Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Advanced Supercomputing”

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 325 records · Page 18

Building Simulation Modelers are we big-data ready?

Recent advances in computing and sensor technologies have pushed the amount of data we collect or generate to limits previously unheard of. Sub-minute resolution data from dozens of channels is becoming increasingly common and is expected to increase with the prevalence of non-intrusive load monitoring. Experts are running larger building simulation experiments and are faced with an increasingly complex data set to analyze and derive meaningful insight. This paper focuses on the data management challenges that building modeling experts may face in data collected from a large array of sensors, or generated from running a large number of building energy/performance simulations. The paper highlights the technical difficulties that were encountered and overcome in order to run 3.5 million EnergyPlus simulations on supercomputers and generating over 200 TBs of simulation output. This extreme case involved development of technologies and insights that will be beneficial to modelers in the immediate future. The paper discusses different database technologies (including relational databases, columnar storage, and schema-less Hadoop) in order to contrast the advantages and disadvantages of employing each for storage of EnergyPlus output. Scalability, analysis requirements, and the adaptability of these database technologies are discussed. Additionally, unique attributes of EnergyPlus output are highlighted which make data-entry non-trivial for multiple simulations. Practical experience regarding cost-effective strategies for big-data storage is provided. The paper also discusses network performance issues when transferring large amounts of data across a network to different computing devices. Practical issues involving lag, bandwidth, and methods for synchronizing or transferring logical portions of the data are presented. A cornerstone of big-data is its use for analytics; data is useless unless information can be meaningfully derived from it. In addition to technical aspects of managing big data, the paper details design of experiments in anticipation of large volumes of data. The cost of re-reading output into an analysis program is elaborated and analysis techniques that perform analysis in-situ with the simulations as they are run are discussed. The paper concludes with an example and elaboration of the tipping point where it becomes more expensive to store the output than re-running a set of simulations.

Sanyal, Jibonananda↗

Transforming Energy through Computational Science: Computing for Clean Energy

This fact sheet discusses the opportunity space for NREL's computational science capabilities to address national clean energy objectives. Achieving a carbon-free power sector by 2035 as a step towards a decarbonized U.S. energy economy in 2050 will require major advances in power generation, autonomous energy systems, transportation, and buildings/communities. Development of integrated modeling approaches for complex energy systems will be essential for deployment. Success requires developments in optimization and control theory, complemented by machine learning (ML) and artificial intelligence (AI), all of which in turn need targeted investments in breadth and scale of computing. This document defines an opportunity space where: embracing computing can link established research and development (R&D) to a decarbonization agenda; pursuing emerging approaches can accelerate the pace of technology advancement across the portfolio; and leading by example could reduce the carbon footprint of computing worldwide.

advanced computing↗

HPC-FAIR: A Framework Managing Data and AI Models for Analyzing and Optimizing Scientific Applications

The increasing reliance on machine learning (ML) to analyze and optimize large-scale scientific applications on supercomputers faces a significant bottleneck: the lack of readily available, high-quality training datasets and the difficulty in reusing existing AI models. This project was motivated by the urgent need to address the “FAIR” principles (Findability, Accessibility, Interoperability, Reusability) for both training datasets and AI models in the high-performance computing (HPC) domain. The project developed HPC-FAIR, a high-performance computing data management framework designed to centralize HPC-related datasets and AI models within a unified hub. To ensure interoperability, the framework established a standardized representation and vocabulary (ontology) for both data and models. HPC-FAIR also implemented automated workflows to streamline data processing, model access, and benchmarking. Additionally, the project focused on optimizing data harnessing efficiency through advanced techniques like deep reuse and compression-based analytics.

