Search NASA⌕ Search

SEARCH · Search NASA

Results for “High performance computing (HPC)”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9

LC-Opt: Benchmarking Reinforcement Learning and Agentic AI for End-to-End Liquid Cooling Optimization in Data Centers

Liquid cooling is critical for thermal management in high-density data centers with the rising AI workloads. However, machine learning-based controllers are essential to unlock greater energy efficiency and reliability, promoting sustainability. We present LC-Opt, a Sustainable Liquid Cooling (LC) benchmark environment, for reinforcement learning (RL) control strategies in energy-efficient liquid cooling of high-performance computing (HPC) systems. Built on the baseline of a high-fidelity digital twin of Oak Ridge National Lab's Frontier Supercomputer cooling system, LC-Opt provides detailed Modelica-based end-to-end models spanning site-level cooling towers to data center cabinets and server blade groups. RL agents optimize critical thermal controls like liquid supply temperature, flow rate, and granular valve actuation at the IT cabinet level, as well as cooling tower (CT) setpoints through a Gymnasium interface, with dynamic changes in workloads. This environment creates a multi-objective real-time optimization challenge balancing local thermal regulation and global energy efficiency, and also supports additional components like a heat recovery unit (HRU). We benchmark centralized and decentralized multi-agent RL approaches, demonstrate policy distillation into decision and regression trees for interpretable control, and explore LLM-based methods that explain control actions in natural language through an agentic mesh architecture designed to foster user trust and simplify system management. LC-Opt democratizes access to detailed, customizable liquid cooling models, enabling the ML community, operators, and vendors to develop sustainable data center liquid cooling control solutions.

Naug, Avisek [Hewlett Packard Enterprise]↗

Extend an innovative HPC-Compatible Multiple Temporal-spatial Resolution Concurrent Finite Element Modeling Approach to Guide Laser Powder Bed Fusion Additive

Laser power bed fusing (PBF) additive manufacturing is a key enabling technology to manufacture highly complex and integrated automotive structures. However, the geometric complexity of PBF-AM technique also leads to highly non-uniform heating and cooling rate in the manufactured part, which may cause flaw formation and produce excessive and nonuniform residual stresses, which increase quality uncertainties and manufacture issues, leading to increases in cost and energy consumption in the form of rejected parts. In this research project, we developed an innovative Multi-Spatial-Temporal-Resolution Finite Element (MUST-FE) method and completed the corresponding high performance computation (HPC) platform-based in-house code, which enables high accuracy prediction of temperature and residual stress fields for component-scale PBF-AM manufacture in efficient computation time. The MUST-FE model is calibrated and validated with a “2D pad” AlSi10Mg experiments by matching the melt pool shape and dimension, and with a “XY-cross” AlSi10Mg experiment by matching the thermal distortion and residual stress. The innovative multi-resolution and concurrent modeling approach adopted in this code ensures accuracy and computational efficiency, which will enable energy-efficient and high-yield, low-cost manufacturing of optimized, qualifiable automotive structures and contribute towards reaching technical targets outlined in AMO’s Program Plan to develop additive manufacturing systems that deliver consistently reliable parts with predictable properties.

36 MATERIALS SCIENCE↗

An empirical study of I/O separation for burst buffers in HPC systems

To meet the exascale I/O requirements for the High-Performance Computing (HPC), a new I/O subsystem, Burst Buffer, based on solid state drives (SSD), has been developed. However, the diverse HPC workloads and the bursty I/O pattern cause severe data fragmentation that requires costly garbage collection (GC) and increases the number of bytes written to the SSD. To address this data fragmentation challenge, a new multi-stream feature has been developed for SSDs. In this work, we develop an I/O Separation scheme called BIOS to leverage this multi-stream feature to group the I/O streams based on the user IDs. We propose a stream-aware scheduling policy based on burst buffer pools in the workload manager, and integrate the BIOS with the workload manager to optimize the I/O separation scheme in burst buffer. We evaluate the proposed framework with a burst buffer I/O traces from Cori Supercomputer including a diverse set of applications. Experimental results show that the BIOS could improve the performance by 1.44x on average and reduce the Write Amplification Factor (WAF) by up to 1.20x. Finally, these demonstrate the potential benefits of the I/O separation scheme for solid state storage systems.

