Durango: Scalable Synthetic Workloard Generation for Extreme-Scale Application Performance Modeling & Simulation
Explore the source record for details and available documents.
SEARCH · Engineering Papers
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.
Explore the source record for details and available documents.
Supercomputers complexity continue to rapidly increase, requiring novel software solutions to aid in addressing the extreme abundance of parallelism, restrictive power constraints, and fault tolerance requirements. While industrial standards like MPI and OpenMP have undergone extensive efforts to prepare for future systems [1, 2], ne grain asynchronous runtimes have demonstrated an innate agility in extracting performance which is otherwise arduous to attain with current execution models
The Oak Ridge Leadership Computing Facility (OLCF) runs the No. 4 supercomputer in the world, supported by a petascale file system, to facilitate scientific discovery. In this paper, using the daily file system metadata snapshots collected over 500 days, we have studied the behavioral trends of 1, 362 active users and 380 projects across 35 science domains. In particular, we have analyzed both individual and collective behavior of users and projects, highlighting needs from individual communities and the overall requirements to operate the file system. We have analyzed the metadata across three dimensions, namely (i) the projects' file generation and usage trends, using quantitative file system-centric metrics, (ii) scientific user behavior on the file system, and (iii) the data sharing trends of users and projects. To the best of our knowledge, our work is the first of its kind to provide comprehensive insights on user behavior from multiple science domains through metadata analysis of a large-scale shared file system. We envision that this OLCF case study will provide valuable insights for the design, operation, and management of storage systems at scale, and also encourage other HPC centers to undertake similar such efforts.
Resilience is one of the key challenges in maintaining high efficiency of future extreme scale supercomputers. Researchers and system practitioners rely on field-data studies to understand reliability characteristics and plan for future HPC systems. In this work, we compare and contrast the reliability characteristics of multiple large-scale HPC production systems. Our study covers more than one billion compute node hours across five different systems over a period of 8 years. We confirm previous findings which continue to be valid, discover new findings, and discuss their implications.
Modern science is inundated with ever increasing data sizes as computational capabilities and image acquisition techniques continue to improve. For example, simulations are tackling ever larger domains with higher fidelity, and high-throughput microscopy techniques generate larger data that are fundamental to gather biologically and medically relevant insights. As the image sizes exceed memory, and even sometimes local disk space, each step in a scientific workflow is impacted. Current software solutions enable data exploration with limited interactivity for visualization and analytic tasks. Furthermore analysis on HPC systems often require complex hand-written parallel implementations of algorithms that suffer from poor portability and maintainability. We present a software infrastructure that simplifies end-to-end visualization and analysis of massive data. First, a hierarchical streaming data access layer enables interactive exploration of remote data, with fast data fetching to test analytics on subsets of the data. Second, a library simplifies the process of developing new analytics algorithms, allowing users to rapidly prototype new approaches and deploy them in an HPC setting. Third, a scalable runtime system automates mapping analysis algorithms to whatever computational hardware is available, reducing the complexity of developing scaling algorithms. Here, we demonstrate the usability and performance of our system using a use case from neuroscience: filtering, registration, and visualization of tera-scale microscopy data. We evaluate the performance of our system using a leadership-class supercomputer, Shaheen II.
Explore the source record for details and available documents.
Architectures are rapidly evolving, and exascale machines are expected to offer billion-way concurrency. We need to rethink algorithms, languages and programming models among other components in order to migrate large scale applications and explore parallelism on these machines. Although directive-based programming models allow programmers to worry less about programming and more about science, expressing complex parallel patterns in these models can be a daunting task especially when the goal is to match the performance that the hardware platforms can offer. One such pattern is wavefront. This paper extensively studies a wavefront-based miniapplication for Denovo, a production code for nuclear reactor modeling. We parallelize the Koch-Baker-Alcouffe (KBA) parallel-wavefront sweep algorithm in the main kernel of Minisweep (the miniapplication) using CUDA, OpenMP and OpenACC. Our OpenACC implementation running on NVIDIA's next-generation Volta GPU boasts an 85.06x speedup over serial code, which is larger than CUDA's 83.72x speedup over the same serial implementation. Our experimental platform includes SummitDev, an ORNL representative architecture of the upcoming Summit supercomputer. Our parallelization effort across platforms also motivated us to define an abstract parallelism model that is architecture independent, with a goal of creating software abstractions that can be used by applications employing the wavefront sweep motif.
Continuous integration, delivery, and deployment (CICD) is widely used in DevOps communities, as it allows for teams of all sizes to deploy rapidly-changing hardware and software resources quickly and confidently. In this paper, we will describe how University of Colorado Boulder Research Computing has adopted these practices on the RMACC Summit supercomputer to allow system engineers and researchers alike to capitalize on the benefits of CICD-centric development workflows. We will introduce the topic of CICD at a high level and describe how such practices can ease common software management challenges for High-Performance Computing (HPC) resources. We will then document the infrastructure deployed for Summit, and explain how software such as Jenkins and Singularity enabled adaptation for an HPC environment. We will conclude with two case studies discussing the use of our CICD infrastructure: one case study from the perspective of a system engineer maintaining user-facing resources, and the other case study from the perspective of a researcher developing, maintaining, and using the MFiX-Exa codebase.