97 MATHEMATICS AND COMPUTING↗

Enabling Large-Scale Condensed-Phase Hybrid Density Functional Theory Based Ab Initio Molecular Dynamics. 1. Theory, Algorithm, and Performance

By including a fraction of exact exchange (EXX), hybrid functionals reduce the self-interaction error in semilocal density functional theory (DFT) and thereby furnish a more accurate and reliable description of the underlying electronic structure in systems throughout biology, chemistry, physics, and materials science. However, the high computational cost associated with the evaluation of all required EXX quantities has limited the applicability of hybrid DFT in the treatment of large molecules and complex condensed-phase materials. To overcome this limitation, we describe a linear-scaling approach that utilizes a local representation of the occupied orbitals (e.g., maximally localized Wannier functions (MLWFs)) to exploit the sparsity in the real-space evaluation of the quantum mechanical exchange interaction in finite-gap systems. In this work, we present a detailed description of the theoretical and algorithmic advances required to perform MLWF-based ab initio molecular dynamics (AIMD) simulations of large-scale condensed-phase systems of interest at the hybrid DFT level. We focus our theoretical discussion on the integration of this approach into the framework of Car–Parrinello AIMD, and highlight the central role played by the MLWF-product potential (i.e., the solution of Poisson’s equation for each corresponding MLWF-product density) in the evaluation of the EXX energy and wave function forces. We then provide a comprehensive description of the exx algorithm implemented in the open-source Quantum ESPRESSO program, which employs a hybrid MPI/OpenMP parallelization scheme to efficiently utilize the high-performance computing (HPC) resources available on current- and next-generation supercomputer architectures. Furthermore, this is followed by a critical assessment of the accuracy and parallel performance (e.g., strong and weak scaling) of this approach when AIMD simulations of liquid water are performed in the canonical (NVT) ensemble. With access to HPC resources, we demonstrate that exx enables hybrid DFT-based AIMD simulations of condensed-phase systems containing 500–1000 atoms (e.g., (H₂O)₂₅₆) with a wall time cost that is comparable to that of semilocal DFT. In doing so, exx takes us one step closer to routinely performing AIMD simulations of complex and large-scale condensed-phase systems for sufficiently long time scales at the hybrid DFT level of theory.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Then and Now: Improving Software Portability, Productivity, and 100× Performance

The US Exascale Computing Project (ECP) has succeeded in preparing applications to run efficiently on the first reported Exascale supercomputers in the world. To achieve this, it modernized the whole leadership software stack, from libraries to simulation codes. In this article, we contrast selected leadership software before and after ECP. We discuss how sustainable research software development for leadership computing can embrace the conversation with the hardware vendors, the leadership computing facilities, the software community, and the domain scientists who are the application developers and integrators of software products. We elaborate on how software needs to take portability as a central design principle and to benefit from interdependent teams; we also demonstrate how moving to programming languages with high momentum, like modern C++, can help improve the sustainability, interoperability, and performance of research software. Finally, we showcase how cross-institutional efforts can enable algorithm advances that are beyond incremental performance optimization.

97 MATHEMATICS AND COMPUTING↗

Low-Overhead In Situ Visualization Using Halo Replay

In situ visualization and analysis is of increasing importance as the compute and I/O gap further widens with the advance to exascale capable computing. Yet, in situ methods impose resource constraints leading to the difficult task of balancing simulation code performance and the quality of analysis. Applications with tightly-coupled in situ visualization often achieve performance through spatial and temporal downsampling, a tradeoff which risks not capturing transient phenomena at sufficient fidelity. Determining a priori visualization parameters such as sampling rate is difficult without time and resource intensive experimentation. We present a method for reducing resource contention between in situ visualization and stencil codes on heterogeneous systems. This method permits full resolution replay through recording halos and the communication-free reconstruction of interior values uncoupled from the main simulation. We apply this method in the computational fluid dynamics (CFD) code HARVEY [1] on the Summit supercomputer. We demonstrate minimal-overhead, in situ visualization relative to simulation alone, and compare the Halo Replay performance to tightly-coupled in situ approaches.