97 MATHEMATICS AND COMPUTING↗

Quantum/AI Topology-Aware Latency-Adaptive HPC Workflow Scheduling Optimization

The growing demand for more powerful high-performance computing (HPC) systems has led to a steady rise in energy consumption by supercomputing worldwide. This study is focused on comparing our Application-Topology Mapper (ATMapper) to the popular Simple Linux Utility for Resource Management (SLURM) for the purpose of exploring methods that can further optimize job-scheduling within HPC systems. ATMapper is an Artificial-Intelligence based approach to job-scheduling that is currently being enhanced with quantum annealing (QA) to generate optimal schedules faster. We are applying QA to speedup our ATMapper process to achieve higher computing efficiency, thereby reducing HPC energy consumption. Here, we examine how four job-scheduling approaches perform in processor node assignment when using an example network architecture of 4 interconnected nodes. Using a specialized script, we are assessing the schedule of a computation flow with 11 interdependent tasks. The data movements among nodes were tracked to count for the number of interactions (network hops) between nodes needed to complete the tasks. The total number of hops and the job completion time were then used to quantify the efficiency of the different mapping approaches. In addition to SLURM, we also compare our ATMapper to the QA-enabled LBNL TIGER and the D-Wave Distributed Computing processor assignment approaches. The preliminary results showed that our topology-aware, latency-adaptive ATMapper is significantly more efficient when compared to the other scheduling approaches due to its load-imbalance network allocation. The scheduler displayed a computing efficiency of 53% by performing significantly fewer network hops than its alternatives. By reducing the number of hops, ATMapper was able to perform all 11 tasks by using only 3 nodes out of given 4. This research indicates the potential to use QA/AI for HPC job-scheduling. Later, we will test a SLURM simulator program to draw further comparisons on the effectiveness of ATMapper's scheduling approach. The results of this comparison will serve as a baseline for later improving SLURM's performance using a QA-enhanced ATMapper approach.

Caraveo, Braulio [University of Huston - Clear Lak↗

Tandem Predictions for HPC Jobs

At the core of the predictive analytics applied to High Performance Computing (HPC), the most prominent tasks are the prediction of job runtimes and the prediction of job queue times, both of which have the potential for informing HPC users during their every-day decision making. Accurate runtime predictions can help users better choose so-called wallclock times at job submission, decreasing the odds of their jobs waiting in queues longer than necessary. The accurate and timely queue time predictions offered for the available partitions can inform the favorable selection of partitions for running jobs. This potential is well understood as we see in the abundance of research studies that propose solutions for these tasks, including the work published in the last several years. These tasks are seemingly receptive to the Machine Learning (ML) solutions, considering that there is no shortage of training data where HPC centers over time run millions and millions of jobs. However, we study the existing research literature, as well as look for examples in the toolchains supported on the exemplar HPC facilities, and, surprisingly, do not find any practical solutions that are ready to be adopted. We interpret this as a manifestation of the shortage of UX/UI efforts that support HPC analytics and also as a sign that the research has not come to the consensus on solving these tasks. In this study, we aim to shed new light on the long-running task of job queue time prediction by exploring the utility of runtime predictions in improving prediction accuracy and, actually, predicting these two metrics together, in tandem. In other words, we show how runtime predictions become valuable input in the queue time modeling. We challenge the existing approaches to feature engineering for the queue time prediction and describe promising results we obtained for a large dataset of HPC jobs from a supercomputer at the National Renewable Energy Laboratory.

HPC↗

An HPC benchmark survey and taxonomy for characterization

The field of High-Performance Computing (HPC) is defined by providing computing devices with highest performance for a variety of demanding scientific users. The tight co-design relationship between HPC providers and users propels the field forward, paired with technological improvements, achieving continuously higher performance and resource utilization. A key device for system architects, architecture researchers, and scientific users are benchmarks, allowing for well-defined assessment of hardware, software, and algorithms. Many benchmarks exist in the community, from individual niche benchmarks testing specific features, to large-scale benchmark suites for whole procurements. We survey the available HPC benchmarks, summarizing them in table form with key details and concise categorization, also through an interactive website. For categorization, we present a benchmark taxonomy for well-defined characterization of benchmarks.