The US Department of Energy deployed the Summit and Sierra supercomputers with the latest state-of-the-art network interconnect technology in 2018 and both systems entered production in 2019. In this paper, we provide an in-depth assessment of the systems' network interconnects that are based on Enhanced Data Rate (EDR) 100 Gb/s Mellanox InfiniBand. Both systems use second-generation EDR Host Channel Adapters (HCAs) and switches with several new features such as Adaptive Routing (AR), switch-based collectives, and HCA-based tag matching. Although based on the same components, Summit's network is "non-blocking" (i.e., a fully provisioned Clos network) and Sierra's network has a 2:1 taper between the racks and aggregation switches. We evaluate the two systems' interconnects using traditional communication benchmarks as well as production applications. We find that the new Adaptive Routing dramatically improves performance but the other new features still need improvement.
Explore the source record for details and available documents.
High-level programming languages such as Python are increasingly used to provide intuitive interfaces to libraries written in lower-level languages and for assembling applications from various components. This migration towards orchestration rather than implementation, coupled with the growing need for parallel computing (e.g., due to big data and the end of Moore's law), necessitates rethinking how parallelism is expressed in programs. Here, we present Parsl, a parallel scripting library that augments Python with simple, scalable, and flexible constructs for encoding parallelism. These constructs allow Parsl to construct a dynamic dependency graph of components that it can then execute efficiently on one or many processors. Parsl is designed for scalability, with an extensible set of executors tailored to different use cases, such as low-latency, high-throughput, or extreme-scale execution. We show, via experiments on the Blue Waters supercomputer, that Parsl executors can allow Python scripts to execute components with as little as 5 ms of overhead, scale to more than 250000 workers across more than 8000 nodes, and process upward of 1200 tasks per second. Other Parsl features simplify the construction and execution of composite programs by supporting elastic provisioning and scaling of infrastructure, fault-tolerant execution, and integrated wide-area data management. We show that these capabilities satisfy the needs of many-task, interactive, online, and machine learning applications in fields such as biology, cosmology, and materials science.
Supervised learning is a promising approach for modeling the performance of applications running on large HPC systems. A key assumption in supervised learning is that the training and testing data are obtained under the same conditions. However, in production HPC systems these conditions might not hold because the conditions of the platform can change over time as a result of hardware degradation, hardware replacement, software upgrade, and configuration updates. These changes could alter the data distribution in a way that affects the accuracy of the predictive performance models and render them less useful; this phenomenon is referred to as concept drift. Ignoring concept drift can lead to suboptimal resource usage and decreased efficiency when those performance models are deployed for tuning and job scheduling in production systems. To address this issue, we propose a concept-drift-aware predictive modeling approach that comprises two components: (1) an online Bayesian changepoint detection method that can automatically identify the location of events that lead to concept drift in near-real time and (2) a moment-matching transformation inspired by transfer learning that converts the training data collected before the drift to be useful for retraining. We use application input/output performance data collected on Cori, a production supercomputing system at the National Energy Research Scientific Computing Center, to demonstrate the effectiveness of our approach. The results show that concept-drift-aware models obtain significant improvement in accuracy; the median absolute error of the best-performing Gaussian process regression improved by 58.8% when the proposed approaches were used.
Universal globally adaptive load-balanced (UGAL) routing has been proposed for various interconnection networks and has been deployed in a number of current-generation supercomputers. Although UGAL-based schemes have been extensively studied, most existing results are based on either simulation or measurement. Without a theoretical understanding of UGAL, multiple questions remain: For which traffic patterns is UGAL most suited? Furthermore, what determines the performance of the UGAL-based scheme on a particular network configuration? Here, we develop a set of throughput models for UGALbased on linear programming. We show that the throughput models are valid across the torus, Dragonfly, and Slim Fly network topologies. Finally, we identify a robust model that can accurately and efficiently predict UGAL throughput for a set of representative traffic patterns across different topologies. Our models not only provide a mechanism to predict UGAL performance on large-scale interconnection networks but also reveal the inner working of UGAL and further our understanding of this type of routing.