Ames, Jeff↗

Full electromagnetic Vlasov code simulation of the Kelvin-Helmholtz instability

Recent advancement in numerical techniques for Vlasov simulations and their application to cross-scale coupling in the plasma universe are discussed. Magnetohydrodynamic (MHD) simulations are now widely used for numerical modeling of global and macroscopic phenomena. In the framework of the MHD approximation, however, diffusion coefficients such as resistivity and adiabatic index are given from empirical models. Thus there are recent attempts to understand first-principle kinetic processes in macroscopic phenomena, such as magnetic reconnection and the Kelvin-Helmholtz (KH) instability via full kinetic particle-in-cell and Vlasov codes. In the present study, a benchmark test for a new four-dimensional full electromagnetic Vlasov code is performed. First, the computational speed of the Vlasov code is measured and a linear performance scaling is obtained on a massively parallel supercomputer with more than 12 000 cores. Second, a first-principle Vlasov simulation of the KH instability is performed in order to evaluate current status of numerical techniques for Vlasov simulations. The KH instability is usually adopted as a benchmark test problem for guiding-center Vlasov codes, in which a cyclotron motion of charged particles is neglected. There is not any full electromagnetic Vlasov simulation of the KH instability; this is because it is difficult to follow E-vectorxB-vector drift motion accurately without approximations. The present first-principle Vlasov simulation has successfully represented the formation of KH vortices and its secondary instability. These results suggest that Vlasov code simulations would be a powerful approach for studies of cross-scale coupling on future Peta-scale supercomputers.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Single-Bubble Flow Boiling Phenomena - Interface Tracking Simulation

Boiling, as one of the most efficient heat transfer mechanisms, is widely used in various engineering systems. Better understanding and modeling of this process remains a major challenge in multiphase flow research. The distribution of vapor in boiling system affects the heat transfer rate and may cause unfavorable conditions, such as heater burn-out. A number of flow regimes have been observed experimentally. In general, there are three regimes under boiling conditions: nucleate boiling, transition boiling, and film boiling. The nucleate boiling regime is categorized into partial boiling regime and fully developed nucleate boiling, according to behavior of bubble dynamics and heat transfer characteristics. In the first region, as the wall temperature increases, the fraction of the wall surface subject to nucleate boiling increases until bubble formation occupies the entire heated surface. Heat transfer in this region is a complex mixture of single-phase forced convection and nucleate boiling. As bubble density increases, bubbles turn to coalesce and form insulating vapor patches, which impedes liquid flow back to the surface. The heat transfer rate under this condition reduces dramatically. This point is called Critical Heat Flux (CHF), which leads to boiling crisis. The behavior of boiling crisis is dependent on local fluid conditions due to bubble growth and departure. Experiments are the ultimate tool for boiling model development and verification, but they still lack some capabilities needed for better understanding of the local mechanism of this phenomena. It's difficult to capture many important details at the pressure/temperature conditions of interest using experimental techniques, such as full three dimensional measurements of velocity and temperature distribution and estimates of heat flux partitioning during the boiling process. The interface tracking simulation (ITS) is one of the promising approaches to describe heat transfer of boiling phenomena and their underlying mechanisms. The advances of high performance computing (HPC) in recent years made it possible to apply DNS to a wide variety of adiabatic flows. Some work has been done to implement evaporation / condensation models as well. Lee and Nydahl performed a numerical simulation of bubble growth and departure assuming that the bubble has the shape of a truncated sphere and a micro-layer exists during bubble growth. A 2-D cylindrical bubbles simulation has been done by Son with a simplified model to predict the heat flux in a thin liquid micro-layer. A direct simulation of 2D film boiling is conducted by Juric and Tryggvason. The film boiling, below critical pressure with a modified Level Set Method, is studied by Son and Dhir. Ose and Kunugi have developed a boiling and condensation model on subcooled boiling phenomena. Fuchs performed numerical simulations of bubble growth and departure considering wall conduction effects. Fully transient heat and fluid flow is modeled with a free surface of a rising bubble and a periodic calculation of the whole cycling of a growing, detaching, and rising bubble. Mukherjee and Kandikar have done a numerical simulation of an evaporating meniscus on a moving heated surface with level set method. The heat transfer associated with the advancing and receding contact angle has been studied. To develop the capability of physics-based boiling process, in the presented work the energy and momentum equations are coupled to simulate single-bubble growth scenarios. For a single bubble simulation in infinite liquid, the analytical solution exists and can be compared with the numerical results for verification purposes. The presented capabilities will be utilized in the large scale DNS of boiling flows in the future. The finite-element based code, PHASTA is used for the simulation. It is a parallel, hierarchic, higher-order accurate, adaptive, stabilized, transient analysis flow solver, which has been shown to be an effective tool for a wide range of single and two-phase flow problems. It has been shown to be highly scalable on top supercomputers ({approx}85% scaling on up to 3 million mesh partitions using a 92 billion element mesh . PHASTA has been shown to reliably predict the details of adiabatic bubbly flows as well as single-phase flows with heat transfer. To develop the capability of physics-based boiling process, in the presented work the energy and momentum equations are coupled to simulate a single-bubble growth scenario. For a single bubble growth in superheated water, the analytical solution exists and can be compared with the numerical results for verification purposes. The second-order accurate spatial and time discretization is used for the flow solver in the presented work. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

