Engineering Papers⌕ Search

SEARCH · Engineering Papers

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 1,459 records · Page 81

Dynamic Non-Hierarchical File Systems for Exascale Storage

This constitutes the final report for “Dynamic Non-Hierarchical File Systems for Exascale Storage”. The ultimate goal of this project was to improve data management in scientific computing and high-end computing (HEC) applications, and to achieve this goal we proposed: to develop the first, HEC-targeted, file system featuring rich metadata and provenance collection, extreme scalability, and future storage hardware integration as core design goals, and to evaluate and develop a flexible non-hierarchical file system interface suitable for providing more powerful and intuitive data management interfaces to HEC and scientific computing users. Data management is swiftly becoming a serious problem in the scientific community – while copious amounts of data are good for obtaining results, finding the right data is often daunting and sometimes impossible. Scientists participating in a Department of Energy workshop noted that most of their time was spent “...finding, processing, organizing, and moving data and it’s going to get much worse”. Scientists should not be forced to become data mining experts in order to retrieve the data they want, nor should they be expected to remember the naming convention they used several years ago for a set of experiments they now wish to revisit. Ideally, locating the data you need would be as easy as browsing the web. Unfortunately, existing data management approaches are usually based on hierarchical naming, a 40 year-old technology designed to manage thousands of files, not exabytes of data. Today’s systems do not take advantage of the rich array of metadata that current high-end computing (HEC) file systems can gather, including content-based metadata and provenance1 information. As a result, current metadata search approaches are typically ad hoc and often work by providing a parallel management system to the “main” file system, as is done in Linux (the locate utility), personal computers, and enterprise search appliances. These search applications are often optimized for a single file system, making it difficult to move files and their metadata between file systems. Users have tried to solve this problem in several ways, including the use of separate databases to index file properties, the encoding of file properties into file names, and separately gathering and managing provenance data, but none of these approaches has worked well, either due to limited usefulness or scalability, or both. Our research addressed several key issues: High-performance, real-time metadata harvesting: extracting important attributes from files dynamically and immediately updating indexes used to improve search; Transparent, automatic, and secure provenance capture: recording the data inputs and processing steps used in the production of each file in the system; Scalable indexing: indexes that are optimized for integration with the file system; Dynamic file system structure: our approach provides dynamic directories similar to those in semantic file systems, but these are the native organization rather than a feature grafted onto a conventional system. In addition to these goals, our research effort will include evaluating the impact of new storage technologies on the file system design and performance. In particular, the indexing and metadata harvesting functions can potentially benefit from the performance improvements promised by new storage class memories.

42 ENGINEERING↗

Optimization of Error-Bounded Lossy Compression for Hard-to-Compress HPC Data

Since today’s scientific applications are producing vast amounts of data, compressing them before storage/transmission is critical. Results of existing compressors show two types of HPC data sets: highly compressible and hard to compress. In this work, we carefully design and optimize the error-bounded lossy compression for hard-tocompress scientific data. We propose an optimized algorithm that can adaptively partition the HPC data into best-fit consecutive segments each having mutually close data values, such that the compression condition can be optimized. Another significant contribution is the optimization of shifting offset such that the XOR-leading-zero length between two consecutive unpredictable data points can be maximized. We finally devise an adaptive method to select the best-fit compressor at runtime for maximizing the compression factor. We evaluate our solution using 13 benchmarks based on real-world scientific problems, and we compare it with 9 other state-of-the-art compressors. Experiments show that our compressor can always guarantee the compression errors within the user-specified error bounds. Most importantly, our optimization can improve the compression factor effectively, by up to 49% for hard-tocompress data sets with similar compression/decompression time cost.