The availability of large-area sensing, scalable algorithms, and high-performance computing has enabled the possibility of urban-scale building energy modeling using new methods not limited to the scalability of manual building energy model creation or retrieval of county-by-county tax assessor’s data. Automatic Building detection and Energy Model creation (AutoBEM) has created 178,368 building energy models for the Electric Power Board (EPB) of Chattanooga, TN, and compared simulation performance to 15-minute data from each building. These models leverage several data sources (e.g. imagery, GIS, utility), software tools to extract building properties (e.g. footprint, height, façade type, window-to-wall ratio, occupancy, building type), simulate at scale on two of the world’s #1 fastest supercomputers, and provide web-based visual analytics.Grid-interactive efficient buildings offer the potential to reduce utility and rate-payer energy costs during each calendar month’s hour of critical generation – when the least efficient, most costly, and often dirtiest generation assets must be brought online. In this paper, EnergyPlus is used to simulate over 150,000 buildings to assess the technical potential of utility-controlled smart thermostats. This is analyzed under a couple scenarios leveraging buildings as thermal batteries via pre-conditioning to coast through peak hours. Results are provided in box and whisker plots assessing the distribution of peak demand reduction at the utility-scale along with breakouts of energy and demand savings by building type and vintage.
There are generally three types of scientific software users: users that solve problems using existing science software tools, researchers that explore new approaches by extending existing code, and educators that teach students scientific concepts. Python is a general-purpose programming language that is accessible to beginners, such as students, but also as a language that has a rich scientific programming ecosystem that facilitates writing research software. Additionally, as high-performance computing (HPC) resources become more readily available, software support for parallel processing becomes more relevant to scientific software.There currently are no Python-based evolutionary computation frameworks that support all three types of scientific software users. Moreover, some support synchronous concurrent fitness evaluation that do not efficiently use HPC resources. We pose here a new Python-based EC framework that uses an established generalized unified approach to EA concepts to provide an easy to use toolkit for users wishing to use an EA to solve a problem, for researchers to implement novel approaches, and for providing a low-bar to entry to EA concepts for students. Additionally, this toolkit provides a scalable asynchronous fitness evaluation implementation friendly to HPC that has been vetted on hardware ranging from laptops to the world’s fastest supercomputer, Summit.
The central importance of large-scale eigenvalue problems in scientific computation necessitates the development of massively parallel algorithms for their solution. Recent advances in dense numerical linear algebra have enabled the routine treatment of eigenvalue problems with dimensions on the order of hundreds of thousands on the world’s largest supercomputers. In cases where dense treatments are not feasible, Krylov subspace methods offer an attractive alternative due to the fact that they do not require storage of the problem matrices. However, demonstration of scalability of either of these classes of eigenvalue algorithms on computing architectures capable of expressing massive parallelism is non-trivial due to communication requirements and serial bottlenecks, respectively. In this work, we introduce the SISLICE method: a parallel shift-invert algorithm for the solution of the symmetric self-consistent field (SCF) eigenvalue problem. The SISLICE method drastically reduces the communication requirement of current parallel shift-invert eigenvalue algorithms through various shift selection and migration techniques based on density of states estimation and k-means clustering, respectively. This work demonstrates the robustness and parallel performance of the SISLICE method on a representative set of SCF eigenvalue problems and outlines research directions that will be explored in future work.
Emerging quantum processors provide an opportunity to explore new approaches for solving traditional problems in the post Moore’s law supercomputing era. However, the limited number of qubits makes it infeasible to tackle massive real-world datasets directly in the near future, leading to new challenges in utilizing these quantum processors for practical purposes. Furthermore, hybrid quantum-classical algorithms that leverage both quantum and classical types of devices are considered as one of the main strategies to apply quantum computing to large-scale problems. In this article, we advocate the use of multilevel frameworks for combinatorial optimization as a promising general paradigm for designing hybrid quantum-classical algorithms. To demonstrate this approach, we apply this method to two well-known combinatorial optimization problems, namely, the Graph Partitioning Problem, and the Community Detection Problem. We develop hybrid multilevel solvers with quantum local search on D-Wave’s quantum annealer and IBM’s gate-model based quantum processor. We carry out experiments on graphs that are orders of magnitude larger than the current quantum hardware size, and we observe results comparable to state-of-the-art solvers in terms of quality of the solution.
Mixed Integer Programming (MIP) is a powerful abstraction in combinatorial optimization that finds real-life application across many significant sectors. The recent proliferation of graphical processing unit (GPU)-based accelerated computing architectures in large-scale parallel computing or supercomputing presents new opportunities as well as challenges in the advancement of MIP solver technology to effectively use the new accelerated computing platforms and scale to large parallel systems. Here, we recount the conventional processor-based strategies and focus on configurations where the most promising intersection lies between parallel MIP solver approaches and the specific strengths of accelerated parallel platforms. We note that the best potential lies in solving problems whose individual matrix sizes (of the linear program relaxation) fit entirely within one accelerator's memory and whose branch-and-bound (or branch-and-cut) trees cannot be fully contained within a small number of computational nodes. Additionally, we identify ideal features of computational linear algebra support on GPU accelerators that would help advance this direction of scalable parallel solution of MIP problems on GPU-based accelerated computing architectures.