LLNL Partners with IBM on Brain-Like Computing Chip

Lawrence Livermore National Laboratory (LLNL) will receive a first-of-a-kind brain-inspired supercomputing platform for deep learning developed by IBM Research. Based on a breakthrough neurosynaptic computer chip called IBM TrueNorth, the scalable platform will process the equivalent of 16 million neurons and 4 billion synapses and consume the energy equivalent of a hearing aid battery – a mere 2.5 watts of power. The brain-like, neural network design of the IBM Neuromorphic System is able to infer complex cognitive tasks such as pattern recognition and integrated sensory processing far more efficiently than conventional chips.

97 MATHEMATICS AND COMPUTING↗

Preparing MPICH for exascale

The advent of exascale supercomputers heralds a new era of scientific discovery, yet it introduces significant architectural challenges that must be overcome for MPI applications to fully exploit its potential. Among these challenges is the adoption of heterogeneous architectures, particularly the integration of GPUs to accelerate computation. Additionally, the complexity of multithreaded programming models has also become a critical factor in achieving performance at scale. The efficient utilization of hardware acceleration for communication, provided by modern NICs, is also essential for achieving low latency and high throughput communication in such complex systems. In response to these challenges, the MPICH library, a high-performance and widely used Message Passing Interface (MPI) implementation, has undergone significant enhancements. Here, this paper presents four major contributions that prepare MPICH for the exascale transition. First, we describe a lightweight communication stack that leverages the advanced features of modern NICs to maximize hardware acceleration. Second, our work showcases a highly scalable multithreaded communication model that addresses the complexities of concurrent environments. Third, we introduce GPU-aware communication capabilities that optimize data movement in GPU-integrated systems. Finally, we present a new datatype engine aimed at accelerating the use of MPI derived datatypes on GPUs. These improvements in the MPICH library not only address the immediate needs of exascale computing architectures but also set a foundation for exploiting future innovations in high-performance computing. By embracing these new designs and approaches, MPICH-derived libraries from HPE Cray and Intel were able to achieve real exascale performance on OLCF Frontier and ALCF Aurora respectively.

Guo, Yanfei [Argonne National Laboratory (ANL), Ar↗

Enhancing Network Visibility and Security through Tensor Analysis

The increasing size, variety, rate of growth and change, and complexity of network data has warranted advanced network analysis and services. Tools that provide automated analysis through traditional or advanced signature-based systems or machine learning classifiers suffer from practical difficulties. These tools fail to provide comprehensive and contextual insights into the network when put to practical use in operational cyber security. In this paper, we present an effective tool for network security and traffic analysis that uses high-performance data analytics based on a class of unsupervised learning algorithms called tensor decompositions. The tool aims to provide a scalable analysis of the network traffic data and also reduce the cognitive load of network analysts and be network-expert-friendly by presenting clear and actionable insights into the network. With this being said, In this paper, we demonstrate the successful use of the tool in two completely diverse operational cyber security environments, namely, (1) security operations center (SOC) for the SCinet network at the SuperComputing (SC) Conference in 2016 and 2017 and (2) Reservoir Labs’ Local Area Network (LAN). In each of these environments, we produce actionable results for cyber security specialists including (but not limited to) (1) finding malicious network traffic involving internal and external attackers using port scans, SSH brute forcing, and NTP amplification attacks, (2) uncovering obfuscated network threats such as data exfiltration using DNS port and using ICMP traffic, and (3) finding network misconfiguration and performance degradation patterns.