97 MATHEMATICS AND COMPUTING↗

Collaborative Visualization for Large-Scale Accelerator Electromagnetic Modeling (Final Report)

This report contains the comprehensive summary of the work performed on the SBIR Phase II, Collaborative Visualization for Large-Scale Accelerator Electromagnetic Modeling at Kitware Inc. in collaboration with Stanford Linear Accelerator Center (SLAC). The goal of the work was to develop collaborative visualization tools for large-scale data as illustrated in the figure below. The solutions we proposed address the typical problems faced by geographicallyand organizationally-separated research and engineering teams, who produce large data (either through simulation or experimental measurement) and wish to work together to analyze and understand their data. Because the data is large, we expect that it cannot be easily transported to each team member's work site, and that the visualization server must reside near the data. Further, we also expect that each work site has heterogeneous resources: some with large computing clients, tiled (or large) displays and high bandwidth; others sites as simple as a team member on a laptop computer. Our solution is based on the open-source, widely used ParaView large-data visualization application. We extended this tool to support multiple collaborative clients who may locally visualize data, and then periodically rejoin and synchronize with the group to discuss their findings. Options for managing session control, adding annotation, and defining the visualization pipeline, among others, were incorporated. We also developed and deployed a Web visualization framework based on ParaView that enables the Web browser to act as a participating client in a collaborative session. The ParaView Web Visualization framework leverages various Web technologies including WebGL, JavaScript, Java and Flash to enable interactive 3D visualization over the web using ParaView as the visualization server. We steered the development of this technology by teaming with the SLAC National Accelerator Laboratory. SLAC has a computationally-intensive problem important to the nations scientific progress as described shortly. Further, SLAC researchers routinely generate massive amounts of data, and frequently collaborate with other researchers located around the world. Thus SLAC is an ideal teammate through which to develop, test and deploy this technology. The nature of the datasets generated by simulations performed at SLAC presented unique visualization challenges especially when dealing with higher-order elements that were addressed during this Phase II. During this Phase II, we have developed a strong platform for collaborative visualization based on ParaView. We have developed and deployed a ParaView Web Visualization framework that can be used for effective collaboration over the Web. Collaborating and visualizing over the Web presents the community with unique opportunities for sharing and accessing visualization and HPC resources that hitherto with either inaccessible or difficult to use. The technology we developed in here will alleviate both these issues as it becomes widely deployed and adopted.

43 PARTICLE ACCELERATORS↗

Extending I/O through High Performance Data Services

The complexity of HPC systems has increased the burden on the developer as applications scale to hundreds of thousands of processing cores. Moreover, additional efforts are required to achieve acceptable I/O performance, where it is important how I/O is performed, which resources are used, and where I/O functionality is deployed. Specifically, by scheduling I/O data movement and by effectively placing operators affecting data volumes or information about the data, tremendous gains can be achieved both in the performance of simulation output and in the usability of output data. Previous studies have shown the value of using asynchronous I/O, of employing a staging area, and of performing select operations on data before it is written to disk. Leveraging such insights, this paper develops and experiments with higher level I/O abstractions, termed 'data services', that manage output data from 'source to sink': where/when it is captured, transported towards storage, and filtered or manipulated by service functions to improve its information content. Useful services include data reduction, data indexing, and those that manage how I/O is performed, i.e., the control aspects of data movement. Our data services implementation distinguishes control aspects - the control plane - from data movement - the data plane, so that both may be changed separably. This results in runtime flexibility not only in which services to employ, but also in where to deploy them and how they use I/O resources. The outcome is consistently high levels of I/O performance at large scale, without requiring application change.