Benchmarking↗

The Kokkos Ecosystem [Brief]

In 2016/2017, the field of High-Performance Computing (HPC) entered a new era driven by fundamental physics challenges to produce ever more energy and cost-efficient processors. Since the convergence on the Message-Passing Interface (MPI) standard in the mid-1990s, application developers enjoyed a seemingly static view of the underlying machine — that of a distributed collection of homogeneous nodes executing in collaboration. However, after almost two decades of dominance, the sole use of MPI to derive parallelism acted as a limiter to improved future performance. While MPI is widely expected to continue to function as the basic mechanism for communication between compute nodes for the immediate future, additional parallelism is required on the computing node itself if high performance and efficiency goals are to be realized. When reviewing the architectures of the top HPC systems today, the change in paradigm is clear: the compute nodes of the leading machines in the world are either powered by many-core chips with a few dozen cores each, or use heterogeneous designs, where traditional CPUs marshal work to massively parallel compute accelerators which has as many as 200,000 processing threads in flight simultaneously. Complicating matters further for application developers, each processor vendor has its own preferred way of writing code for their architecture.The Kokkos EcoSystem was released by Sandia in 2017 to address this new era in HPC system design by providing a vendor independent performance portable programming system for scientific, engineering, and mathematical software applications written in the C++ programming language. Using Kokkos, application developers can be more productive because they will not have to create and maintain separate versions of their software for each architecture, nor will they have to be experts in each architecture's peculiar requirements. Instead, they will have a single method of programming for the diverse set of modern HPC architectures. While Kokkos started in 2011 as a programming model only, it soon became clear that complex applications needed more. It is also critical to have a portable mathematical functions and developers need tools to debug their applications, gain insight into the performance characteristics of their codes and tune algorithm performance parameters through automated processes. The Kokkos EcoSystem addresses those needs through its three main components: the Kokkos Core programming model, the Kokkos Kernels math library, and the Kokkos Tools project.

97 MATHEMATICS AND COMPUTING↗

Design and implementation of I/O performance prediction scheme on HPC systems through large-scale log analysis

Abstract Large-scale high performance computing (HPC) systems typically consist of many thousands of CPUs and storage units used by hundreds to thousands of users simultaneously. Applications from large numbers of users have diverse characteristics, such as varying computation, communication, memory, and I/O intensity. A good understanding of the performance characteristics of each user application is important for job scheduling and resource provisioning. Among these performance characteristics, I/O performance is becoming increasingly important as data sizes rapidly increase and large-scale applications, such as simulation and model training, are widely adopted. However, predicting I/O performance is difficult because I/O systems are shared among all users and involve many layers of software and hardware stack, including the application, network interconnect, operating system, file system, and storage devices. Furthermore, updates to these layers and changes in system management policy can significantly alter the I/O behavior of applications and the entire system. To improve the prediction of the I/O performance on HPC systems, we propose integrating information from several different system logs and developing a regression-based approach to predict the I/O performance. Our proposed scheme can dynamically select the most relevant features from the log entries using various feature selection algorithms and scoring functions, and can automatically select the regression algorithm with the best accuracy for the prediction task. The evaluation results show that our proposed scheme can predict the write performance with up to 90% prediction accuracy and the read performance with up to 99% prediction accuracy using the real logs from the Cori supercomputer system at NERSC.

97 MATHEMATICS AND COMPUTING↗

Toward lower-diameter large-scale HPC and data center networks with co-packaged optics

We investigate the advantages of using co-packaged optics for building low-diameter, large-scale high-performance computing (HPC) and data center networks. The increased escape bandwidth offered by co-packaged optics can enable high-radix switch implementations of more than 150 switch ports, which can be combined with data rates of up to 400 Gb/s per port. From the network architecture perspective, the key benefits of using co-packaged optics in future fat-tree networks include (a) the ability to implement large-scale topologies of > <#comment/> 11 , 000 end points by eliminating the need for a third switching layer and (b) the ability to provide up to 4 × <#comment/> higher bisection bandwidth compared to existing solutions, reducing at the same time the number of required switch application-specific integrated circuits by > <#comment/> 80 % <#comment/> . From the network operation perspective, both reduced energy consumption and lower packet delays can be achieved since fewer hops are required; i.e., packets need to traverse fewer serializer/deserializer lanes and fewer switch buffers, which reduces the probability of contending with other packets and improves the tolerance of network congestion. The performance of the proposed architecture is evaluated via discrete-event simulations for a wide range of representative HPC synthetic-traffic cases that include both hotspot and non-hotspot scenarios. The simulation results suggest that co-packaged optics form a promising solution to keep up with bandwidth scaling in future networks, while the reduced number of switching layers can lead to significant mean packet delay improvements that start from 30% and reach up to 74% for high-load conditions.