97 MATHEMATICS AND COMPUTING↗

Enabling Low-Temperature (LTP) Ignition Technologies for Multi-Mode Engines through the Development of a Validated High-Fidelity LTP Model for Predicative Simulations Tools

The goal of multi-mode engine architectures is to extend current lean-burn dilution limits with renewable fuels, which requires spark plugs to deposit high energies (hundreds of mJ) in order to initiate ignition and complete combustion. At elevated energy deposition rates, spark plugs experience increased electrode erosion and thermal losses, which ultimately shortens the spark-plug lifetime and lowers ignition efficiency. As such, in order to safeguard the efficiency gains of multi-mode concepts, new and improved ignition technologies are required. Recently, non-equilibrium low-temperature plasmas (LTP) have been shown to promote energy-efficient ignition via quenching and transport of electronically excited atoms and molecules, selective radical production and fast heating of hydrocarbon/air mixtures [1-2]. Thus, LTP is seen as a technology that can potentially improve the energy extraction efficiency of fuels, while enabling kinetically controlled combustion modes towards fuel leaner conditions to realize current DOE VTO goals of improving the sustainability of future mobility [3]. Although many previous studies have demonstrated the efficacy of plasma-assisted ignition to enhance combustion, the detailed enhancement mechanisms remain largely unknown, especially for oxygenated fuels and at elevated pressures that are most relevant to practical engine conditions. These barriers hinder the development of accurate and comprehensive numerical models that seek to describe LTP-based ignition in existing engine design software tools and methods. Current state-of-the-art simulation capabilities for LTP ignition systems are in need of improvements since they deliver qualitative results only due to important limitations of existing approaches. Firstly, validated kinetic models with elementary steps for plasma discharges in oxygenated fuel/air mixtures of relevance to the transportation sector are required. Such kinetic models do not exist at present and will be developed and validated within this project. Secondly, plasma discharges and reactive mixture ignition are multi-scale, unsteady processes requiring high-performance numerical methods and software that execute efficiently on DOE supercomputers. Such software does not exist at present and will be developed and applied to practical LTP ignition scenarios as part of this project. Thirdly, experimental databases that are tailored to serve as benchmark in support of the development of predictive computational models of LTP ignition do not exist and will be part of this project.

33 ADVANCED PROPULSION SYSTEMS↗

Cooperative Research and Development Agreement With Georgetown University Report: National Institutes of Health, National Center for Advancing Translational Sciences Clinical and Translational Science Award

Oak Ridge National Laboratory (ORNL) is participating with Georgetown University (GU) as a subrecipient in response to the National Institutes of Health (NIH), National Center for Advancing Translational Sciences (NCATS) Clinical and Translational Science Award (CTSA) (U54 Clinical Trial Optional) funding opportunity announcement. This Cooperative Research and Development Agreement is put in to place to facilitate the development and implementation of clinical interventions that demonstrably improve human health is currently a complex, recursive, and inefficient process that leads to delays of years or decades before discoveries in biomedical research result in health benefits for patients and communities. NCATS conducts and supports research in the science of translation, to discover the mechanistic and operational principles of the intervention development and dissemination process, thereby providing the scientific foundation for improvements in translational efficiency that will accelerate the realization of interventions that improve human health. Under NCATS’ leadership, the CTSA Program supports a national network of medical research institutions called hubs. GU is the lead institution in one of the NIH hubs that was created as a result of a previous NIH CTSA. The missions of the GU have historically included the advancement of health through research in the clinical and biomedical sciences, the education of future leaders in medical and nursing practice and academia, and the provision of compassionate and scientifically competent patient care and service to the Washington, DC community and the nation. GU is the lead institution for the Georgetown-Howard Universities Center for Clinical and Translational Science (GHUCCTS), a multi-institutional partnership of medical research institutions forged from a desire to promote clinical research and translational science. Through multiple collaborations among these institutions, GHUCCTS is transforming clinical research and translational science in order to bring new scientific advances to health care. Oak Ridge National Laboratory is the Department of Energy's (DOE) largest science and energy laboratory. Managed since April 2000 by a partnership of the University of Tennessee and Battelle, ORNL was established in 1943 as a part of the secret Manhattan Project to pioneer a method for producing and separating plutonium. During the 1950s and 1960s, ORNL became an international center for the study of nuclear energy and related research in the physical and life sciences. With the creation of DOE in the 1970s, ORNL's mission broadened to include a variety of energy technologies and strategies. Today the laboratory supports the nation with a peacetime science and technology mission that is just as important as, but very different from, its role during the Manhattan Project. ORNL is home to the world's premier center for high performance supercomputing to enable scientific discovery. ORNL has extensive expertise in various areas of computer science that are uniquely situated to support GU. Additionally, ORNL’s leading computational user facilities present a unique opportunity to leverage the largest scale machines for open science in support of the stated mission of the NCATS CTSA. ORNL's partnership with GU will offer unparalleled opportunity in data analytics, deep-learning, artificial intelligence, and urban dynamics.