97 MATHEMATICS AND COMPUTING↗

An Early Investigation of the HHL Quantum Linear Solver for Scientific Applications

In this paper, we explore using the Harrow–Hassidim–Lloyd (HHL) algorithm to address scientific and engineering problems through quantum computing, utilizing the NWQSim simulation package on a high-performance computing platform. Focusing on domains such as power-grid management and climate projection, we demonstrate the correlations of the accuracy of quantum phase estimation, along with various properties of coefficient matrices, on the final solution and quantum resource cost in iterative and non-iterative numerical methods such as the Newton–Raphson method and finite difference method, as well as their impacts on quantum error correction costs using the Microsoft Azure Quantum resource estimator. We summarize the exponential resource cost from quantum phase estimation before and after quantum error correction and illustrate a potential way to reduce the demands on physical qubits. This work lays down a preliminary step for future investigations, urging a closer examination of quantum algorithms’ scalability and efficiency in domain applications.

hybrid software for QC-HPC↗

Effective Software Design and Development for the New Graph Architecture HPC Machines

Software applications need to change and adapt as modern architectures evolve. Nowadays advancement in chip design translates to increased parallelism. Exploiting such parallelism is a major challenge in modern software engineering. Multicore processors are about to introduce a significant change in the way we design and use fundamental data structures. In this work we describe the design and programming principles of a software library of highly concurrent scalable and nonblocking data containers. In this project we have created algorithms and data structures for handling fundamental computations in massively multithreaded contexts, and we have incorporated these into a usable library with familiar look and feel. In this work we demonstrate the first design and implementation of a wait-free hash table. Our multiprocessor data structure design allows a large number of threads to concurrently insert, remove, and retrieve information. Non-blocking designs alleviate the problems traditionally associated with the use of mutual exclusion, such as bottlenecks and thread-safety. Lock-freedom provides the ability to share data without some of the drawbacks associated with locks, however, these designs remain susceptible to starvation. Furthermore, wait-freedom provides all of the benefits of lock-free synchronization with the added assurance that every thread makes progress in a finite number of steps. This implies deadlock-freedom, livelock-freedom, starvation-freedom, freedom from priority inversion, and thread-safety. The challenges of providing the desirable progress and correctness guarantees of wait-free objects makes their design and implementation difficult. There are few wait-free data structures described in the literature. Using only standard atomic operations provided by the hardware, our design is portable; therefore, it is applicable to a variety of data-intensive applications including the domains of embedded systems and supercomputers.Our experimental evaluation shows that our hash table design outperforms the most advanced locking solution, provided by Intel's TBB library, by 22%. When compared to more traditional locking designs we show a performance improvement by a factor of 7.92. When compared to alternative non-blocking designs, our hash table demonstrates solid performance gains in a large majority of cases, typically by a factor of 3.44.

99 GENERAL AND MISCELLANEOUS↗

Formal Definitions and Performance Comparison of Consistency Models for Parallel File Systems

The semantics of HPC storage systems are defined by the consistency models to which they abide. Storage consistency models have been less studied than their counterparts in memory systems, with the exception of the POSIX standard and its strict consistency model. The use of POSIX consistency imposes a performance penalty that becomes more significant as the scale of parallel file systems increases and the access time to storage devices, such as node-local solid storage devices, decreases. While some efforts have been made to adopt relaxed storage consistency models, these models are often defined informally and ambiguously as by-products of a particular implementation. Here in this work, we establish a connection between memory consistency models and storage consistency models and revisit the key design choices of storage consistency models from a high-level perspective. Further, we propose a formal and unified framework for defining storage consistency models and a layered implementation that can be used to easily evaluate their relative performance for different I/O workloads. Finally, we conduct a comprehensive performance comparison of two relaxed consistency models on a range of commonly seen parallel I/O workloads, such as checkpoint/restart of scientific applications and random reads of deep learning applications. We demonstrate that for certain I/O scenarios, a weaker consistency model can significantly improve the I/O performance. For instance, in small random reads that are typically found in deep learning applications, session consistency achieved a 5x improvement in I/O bandwidth compared to commit consistency, even at small scales.