Maniotis, Pavlos (ORCID:0000000244905253)↗

Optimizing I/O Performance of HPC Applications with Autotuning

Parallel Input output is an essential component of modern high-performance computing (HPC). Obtaining good I/O performance for a broad range of applications on diverse HPC platforms is a major challenge, in part, because of complex inter dependencies between I/O middleware and hardware. The parallel file system and I/O middleware layers all offer optimization parameters that can, in theory, result in better I/O performance. Unfortunately, the right combination of parameters is highly dependent on the application, HPC platform, problem size, and concurrency. Scientific application developers do not have the time or expertise to take on the substantial burden of identifying good parameters for each problem configuration. They resort to using system defaults, a choice that frequently results in poor I/O performance. We expect this problem to be compounded on exascale-class machines, which will likely have a deeper software stack with hierarchically arranged hardware resources. We present as a solution to this problem an autotuning system for optimizing I/O performance, I/O performance modeling, I/O tuning, and I/O patterns. We demonstrate the value of this framework across several HPC platforms and applications at scale.

Behzad, Babak↗

Towards performance-portability of the Albany Finite Element analysis code using the Kokkos library of Trilinos

Performance portability on heterogeneous high-performance computing (HPC) systems is a major challenge faced today by code developers: parallel code needs to execute correctly as well as with high performance on machines with different architectures, operating systems, and software libraries. The Finite Element Method (FEM) is a popular and flexible method for discretizing partial differential equations arising in a wide variety of scientific, engineering, and industry applications that require HPC. This paper presents some preliminary results pertaining to our development of a performance portable implementation of the FEM-based Albany code. Performance portability is achieved using the Kokkos library of Trilinos. We present performance results for two different physics simulations modules in Albany: the Aeras global atmosphere dynamical code and the FELIX land-ice solver. As a result, numerical experiments show that our single code implementation gives reasonable performance across two multi-core/many-core architectures: NVIDIA GPUs and multi-core CPUs.

97 MATHEMATICS AND COMPUTING↗

2019 Budget Request for the DOE Computational Science Graduate Fellowship (CSGF) Grant

The Department of Energy Computational Science Graduate Fellowship (DOE CSGF) is necessary to meet the continual challenging national workforce needs that arise as computational science and engineering problems continue to grow in scope and complexity. Computational science and engineering (CSE) is a multidisciplinary approach that uses scientific computing to solve practical problems methods and to supply technical tools across the scientific discovery spectrum. In particular, the DOE CSGF emphasizes high-performance computing (HPC) that enables CSE that advances science and engineering in directions important to the DOE and the economy in general. Over the past half-century, HPC has been an essential tool for DOE’s success. During this period, important missions, such as nuclear stockpile stewardship, have turned to HPC as an essential technology. Entire science disciplines, such as biology and cosmology, have been transformed through the augmentation of scientific observation via HPC. At government laboratories and in industry, DOE CSGF alumni are helping push traditional HPC boundaries while contributing to discoveries in high-energy physics, renewable energy, fusion-reactor design, additive manufacturing, nanomaterials for next-generation batteries and transistors, and turbine and advanced nuclear reactor modeling. In addition, HPC is used to address national health needs that will eventually point to cures both by helping cancer researchers manage and analyze huge troves of data, by simulating biological mechanisms, and by accelerating drug development — including continuing to rise to the challenge of pandemic-related research. A 2023 report from the ASCAC Subcommittee on American Competitiveness and Innovation to the ASCR office, “Can the United States Maintain Its Leadership in High-Performance Computing?” says of the Program, “The CSGF program provides a barometer for disciplines that will be of interest to future DOE computing.” An explosion in scientific and technological data has driven the need for increasingly sophisticated HPC to transform those data into scientific understanding. With access to more and more data and the proliferation of HPC, Machine Learning and Artificial Intelligence are experiencing a renaissance, complementing the now well-established use of computational simulation. Indeed, in its September 2020 subcommittee report on “AI/ML, Data Intensive Science and High-Performance Computing”, the DOE Advanced Scientific Computing Advisory Committee (ASCAC) explicitly called for a fellowship program to train computational and data scientists to tackle exascale and data-intensive computing challenges. This collaboration of empirical and theory-based modeling will increasingly inform federal policymakers whose decisions affect American society and future generations, and it requires highly skilled and intellectually agile computational scientists who can support the fast-moving DOE National Laboratory research environment. In fact, the DOE CSGF program has explicitly and consistently addressed this need.