59 BASIC BIOLOGICAL SCIENCES↗

Scalable training of graph convolutional neural networks for fast and accurate predictions of HOMO-LUMO gap in molecules

Abstract Graph Convolutional Neural Network (GCNN) is a popular class of deep learning (DL) models in material science to predict material properties from the graph representation of molecular structures. Training an accurate and comprehensive GCNN surrogate for molecular design requires large-scale graph datasets and is usually a time-consuming process. Recent advances in GPUs and distributed computing open a path to reduce the computational cost for GCNN training effectively. However, efficient utilization of high performance computing (HPC) resources for training requires simultaneously optimizing large-scale data management and scalable stochastic batched optimization techniques. In this work, we focus on building GCNN models on HPC systems to predict material properties of millions of molecules. We use HydraGNN, our in-house library for large-scale GCNN training, leveraging distributed data parallelism in PyTorch. We use ADIOS, a high-performance data management framework for efficient storage and reading of large molecular graph data. We perform parallel training on two open-source large-scale graph datasets to build a GCNN predictor for an important quantum property known as the HOMO-LUMO gap. We measure the scalability, accuracy, and convergence of our approach on two DOE supercomputers: the Summit supercomputer at the Oak Ridge Leadership Computing Facility (OLCF) and the Perlmutter system at the National Energy Research Scientific Computing Center (NERSC). We present our experimental results with HydraGNN showing (i) reduction of data loading time up to 4.2 times compared with a conventional method and (ii) linear scaling performance for training up to 1024 GPUs on both Summit and Perlmutter.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Scale-up Unlearnable Examples Learning with High-performance Computing

Recent advancements in AI models, like ChatGPT, are structured to retain user interactions, which could inadvertently include sensitive healthcare data. In the healthcare field, particularly when radiologists use AI-driven diagnostic tools hosted on online platforms, there is a risk that medical imaging data may be repurposed for future AI training without explicit consent, spotlighting critical privacy and intellectual property concerns around healthcare data usage. Addressing these privacy challenges, a novel approach known as Unlearnable Examples (UEs) has been introduced, aiming to make data unlearnable to deep learning models. A prominent method within this area, called Unlearnable Clustering (UC), has shown improved UE performance with larger batch sizes but was previously limited by computational resources (e.g., a single workstation). To push the boundaries of UE performance with theoretically unlimited resources, we scaled up UC learning across various datasets using Distributed Data Parallel (DDP) training on the Summit supercomputer. Our goal was to examine UE efficacy at high-performance computing (HPC) levels to prevent unauthorized learning and enhance data security, particularly exploring the impact of batch size on UE’s unlearnability. Utilizing the robust computational capabilities of the Summit, extensive experiments were conducted on diverse datasets such as Pets, MedMNist, Flowers, and Flowers102. Our findings reveal that both overly large and overly small batch sizes can lead to performance instability and affect accuracy. However, the relationship between batch size and unlearnability varied across datasets, highlighting the necessity for tailored batch size strategies to achieve optimal data protection. The use of Summit’s high-performance GPUs, along with the efficiency of the DDP framework, facilitated rapid updates of model parameters and consistent training across nodes. Our results underscore the critical role of selecting appropriate batch sizes based on the specific characteristics of each dataset to prevent learning and ensure data security in deep learning applications. The source code is publicly available at https: // github. com/ hrlblab/ UE_ HPC .