97 MATHEMATICS AND COMPUTING↗

Performance on HPC Platforms Is Possible Without C++

Computing at large scales has become extremely challenging due to increasing heterogeneity in both hardware and software. More and more scientific workflows must tackle a range of scales and use machine learning and AI intertwined with more traditional numerical modeling methods, placing more demands on computational platforms. These constraints indicate a need to fundamentally rethink the way computational science is done and the tools that are needed to enable these complex workflows. The current set of C++-based solutions may not suffice, and relying exclusively upon C++ may not be the best option, especially because several newer languages and boutique solutions offer more robust design features to tackle the challenges of heterogeneity. In June 2023, we held a mini symposium that explored the use of newer languages and heterogeneity solutions that are not tied to C++ and that offer options beyond template metaprogramming and Parallel. For for performance and portability. In conclusion, we describe some of the presentations and discussion from the mini symposium in this article.

97 MATHEMATICS AND COMPUTING↗

Landrush: Rethinking In-Situ Analysis for GPGPU Workflows

In-situ analysis on the output data of scientific simulations has been made necessary by ever-growing output data volumes and increasing costs of data movement as supercomputing is moving towards exascale. With hardware accelerators like GPUs becoming increasingly common in high end machines, new opportunities arise to co-locate scientific simulations and online analysis performed on the scientific data generated by the simulations. However, the asynchronous nature of GPGPU programming models and the limited context-switching capabilities on the GPU pose challenges to co-locating the scientific simulation and analysis on the same GPU. This paper dives deeper into these challenges to understand how best to co-locate analysis with scientific simulations on the GPUs in HPC clusters. Specifically, our 'Landrush' approach to GPU sharing proposes a solution that utilizes idle cycles on the GPU to provide an improved time-to-answer, that is, the total time to run the scientific simulation and analysis of the generated data. Landrush is demonstrated with experimental results obtained from leadership high-end applications on ORNL's Titan supercomputer, which show that (i) GPU-based scientific simulations have varying degrees of idle cycles to afford useful analysis task co-location, and (ii) the inability to context switch on the GPU at instruction granularity can be overcome by careful control of the analysis kernel launches and software-controlled early completion of analysis kernel executions. Results show that Landrush is superior in terms of time-to-answer compared to serially running simulations followed by analysis or by relying on the GPU driver and hardwired thread dispatcher to run analysis concurrently on a single GPU.

Computer Science↗

StoreRush: An Application-Level Approach to Harvesting Idle Storage in a Best Effort Environment

For a production HPC system where storage devices are shared between multiple applications and managed in a best effort manner, contention is often a major problem leading to some storage devices being more loaded than others and causing a significant reduction in I/O throughput. In this paper, we describe our latest efforts StoreRush to resolve this practical issue at the application level without requiring modification to the file and storage system. The proposed scheme uses a two-level messaging system to harvest idle storage via re-routing I/O requests to a less congested storage location so that write performance is improved while limiting the impact on read by throttling re-routing if deemed too much. An analytical model is derived to guide the setup of optimal throttling factor. The proposed scheme is verified against production applications Pixie3D, XGC1 and QMCPack during production windows, which very well demonstrated the effectiveness (e.g., up to 1.8x improvement in write) and scalability of our approach (up to 131,072 cores).

97 MATHEMATICS AND COMPUTING↗

Development of HPC based phase field simulations tool for modification of alloy morphology to enhance material properties during additive manufacturing (AM) process

