloads from Lustre to the burst buffer.

I/O efficiency is essential to productivity in scientific computing, especially as many scientific domains become more data-intensive. Many characterization tools have been used to elucidate specific aspects of parallel I/O performance, but analyzing components of complex I/O subsystems in isolation fails to provide insight into critical questions: how do the I/O components interact, what are reasonable expectations for application performance, and what are the underlying causes of I/O performance problems? To address these questions while capitalizing on existing component-level characterization tools, we propose an approach that combines on-demand, modular synthesis of I/O characterization data into a unified monitoring and metrics interface (UMAMI) to provide a normalized, holistic view of I/O behavior. We evaluate the feasibility of this approach by applying it to a month-long benchmarking study on two distinct largescale computing platforms. We present three case studies that highlight the importance of analyzing application I/O performance in context with both contemporaneous and historical component metrics, and we provide new insights into the factors affecting I/O performance. By demonstrating the generality of our approach, we lay the groundwork for a production-grade framework for holistic I/O analysis.