Zhu, Yanfan [Vanderbilt University, Nashville, TN,↗

Revisiting Huffman Coding: Toward Extreme Performance on Modern GPU Architectures

Today's high-performance computing (HPC) applications are producing vast volumes of data, which are challenging to store and transfer efficiently during the execution, such that data compression is becoming a critical technique to mitigate the storage burden and data movement cost. Huffman coding is arguably the most efficient Entropy coding algorithm in information theory, such that it could be found as a fundamental step in many modern compression algorithms such as DEFLATE. On the other hand, today's HPC applications are more and more relying on the accelerators such as GPU on supercomputers, while Huffman encoding suffers from low throughput on GPUs, resulting in a significant bottleneck in the entire data processing. In this paper, we propose and implement an efficient Huffman encoding approach based on modern GPU architectures, which addresses two key challenges: (1) how to parallelize the entire Huffman encoding algorithm, including codebook construction, and (2) how to fully utilize the high memory-bandwidth feature of modern GPU architectures. The detailed contribution is fourfold. (1) We develop an efficient parallel codebook construction on GPUs that scales effectively with the number of input symbols. (2) We propose a novel reduction based encoding scheme that can efficiently merge the codewords on GPUs. (3) We optimize the overall GPU performance by leveraging the state-of-the-art CUDA APIs such as Cooperative Groups. (4) We evaluate our Huffman encoder thoroughly using six real-world application datasets on two advanced GPUs and compare with our implemented multi-threaded Huffman encoder. Experiments show that our solution can improve the encoding throughput by up to 5.0x and 6.8x on NVIDIA RTX 5000 and V100, respectively, over the state-of-the-art GPU Huffman encoder, and by up to 3.3x over the multi-thread encoder on two 28-core Xeon Platinum 8280 CPUs.

Tian, Jiannan↗

Revisiting Huffman Coding: Toward Extreme Performance on Modern GPU Architectures

Today's high-performance computing (HPC) applications are producing vast volumes of data, which are challenging to store and transfer efficiently during the execution, such that data compression is becoming a critical technique to mitigate the storage burden and data movement cost. Huffman coding is arguably the most efficient Entropy coding algorithm in information theory, such that it could be found as a fundamental step in many modern compression algorithms such as DEFLATE. On the other hand, today's HPC applications are more and more relying on the accelerators such as GPU on supercomputers, while Huffman encoding suffers from low throughput on GPUs, resulting in a significant bottleneck in the entire data processing. In this paper, we propose and implement an efficient Huffman encoding approach based on modern GPU architectures, which addresses two key challenges: (1) how to parallelize the entire Huffman encoding algorithm, including codebook construction, and (2) how to fully utilize the high memory-bandwidth feature of modern GPU architectures. The detailed contribution is fourfold. (1) We develop an efficient parallel codebook construction on GPUs that scales effectively with the number of input symbols. (2) We propose a novel reduction based encoding scheme that can efficiently merge the codewords on GPUs. (3) We optimize the overall GPU performance by leveraging the state-of-the-art CUDA APIs such as Cooperative Groups. (4) We evaluate our Huffman encoder thoroughly using six real-world application datasets on two advanced GPUs and compare with our implemented multithreaded Huffman encoder. Experiments show that our solution can improve the encoding throughput by up to 5.0× and 6.8× on NVIDIA RTX 5000 and V100, respectively, over the state-of-the-art GPU Huffman encoder, and by up to 3.3× over the multithread encoder on two 28-core Xeon Platinum 8280 CPUs.

Tian, Jiannan↗

ExM:System Support for Extreme-Scale, Many-Task Applications

The ever-increasing power of supercomputer systems is both driving and enabling the emergence of new problem-solving methods that require the effi cient execution of many concurrent and interacting tasks. Methodologies such as rational design (e.g., in materials science), uncertainty quanti fication (e.g., in engineering), parameter estimation (e.g., for chemical and nuclear potential functions, and in economic energy systems modeling), massive dynamic graph pruning (e.g., in phylogenetic searches), Monte-Carlo- based iterative fi xing (e.g., in protein structure prediction), and inverse modeling (e.g., in reservoir simulation) all have these requirements. These many-task applications frequently have aggregate computing needs that demand the fastest computers. For example, proposed next-generation climate model ensemble studies will involve 1,000 or more runs, each requiring 10,000 cores for a week, to characterize model sensitivity to initial condition and parameter uncertainty. The goal of the ExM project is to achieve the technical advances required to execute such many-task applications efficiently, reliably, and easily on petascale and exascale computers. In this way, we will open up extreme-scale computing to new problem solving methods and application classes. In this document, we report on combined technical progress of the collaborative ExM project, and the institutional financial status of the portion of the project at University of Chicago, over the rst 8 months (through April 30, 2011)

97 MATHEMATICS AND COMPUTING↗