Currently, additive manufacturing (AM) technology is extensively used to manufacture aerospace components because of the unique advantages of the technique in producing custom designs, complex geometries in a near-net fashion that significantly reduces the manufacturing cost, and improved energy efficiency. However, the thermodynamic properties of existing aerospace alloys and the thermal conditions that prevail in most of the powder-bed based additive manufacturing processes lead to undesirable columnar / columnar-dendritic microstructures that promote solidification cracking during the process, and the development of anisotropy in the mechanical properties of the as-built components. Time consuming, energy-intensive post AM treatments are required to recover the mechanical properties of the wrought components for which the existing alloy compositions were designed for. However, novel alloy compositions, that exploit the AM thermal conditions to produce fine, equiaxed grain structures in the as-built condition, with mechanical properties equal to or exceeding those of the wrought work-horse alloy, Ti-6Al-4V has recently been demonstrated in Ti-Cu alloys. Significant energy savings up to 66% can be obtained for a typical aerospace component through digitally designing the AM process, process parameters, and alloy composition. The current project focused on developing a fundamental understanding of the evolution of the columnar to equiaxed transition (CET) occurring during solidification under AM thermal conditions in Ti-Cu and Ti-Cu-X alloys using high-fidelity phase field (PF) simulations using the leadership class computing facilities that exist in the national laboratories. The simulations were able to capture the effect of alloy and process conditions on CET, and clearly showed the beneficial effect of a ternary solute addition to Ti-Cu binary alloys on CET. The simulations indicated that bulk nucleation in the liquid ahead of the solidifying epitaxial front was promoted by a large, transient, thermal undercooling promoted by the mismatch between the growth rate of the dendrite tips in the epitaxial front and the imposed solidification rate. AM experiments performed at Raytheon Technologies Research Center using the same thermal conditions used in the PF simulations indicated the formation of equiaxed grains, although the grain size was bigger than the ones reported in the literature. On the other hand, the equiaxed grain size predicted by PF simulations were smaller than the grain size of Ti-Cu alloys reported in the literature. The industry will follow up the existing work in a future program that would involve further optimization of the process for the Ti-Cu-X alloys to demonstrate the application of the process and the alloy to a specific aerospace component.

36 MATERIALS SCIENCE↗

ExaFEL: extreme-scale real-time data processing for X-ray free electron laser science

ExaFEL is an HPC-capable X-ray Free Electron Laser (XFEL) data analysis software suite for both Serial Femtosecond Crystallography (SFX) and Single Particle Imaging (SPI) developed in collaboration with the Linac Coherent Lightsource (LCLS), Lawrence Berkeley National Laboratory (LBNL) and Los Alamos National Laboratory. ExaFEL supports real-time data analysis via a cross-facility workflow spanning LCLS and HPC centers such as NERSC and OLCF. Our work therefore constitutes initial path-finding for the US Department of Energy's (DOE) Integrated Research Infrastructure (IRI) program. We present the ExaFEL team's 7 years of experience in developing real-time XFEL data analysis software for the DOE's exascale supercomputers. We present our experiences and lessons learned with the Perlmutter and Frontier supercomputers. Furthermore we outline essential data center services (and the implications for institutional policy) required for real-time data analysis. Finally we summarize our software and performance engineering approaches and our experiences with NERSC's Perlmutter and OLCF's Frontier systems. This work is intended to be a practical blueprint for similar efforts in integrating exascale compute resources into other cross-facility workflows.

59 BASIC BIOLOGICAL SCIENCES↗

Inference-Optimized AI and High Performance Computing for Gravitational Wave Detection at Scale

We introduce an ensemble of artificial intelligence models for gravitational wave detection that we trained in the Summit supercomputer using 32 nodes, equivalent to 192 NVIDIA V100 GPUs, within 2 h. Once fully trained, we optimized these models for accelerated inference using NVIDIA TensorRT. We deployed our inference-optimized AI ensemble in the ThetaGPU supercomputer at Argonne Leadership Computer Facility to conduct distributed inference. Using the entire ThetaGPU supercomputer, consisting of 20 nodes each of which has 8 NVIDIA A100 Tensor Core GPUs and 2 AMD Rome CPUs, our NVIDIA TensorRT-optimized AI ensemble processed an entire month of advanced LIGO data (including Hanford and Livingston data streams) within 50 s. Our inference-optimized AI ensemble retains the same sensitivity of traditional AI models, namely, it identifies all known binary black hole mergers previously identified in this advanced LIGO dataset and reports no misclassifications, while also providing a 3X inference speedup compared to traditional artificial intelligence models. We used time slides to quantify the performance of our AI ensemble to process up to 5 years worth of advanced LIGO data. In this synthetically enhanced dataset, our AI ensemble reports an average of one misclassification for every month of searched advanced LIGO data. We also present the receiver operating characteristic curve of our AI ensemble using this 5 year long advanced LIGO dataset. This approach provides the required tools to conduct accelerated, AI-driven gravitational wave detection at scale.

