Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Heterogeneous architectures”

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 127 records · Page 7

A High-Performance Design for Hierarchical Parallelism in the QMCPACK Monte Carlo code

We introduce a new high-performance design for parallelism within the Quantum Monte Carlo code QMCPACK. We demonstrate that the new design is better able to exploit the hierarchical parallelism of heterogeneous architectures compared to the previous GPU implementation. The new version is able to achieve higher GPU occupancy via the new concept of crowds of Monte Carlo walkers, and by enabling more host CPU threads to effectively offload to the GPU. The higher performance is expected to be achieved independent of the underlying hardware, significantly improving developer productivity and reducing code maintenance costs. Scientific productivity is also improved with full support for fallback to CPU execution when GPU implementations are not available or CPU execution is more optimal.

Luo, Ye↗

Efficient Implementation of Many-body Quantum Chemical Methods on the Intel Xeon Phi Coprocessor

This paper presents the implementation and performance of the highly accurate CCSD(T) quantum chemistry method on the Intel Xeon Phi coprocessor within the context of the NWChem computational chemistry package. The widespread use of highly correlated methods in electronic structure calculations is contingent upon the interplay between advances in theory and the possibility of utilizing the ever-growing computer power of emerging heterogeneous architectures. We discuss the design decisions of our implementation as well as the optimizations applied to the compute kernels and data transfers between host and coprocessor. We show the feasibility of adopting the Intel Many Integrated Core Architecture and the Intel Xeon Phi coprocessor for developing efficient computational chemistry modeling tools. Remarkable scalability is demonstrated by benchmarks. Our solution scales up to a total of 62560 cores with the concurrent utilization of Intel Xeon processors and Intel Xeon Phi coprocessors.

Environmental Molecular Sciences Laboratory↗

DeepWelding: a Deep Learning Enhanced Approach to GTAW Using Multi-source Sensing Images

Deep learning has great potential to reshape manufacturing industries. In this paper, we present DeepWelding, a novel framework that applies deep learning techniques to improve gas tungsten arc welding (GTAW) process monitoring and penetration detection using multi-source sensing images. The framework is capable of analyzing multiple types of optical sensing images synchronously and consists of three deep learning enhanced consecutive phases: image preprocessing, image selection, and weld penetration classification. Specifically, we adopted generative adversarial networks (pix2pix) for image denoising and classic convolutional neural networks (AlexNet) for image selection. Both pix2pix and AlexNet delivered satisfactory performance. However, five individual neural networks with heterogeneous architectures demonstrated inconsistent generalization capabilities in the classification phase when holding out multi-source images generated with specific experimental settings. Therefore, two ensemble methods combining multiple neural networks were designed to improve the model performance on unseen data collected from different experimental settings. In conclusion, we also found that the quality of model prediction was heavily influenced by the data stream collection environment. We think these findings are beneficial for the broad intelligent welding community.

42 ENGINEERING↗

A Block-Based Triangle Counting Algorithm on Heterogeneous Environments

Triangle counting is a fundamental building block in graph algorithms. In this article, we propose a block-based triangle counting algorithm to reduce data movement during both sequential and parallel execution. Our block-based formulation makes the algorithm naturally suitable for heterogeneous architectures. The problem of partitioning the adjacency matrix of a graph is well-studied. Our task decomposition goes one step further: it partitions the set of triangles in the graph. By streaming these small tasks to compute resources, we can solve problems that do not fit on a device. We demonstrate the effectiveness of our approach by providing an implementation on a compute node with multiple sockets, cores and GPUs. The current state-of-the-art in triangle enumeration processes the Friendster graph in 2.1 seconds, not including data copy time between CPU and GPU. Using that metric, our approach is 20 percent faster. When copy times are included, our algorithm takes 3.2 seconds. This is 5.6 times faster than the fastest published CPU-only time.

97 MATHEMATICS AND COMPUTING↗

Generalizable coordination of large multiscale workflows: challenges and learnings at scale

The advancement of machine learning techniques and the heterogeneous architectures of most current supercomputers are propelling the demand for large multiscale simulations that can automatically and autonomously couple diverse components and map them to relevant resources to solve complex problems at multiple scales. Nevertheless, despite the recent progress in workflow technologies, current capabilities are limited to coupling two scales. In the first-ever demonstration of using three scales of resolution, we present a scalable and generalizable framework that couples pairs of models using machine learning and in situ feedback. We expand upon the massively parallel Multiscale Machine-Learned Modeling Infrastructure (MuMMI), a recent, award-winning workflow, and generalize the framework beyond its original design. We discuss the challenges and learnings in executing a massive multiscale simulation campaign that utilized over 600,000 node hours on Summit and achieved more than 98% GPU occupancy for more than 83% of the time. We present innovations to enable several orders of magnitude scaling, including simultaneously coordinating 24,000 jobs, and managing several TBs of new data per day and over a billion files in total. Finally, we describe the generalizability of our framework and, with an upcoming open-source release, discuss how the presented framework may be used for new applications.