97 MATHEMATICS AND COMPUTING↗

Spacer Grid and Mixing Vanes Effects Study on Two-phase Flow Using Bubble Tracking Simulations

In Light Water Reactor (LWR) cores, the nuclear fuel rods are arranged together as rod bundles in a triangular or square pattern. The fuel rod bundles are supported by the spacer grids with mixing vanes. An understanding of spacer grid and mixing vanes effect is important since it has a direct impact on the turbulence structure in the flow and the heat removal efficiency from the fuel rod surface. Although a considerable amount of literature has been published on the experimental investigation of spacer grid and mixing vanes effect, the measurements were mainly focused on single-phase flow. The influence of spacer grid and mixing vanes on two-phase flow is yet to be fully understood. Some preliminary research efforts have been invested in nuclear community. For example, Wheeler et al. experimentally studied the effects of spacer grid in air-water two-phase flow. Although mixing vanes are not considered, they presented interesting results of time-averaged two-phase flow quantities at upstream and downstream locations of the spacer grid. Over the past decades, the rapid development of High-Performance Computing (HPC) is allowing for the increasingly larger-scale high-fidelity simulations. Thanks to this favorable historical trend, Direct Numerical Simulation (DNS), integrated with Interface Tracking Methods (ITM), has been emerging as a valuable tool to complement and expand our ability to understand the two-phase flow phenomenon. Meanwhile, in order to make best use of the large-scale two-phase flow simulations, advanced data analysis techniques are being developed, as they enable the extraction of detailed parameters about individual bubble behavior. The level set ITM is utilized in our DNS flow solver, which makes use of a signed distance field. The gas-liquid interface is modeled by the zero level set. The level set method can be readily used to distinguish different phases based on the sign of corresponding level set value (e.g. the sign is positive in the liquid phase while negative in the gas phase.). However, the traditional level set approach is not able to collect calculated values and associate with specific bubbles when multiple (even thousands of) bubbles present in the simulations of two-phase flow. This drawback hinders the collection of useful bubble information, which can give us valuable insights about bubbly flows. For example, how the different local fluid conditions could affect bubble interfacial forces, bubble deformation level, and eventually the bubble distribution throughout the whole domain. A bubble tracking methodology has been developed to address the drawbacks and improve the data collection efficiency of level set based two-phase flow simulations. In the present work, a single Pressurized Water Reactor (PWR) subchannel with reduced size spacer grid and mixing vanes is chosen as the computational domain. The springs and dimples of the spacer grid are not represented for simplicity in the present investigation. The bubble tracking methodology is applied to the simulations of two-phase bubbly flow through spacer grid and mixing vanes region. In the intended simulation, the liquid turbulence will be fully resolved by DNS while the two-phase behaviors are captured by level set ITM. Bubble tracking data are to be processed to investigate the spacer grid and mixing vanes effect on two-phase bubbly flow. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Advancing Sustainability in Data Centers: Evaluation of Hybrid Air/Liquid Cooling Schemes for IT Payload Using Sea Water

Abstract-The growth in cloud computing, Big Data, AI and high-performance computing (HPC) necessitate the deployment of additional data centers (DC's) with high energy demands. The unprecedented increase in the Thermal Design Power (TDP) of the computing chips will require innovative cooling techniques. Furthermore, DC's are increasingly limited in their ability to add powerful GPU servers by power capacity constraints. As cooling energy use accounts for up to 40% of DC energy consumption, creative cooling solutions are urgently needed to allow deployment of additional servers, enhance sustainability and increase energy efficiency of DC's. The information in this study is provided from Start Campus' Sines facility supported by Alfa Laval for the heat exchanger and CO 2 emission calculations. The study evaluates the performance and sustainability impact of various data center cooling strategies including an air-only deployment and a subsequent hybrid air/water cooling solution all utilizing sea water as the cooling source. Here we evaluate scenarios from 3 MW to 15+1 MW of IT load in 3 MW increments which correspond to the size of heat exchangers used in the Start Campus' modular system design. This study also evaluates the CO 2 emissions compared to a conventional chiller system for all the presented scenarios. Results indicate that the effective use of the sea water cooled system combined with liquid cooled systems improve the efficiency of the DC, plays a role in decreasing the CO 2 emissions and supports in achieving sustainability goals.