97 MATHEMATICS AND COMPUTING↗

Hyperspectral Aquatic Radiative Transfer Modeling Using a High-Performance Cluster Computing-Based Approach

Abstract For aquatic studies, radiative transfer (RT) modeling can be used to compute hyperspectral above-surface remote sensing reflectance that can be utilized for inverse model development. Inverse models can provide bathymetry and inherent-and bottom-optical property estimation. Because measured oceanic field/organic datasets are often spatio-temporally sparse, synthetic data generation is useful in yielding sufficiently large datasets for inversion model development; however, these forward-modeled data are computationally expensive and time-consuming to generate. This study establishes the magnitude of wall-clock-time savings achieved for performing large, aquatic RT batch-runs using parallel computing versus a sequential approach. Given 2,600 simulations and identical compute-node characteristics, sequential architecture required ~100 hours until termination, whereas a parallel approach required only ~2.5 hours (42 compute nodes) a 40x speed-up. Tools developed for this parallel execution are discussed.

58 GEOSCIENCES↗

Artificial Intelligence/Machine Learning Technologies for Advanced Reactors (Workshop Summary Report)

A workshop on artificial intelligence and machine learning (AI/ML) for advanced reactors (AR) was held October 5-6, 2021. The workshop was to be attended in-person at ANL but COVID restrictions forced the workshop to go virtual. The objectives of the workshop were to identify the most promising AI/ML opportunities for improving advanced reactor design, optimizing plant performance, and enhancing economic competitiveness and to develop an understanding of the scientific, engineering and licensing challenges facing their application. The workshop planning committee included GAIN, EPRI and NEI and members of three national laboratories (ANL, INL, and ORNL). The workshop was attended by more than 200 individuals representing academic and scientific institutions and the nuclear power industry. The definition put forth for an AI/ML system was one that perceives its environment and takes actions that maximize its chance of achieving its goals. In this report AI/ML refers to next generation algorithms that include deep learning, statistical analysis and data analytics and associated scientific computing and their potential application to the design, licensing, operation and maintenance of ARs. These methods typically incorporate models built from process data and may also include data generated by simulations that represent the behavior of a system. The workshop was organized in response to the growing interest in application of AI/ML for improving the economic competitiveness of nuclear energy. Increasingly more resources are being allocated to investigating the benefits of AI/ML methods. The DOE created the Artificial Intelligence & Technology Office to promote their development. And within the Office of Nuclear Energy, resources have been allocated to explore and understand the potential benefits of AI/ML. Additionally, the national laboratories are strategically positioned with DOE computing facilities such as Summit, Perlmutter, Aurora and Frontier that support large-scale simulations, hybrid HPC models with AI surrogates, and the exploration of new types of generative models emerging from multi-model data streams and sources. The workshop was organized with members of the AR community to understand the effort and to identify the level of interest and progress in this emerging technology. The workshop discussions focused on identifying opportunities for AI/ML across diverse areas of the nuclear industry and identifying current scientific and engineering challenges for advanced reactors that might be addressed through transformational uses of AI/ML. Discussion panels focused on four high-interest technical domains for advanced reactors: design, maintenance and operations, energy storage, and materials. The results of those discussions are summarized in this report. This includes opportunities that were identified for exploiting AI techniques and methods to improve the efficacy and efficiency of reactor analysis and to improve the operation and optimization of advanced reactors. Advanced reactor developers expressed an interest in learning more about AI/ML methods and their application. This included understanding whether ML methods can provide an advantage over existing nonlinear data regression methods for collapsing high-fidelity simulation results into faster running models. A consensus emerged that AR advances planned for the next decade will benefit from the use of AI/ML tools. The need exists to understand and model complex systems across length scales and modalities. AI/ML is a tool for discovery that can yield a set of engineering principles for use by nuclear engineers, licensing bodies, and operators to solve problems in plant design, safety analyses, autonomous operation, and predictive maintenance. While AI/ML represents a new set of tools, an awareness by the nuclear community of the full potential is still in the early stages so there is a need to increase awareness. It appears that the wide-spread adoption of AI/ML tools for ARs would be facilitated by future educational workshops that describe foundational methods and capabilities and describe successful applications.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Challenges and Opportunities of User-Level File Systemsfor HPC