Bhatia, Harsh↗

Preliminary Study on Fine-Grained Power and Energy Measurements on Grace Hopper GH200 with Open-Source Performance Tools

The increasing adoption of tightly integrated, heterogeneous architectures, combined with the slowdown of Moore’s law, has made application power and energy-driven optimizations critical to efficiently use high-performance computing systems. This paper introduces a newly developed open-source toolkit that seamlessly integrates the Linux real-time hardware monitoring program hwmon with the Performance Application Programming Interface and the Score-P performance measurement system, thereby enabling fine-grained power and energy measurements for high-performance computing applications. Our primary target platform is the Wombat test bed, which is a system based on the NVIDIA GH200 superchip. The toolkit can capture transient power peaks with high temporal resolution (50 ms) and, thanks to Score-P integration, can map power metrics to specific code regions, thereby providing actionable information on power-intensive operations and inefficiencies. The toolkit also provides a holistic view of both the power and the energy consumption of the entire GH200 superchip by covering all major components: the Grace CPU, the Hopper GPU, and the I/O subsystem. Experiments that use Locally Self-consistent Multiple Scattering, which is an application for first-principles calculations of materials developed at Oak Ridge National Laboratory, have demonstrated the tool’s ability to identify transient power spikes and uncover opportunities for energy-aware optimizations. Additionally, we introduce a Python-based utility for converting Open Trace Format 2 traces to Parquet format, thus enabling advanced data analysis for numerical integration methods applied to power data for accurate energy profiling.

Hernandez Mendoza, Oscar [ORNL] (ORCID:00000002538↗

Scaling Uintah on the Aurora Exascale System up to 122,880 Intel Ponte Vecchio Xe Stacks

The challenge of being able to scale application codes based on the Asynchronous Many-Task (AMT) Uintah framework on the Department of Energy (DOE) Aurora exascale system is addressed in this work by considering a challenging Reverse Monte Carlo Ray Tracing radiation benchmark calculation. This benchmark involves potentially global all-to-all communication and uses adaptive mesh refinement and ray tracing to achieve scalability. This benchmark has been used as part of previous scalability studies on a number of pre-exascale systems and on the DOE Frontier exascale system. This paper describes steps taken to enable this benchmark to run successfully on up to 10,240 nodes and 122,880 Intel® Ponte Vecchio Xe stacks on the DOE Aurora exascale system. This scalability was achieved through a limited number of experiments on Aurora, given machine loads and its uniqueness. These experiments constitute valuable lessons learned to achieve scalability at this level. The resulting scalability runs, while few in number, demonstrate relatively good strong-scaling characteristics. A detailed analysis of these results provides important indications about the path to scalability on Aurora for future work. Overall, these results continue the remarkable ability of this AMT approach to produce scalable solutions for challenging problems at extreme scale on heterogeneous architectures.

Garcia, Marta [Argonne National Laboratory (ANL)]↗

Scoreboard

Emerging HPC machines have given rise to enhanced compute power that far outstrips the machine's ability to save large scale results for post-processing. To combat this, in situ data analysis techniques are slowly being adopted. With in situ data management favoring workflows composed of multiple simulations and analyses connected in transit on heterogeneous machines, scientists and engineers need a tool that enables them to create data extracts, visualizations, and interactively monitor and steer their simulations. Scoreboard Phase II is a next generation analysis software that supports composite in transit workflows on heterogeneous architectures and restores interactivity to in situ data analysis through simulation monitoring and computational steering. Scoreboard provides a simulation dashboard with graphs of metrics over time, controls for setting custom simulation steering parameters, controls for managing the set of data extracts being produced in the simulation, as well as the ability to explore data extracts, all from a web browser. Realizing the vision outlined in this project required research into making a system that integrates end to end from simulations all the way to the user. In situ tools generally suffer from complexity and excessive software dependencies. Scoreboard, by contrast, is easy to build and integrate into simulation codes and it provides first class FORTRAN support. The Scoreboard library is capable of in situ and in transit data analysis that can produce data extracts commonly needed for Computational Fluid Dynamics (CFD) analysis. Simulations can transparently stage data in transit to a Scoreboard Endpoint program, which can accept their data and produce the requested data extracts. This lets simulations return to their work while the Endpoint works on the analysis. Efficiently staging the data at scale was a topic of this research. Scoreboard provides the means to let the user manage data extracts and monitor/steer many simulations from a web browser. This area of the research focused on discovery of in transit network components to expose and control their steering parameters within an interactive browser-based user interface that includes: system topology, gathered metrics, notifications, dynamically-generated steering controls, and exploration of visualization data products.