97 MATHEMATICS AND COMPUTING↗

Trends in data locality abstractions for HPC systems

The cost of data movement has always been an important concern in high performance computing (HPC) systems. It has now become the dominant factor in terms of both energy consumption and performance. Support for expression of data locality has been explored in the past, but those efforts have had only modest success in being adopted in HPC applications for various reasons. However, with the increasing complexity of the memory hierarchy and higher parallelism in emerging HPC systems, locality management has acquired a new urgency. Developers can no longer limit themselves to low-level solutions and ignore the potential for productivity and performance portability obtained by using locality abstractions. Fortunately, the trend emerging in recent literature on the topic alleviates many of the concerns that got in the way of their adoption by application developers. Data locality abstractions are available in the forms of libraries, data structures, languages and runtime systems; a common theme is increasing productivity without sacrificing performance. Furthermore, this paper examines these trends and identifies commonalities that can combine various locality concepts to develop a comprehensive approach to expressing and managing data locality on future large-scale high-performance computing systems.

97 MATHEMATICS AND COMPUTING↗

Highly interactive, steered scientific workflows on HPC systems: optimizing design solutions

Scientific workflows are becoming increasingly important in high performance computing (HPC) settings, as the feasibility and appeal of many simultaneous heterogeneous tasks increases with increasing hardware capabilities. Currently no HPC-based workflow platform supports a dynamically adaptable workflow with interactive steering and analysis at run-time. Furthermore, for most workflow programs, compute resources are fixed for a given instance, resulting in a possible waste of expensive allocation resources when tasks are spawned and killed. Here we describe the design and testing of a run-time-interactive, adaptable, steered workflow tool capable of executing thousands of parallel tasks without an MPI programming model, using a database management system to facilitate task management through multiple live connections. We find that on the Oak Ridge Leadership Computing Facility pre-exascale Summit supercomputer it is possible to launch and interactively steer workflows with thousands of simultaneous tasks with negligible latency. For the case of particle simulation and analysis tasks that run for minutes to hours, this paradigm offers the prospect of a robust and efficient means to perform simulation-space exploration with on-the-fly analysis and adaptation.

Ossyra, John R↗

Exploiting Redundancy and Application Scalability for Cost-Effective, Time-Constrained Execution of HPC Applications on Amazon EC2

The use of clouds to execute high-performance computing (HPC) applications has greatly increased recently. Clouds provide several potential advantages over traditional supercomputers and in-house clusters. The most popular cloud is currently Amazon EC2, which provides fixed-cost and variable-cost, auction-based options. The auction market trades lower cost for potential interruptions that necessitate checkpointing; if the market price exceeds the bid price, a node is taken away from the user without warning. We explore techniques to maximize performance per dollar given a time constraint within which an application must complete. Specifically, we design and implement multiple techniques to reduce expected cost by exploiting redundancy in the EC2 auction market. We then design an adaptive algorithm that selects a scheduling algorithm and determines the bid price. We show that our adaptive algorithm executes programs up to seven times cheaper than using the on-demand market and up to 44 percent cheaper than the best non-redundant, auction-market algorithm. We extend our adaptive algorithm to incorporate application scalability characteristics for further cost savings. In conclusion, we show that the adaptive algorithm informed with scalability characteristics of applications achieves up to 56 percent cost savings compared to the expected cost for the base adaptive algorithm run at a fixed, user-defined scale.

97 MATHEMATICS AND COMPUTING↗

/Scratch as a Cache: Rethinking HPC Center Scratch Storage

To sustain emerging data-intensive scientific applications, High Performance Computing (HPC) centers invest a notable fraction of their operating budget on a specialized fast storage system, scratch space, which is designed for storing the data of currently running and soon-to-run HPC jobs. Instead, it is often used as a standard file system, wherein users arbitrarily store their data, without any consideration to the center's overall performance. To remedy this, centers periodically scan the scratch in an attempt to purge transient and stale data. This practice of supporting a cache workload using a file system and disjoint tools for staging and purging results in suboptimal use of the scratch space. In this paper, we address the above issues by proposing a new perspective, where the HPC scratch space is treated as a cache, and data population, retention, and eviction tools are integrated with scratch management. In our approach, data is moved to the scratch space only when it needed, and unneeded data is removed as soon as possible. We also design a new job-workflow-aware caching policy that leverages user-supplied hints for managing the cache. Our evaluation using three-year job logs from the Jaguar supercomputer, shows that compared to the widely-used purge approach, workflow-aware caching optimizes scratch utilization by reducing the average amount of data read by 9.3%, and by reducing job scheduling delays associated with data staging, on average, by 282.0%.

99 GENERAL AND MISCELLANEOUS↗