The performance gap between magnetic disks and data processing on HPC systems has become that huge that an efficient data processing can only be achieved by introducing non-volatile memory (NVRAM) as a new storage tier. Although the benefits of hierarchical storage have been adequately demonstrated to the point that the newest leadership class HPC systems will employ burst buffers, critical questions remain for supporting hierarchical storage systems, including: How should we present hierarchical storage systems to user applications, such that they are easy to use and that application code is portable across systems? How should we manage data movement through a storage hierarchy for best performance and resilience of data? How do the particular I/O use cases mandate the way we manage data? There have been many efforts to explore this space in the form of file systems, with increasingly more implemented at the user level. This is because it is relatively easy to swap in new, specialized user-level file systems for use by applications on a case-by-case basis, as opposed to the current mainstream approach of using general-purpose, system-level file systems which may not be optimized for HPC workloads and must be installed by administrators. In contrast, file systems at the user level can be tailored for specific HPC workloads for high performance and can be used by applications without administrator intervention. Many user-level file system developers have found themselves “having to reinvent the wheel” to implement various optimizations in their file systems. Thus, a main goal of this meeting was to bring together experts in I/O performance, file systems, and storage, and collectively explore the space of current and future problems and solutions for I/O on hierarchical storage systems in order to begin a community effort in enabling user-level file system support for HPC systems. We had a lively week of learning about each other’s approaches as well as unique I/O use cases that can influence the design of a community-driven file and storage system standards. The agenda for this meeting contained talks from participants on the following high level topics: HPC storage and I/O support today; what do HPC users need for I/O; existing userlevel file system efforts; object stores and other alternative storage systems; and components for building user-level file systems. The talks were short and intended to be conversation starters for more in-depth discussions with the whole group. The participants engaged in lengthy discussions on various questions that arose from the talks including: Are we ready to program to a memory hierarchy versus block devices? Are the needs of HPC users reflected in our existing file systems and storage systems? Should we drop or keep POSIX moving forward? What do we mean when we say “user-level file system”? Do we all mean the same thing? How should the IO 500 benchmark be defined so it is fair and useful? and How are stage-in and stage-out actually going to work? The report for this seminar contains a record of the talks from the participants as well as the resulting discussions. Our hope is that the effort initiated during this seminar will result in long-term collaborations that will benefit the HPC community as a whole.