Whitlock, BradJoseph [Intelligent Light] (00000001↗

MARBLES (Multi-scale Adaptively Refined Boltzmann LatticE Solver) [SWR-23-37]

MARBLES (Multi-scale Adaptively Refined Boltzmann LatticE Solver) is an open-source computational fluid dynamics package powered by the lattice Boltzmann equations and built on AMReX. In the lattice Boltzmann method, local collisions between meso-scale fictitious particles drive the governing equations which enables MARBLES to easily simulate flow around complex and/or moving geometry without the generation of a body-conforming mesh. Using AMReX data structures and operations ensures a high level of computational performance and parallel scaling on heterogenous architectures while also naturally supporting locally enhanced grid resolution and fidelity through automatic mesh refinement. New domains and problem definitions are easily specified through an input file with examples and guidance on all options and variables provided in the MARBLES documentation.

Henry de Frahan, Marc↗

Kokkos Programming Model and Library

High performance computing is in the midst of a many-core revolution driven by rapidly evolving microchips with diverse and heterogeneous architectures, increasing concurrency (thread count), and decreasing memory per thread. Our HPC applications face a major challenge to effectively use these microchips with a single code base that is performance portable across these evolving architectures. Kokkos was developed to enable on-chip concurrency and manage architecture-dependent constraints on memory access patterns. The goal is to maximize the amount of application code that is portable across diverse architectures and still obtain the same (or nearly) performance as a variant of that code written specifically for that architecture. The Kokkos team has ongoing R&D co-design collaborations with Sandia's and vendors' microchip and system engineers to rapidly incorporate evolving computer architecture features into Kokkos. The team is also providing leadership for DOE laboratories' participation on the ISO/C++ language standard committee to subsuming Kokkos features into a future standard.

97 MATHEMATICS AND COMPUTING↗

High Performance Computing Facility Operational Assessment 2015: Oak Ridge Leadership Computing Facility

Oak Ridge National Laboratory’s (ORNL’s) Leadership Computing Facility (OLCF) continues to surpass its operational target goals: supporting users; delivering fast, reliable systems; creating innovative solutions for high-performance computing (HPC) needs; and managing risks, safety, and security aspects associated with operating one of the most powerful computers in the world. The results can be seen in the cutting-edge science delivered by users and the praise from the research community. Calendar year (CY) 2015 was filled with outstanding operational results and accomplishments: a very high rating from users on overall satisfaction that ties the highest-ever mark set in CY 2014; the greatest number of core-hours delivered to research projects; the largest percentage of capability usage since the OLCF began tracking the metric in 2009; and success in delivering on the allocation of 60, 30, and 10% of core hours offered for the INCITE (Innovative and Novel Computational Impact on Theory and Experiment), ALCC (Advanced Scientific Computing Research Leadership Computing Challenge), and Director’s Discretionary programs, respectively. These accomplishments, coupled with the extremely high utilization rate, represent the fulfillment of the promise of Titan: maximum use by maximum-size simulations. The impact of all of these successes and more is reflected in the accomplishments of OLCF users, with publications this year in notable journals Nature, Nature Materials, Nature Chemistry, Nature Physics, Nature Climate Change, ACS Nano, Journal of the American Chemical Society, and Physical Review Letters, as well as many others. The achievements included in the 2015 OLCF Operational Assessment Report reflect first-ever or largest simulations in their communities; for example Titan enabled engineers in Los Angeles and the surrounding region to design and begin building improved critical infrastructure by enabling the highest-resolution Cybershake map for Southern California to date. The Titan system provides the largest extant heterogeneous architecture for computing and computational science. Usage is high, delivering on the promise of a system well-suited for capability simulations for science. This success is due in part to innovations in tracking and reporting the activity on the compute nodes, and using this information to further enable and optimize applications, extending and balancing workload across the entire node. The OLCF continues to invest in innovative processes, tools, and resources necessary to meet continuing user demand. The facility’s leadership in data analysis and workflows was featured at the Department of Energy (DOE) booth at SC15, for the second year in a row, highlighting work with researchers from the National Library of Medicine coupled with unique computational and data resources serving experimental and observational data across facilities. Effective operations of the OLCF play a key role in the scientific missions and accomplishments of its users. Building on the exemplary year of 2014, as shown by the 2014 Operational Assessment Report (OAR) review committee response in Appendix A, this OAR delineates the policies, procedures, and innovations implemented by the OLCF to continue delivering a multi-petaflop resource for cutting-edge research. This report covers CY 2015, which, unless otherwise specified, denotes January 1, 2015, through December 31, 2015.

97 MATHEMATICS AND COMPUTING↗

FY17 ASC P&EM L2 Milestone 6009: Demonstrate Thread Scalability within Aria on Both Sides of Trinity

The use of next-generation platforms (NGPs), also known as advanced technology systems (ATS), that incorporate many-core and heterogeneous architectures for scientific computing represent a tectonic shift in computing hardware design that will require massive development work within the Sierra applications to harness these systems to their full potential. The completion of this milestone represents a first step towards this effort by threading many of the computational kernels within Sierra/Aria.