97 MATHEMATICS AND COMPUTING↗

Designing and prototyping extensions to the Message Passing Interface in MPICH

As HPC system architectures and the applications running on them continue to evolve, the MPI standard itself must evolve. The trend in current and future HPC systems toward powerful nodes with multiple CPU cores and multiple GPU accelerators makes efficient support for hybrid programming critical for applications to achieve high performance. However, the support for hybrid programming in the MPI standard has not kept up with recent trends. The MPICH implementation of MPI provides a platform for implementing and experimenting with new proposals and extensions to fill this gap and to gain valuable experience and feedback before the MPI Forum can consider them for standardization. Here, in this work, we detail six extensions implemented in MPICH to increase MPI interoperability with other runtimes, with a specific focus on heterogeneous architectures. First, the extension to MPI generalized requests lets applications integrate asynchronous tasks into MPI’s progress engine. Second, the iovec extension to datatypes lets applications use MPI datatypes as a general-purpose data layout API beyond just MPI communications. Third, a new MPI object, MPIX_Stream, can be used by applications to identify execution contexts beyond MPI processes, including threads and GPU streams. MPIX stream communicators can be created to make existing MPI functions thread-aware and GPU-aware, thus providing applications with explicit ways to achieve higher performance. Fourth, MPIX Streams are extended to support the enqueue semantics for offloading MPI communications onto a GPU stream context. Fifth, thread communicators allow MPI communicators to be constructed with individual threads, thus providing a new level of interoperability between MPI and on-node runtimes such as OpenMP. Lastly, we present an extension to invoke MPI progress, which lets users spawn progress threads with fine-grained control to adapt the communication performance to their application designs. We describe the design and implementation of these extensions, provide usage examples, and highlight their expected benefits with performance results.

97 MATHEMATICS AND COMPUTING↗

SMC 2021 : Analyzing Resource Utilization and User Behavior on Titan Supercomputer

Resource utilization statistics of submitted jobs on a supercomputer can help us understand how users from various scientific domains use HPC platforms and better design a job scheduler. We explore to generate insight regarding workload distribution and usage pattern domains from job scheduler trace, GPU failure information, and project-specific information collected from Titan supercomputer. Furthermore, we want to know how the scheduler performance varies over time and how the users' scheduling behavior changes following a system failure. These observations have the potential to provide valuable insight, which is helpful to prepare for system failures. These practices will help us develop and apply novel machine learning algorithms in understanding system behavior, requirement, and better scheduling of HPC systems. There are two datasets, RUR and GPU. RUR: This dataset is the job scheduler traces collected from the Titan supercomputerfrom 01/01/2015 to 07/31/2019 (2015.csv - 2019.csv). These were collected usingResource Utilization Report (RUR), a Cray-developed resource-usage data collectionand reporting system. It contains the usage information of its critical resources (CPU,Memory, GPU, and I/O) of each running job on Titan during that period [2]. ProjectAreas: Every job is associated with a project ID. TheProjectAreas.csvdatasetprovides a mapping of the project ID to its domain science. GPU: There have been some hardware-related issues in the GPUs in Titan that caused some GPUs to fail, sometimes irrecoverably during some job runs. This dataset provides information regarding these failures during the execution of the submitted jobs. GPUs on Titan are uniquely identified by a serial number (SN), and they are installed in a location. A GPU can be installed in a location, then removed from that location following a failure, and then re-installed in a different location after fixing the problem. If the failure can't be recovered, the GPU might be removed entirely from Titan. There are two prominent types of failures that resulted in the removal of GPUs from Titan: Double Bit Error (DBE) and Out of the Bus (OTB). The dataset (gc_full.csv) has the following fields: 1. SN : Serial number of a GPU 2. location : The location where it is installed 3. insert : The time when it was inserted into that location 4. remove : The time when it was removed from that location 5. duration : Amount of time the GPU spent in this location 6. out : If the device was taken out entirely w/o a re-installment into a new location. 7. event : If the GPU was taken out entirely, the reason for its removal. To learn more about this dataset, please refer to the git repositoryhttps://github.com/olcf/TitanGPULifeand the related publication [1]. References [1] George Ostrouchov, Don Maxwell, Rizwan A Ashraf, Christian Engelmann, MallikarjunShankar, and James H Rogers. Gpu lifetimes on titan supercomputer: Survival analysisand reliability. InSC20: International Conference for High Performance Computing,Networking, Storage and Analysis, pages 1-14. IEEE, 2020. [2] Feiyi Wang, Sarp Oral, Satyabrata Sen, and Neena Imam. Learning from five-yearresource-utilization data of titan system. In2019 IEEE International Conference onCluster Computing (CLUSTER), pages 1-6. IEEE, 2019.

42 ENGINEERING↗