97 MATHEMATICS AND COMPUTING↗

PIPER: Performance Insight for Programmers and Exascale Runtimes: Guiding the Development of the Exascale Software Stack

The PIPER project set out to develop methodologies and software for measurement, analysis, attribution, and presentation of performance data for extreme-scale systems. Goals of the project were to support analysis of massive multi-scale parallelism, heterogeneous architectures, multi-faceted performance concerns, and to support both post-mortem performance analysis to identify program features that contribute to problematic performance and on-line performance analysis to drive adaptation. This final report summarizes the research and development activity at Rice University as part of the PIPER project. Producing a complete suite of performance tools for exascale platforms during the course of this project was impossible since both hardware and software for exascale systems is still a moving target. For that reason, the project focused broadly on the development of new techniques for measurement and analysis of performance on modern parallel architectures, enhancements to HPCToolkit’s software infrastructure to support our research goals or use on sophisticated applications, engaging developers of multithreaded runtimes to explore how support for tools should be integrated into their designs, engaging operating system developers with feature requests for enhanced monitoring support, engaging vendors with requests that they add hardware measure- ment capabilities and software interfaces needed by tools as they design new components of HPC platforms including processors, accelerators and networks, and finally collaborations with partners interested in using HPCToolkit to analyze and tune scalable parallel applications.

97 MATHEMATICS AND COMPUTING↗

A Block-Based Triangle Counting Algorithm on Heterogeneous Environments

Triangle counting is a fundamental building block in graph algorithms. In this paper, we propose a block-based triangle counting algorithm to reduce data movement during both sequential and parallel execution. Our block-based formulation makes the algorithm naturally suitable for heterogeneous architectures. The problem of partitioning the adjacency matrix of a graph is well-studied. Our task decomposition goes one step further: it partitions the set of triangles in the graph. By streaming these small tasks to compute resources, we can solve problems that do not fit on a device. We demonstrate the effectiveness of our approach by providing an implementation on a compute node with multiple sockets, cores and GPUs. The current state-of-the-art in triangle enumeration processes the Friendster graph in 2.1 seconds, not including data copy time between CPU and GPU. Using that metric, our approach is 20 percent faster. When copy times are included, our algorithm takes 3.2 seconds. This is 5.6 times faster than the fastest published CPU-only time.

97 MATHEMATICS AND COMPUTING↗

Towards Use of Mixed Precision in ECP Math Libraries [Exascale Computing Project]

The use of multiple types of precision in mathematical software has the potential to increase its performance on new heterogeneous architectures. The xSDK project focuses both on the investigation and development of multiprecision algorithms as well as their inclusion into xSDK member libraries. This report summarizes current efforts on including and/or using mixed precision capabilities in the math libraries Ginkgo, heFFTe, hypre, MAGMA, PETSc/TAO, SLATE, SuperLU, and Trilinos, including KokkosKernels. It contains both numerical results from libraries that already provide mixed precision capabilities, as well as descriptions of the strategies to incorporate multiprecision into established libraries.

97 MATHEMATICS AND COMPUTING↗

Towards Use of Mixed Precision in ECP Math Libraries

The use of multiple types of precision in mathematical software has the potential to increase its performance on new heterogeneous architectures. The xSDK project focuses both on the investigation and development of multiprecision algorithms as well as their inclusion into xSDK member libraries. This report summarizes current efforts on including and/or using mixed precision capabilities in the math libraries Ginkgo, heFFTe, hypre, MAGMA, PETSc/TAO, SLATE, SuperLU, and Trilinos, including KokkosKernels. It contains both numerical results from libraries that already provide mixed precision capabilities, as well as descriptions of the strategies to incorporate multiprecision into established libraries.

97 MATHEMATICS AND COMPUTING↗

Chromium Tolerant, Highly Active and Stable Electrocatalytic Internal Surface Coating for Cathode of Commercial SOFCs (Final Report)

This project is aimed to develop a chromium (Cr) tolerant, highly active, and stable coating layer on the internal surfaces of the porous composite cathode from commercially available SOFCs. Such coating layer was developed using the additive manufacturing process of Atomic Layer Deposition (ALD) and has been applied on the cathode consisting of either an electronic conductor of LaxSr 1-x MnyO 3-δ (LSM) or mixed ionic and electronic conducting La x Sr 1-x Co y Fe 1-y O 3-δ (LSCF). PI's work has demonstrated that the internal surface of cathode from the commercial cells, can be further tailored using ALD coating to dramatically enhance the cell performance. For instance, ALD layer consisting heterostructured nano composite of nano-Pt and nano-(Mn 0.8 Co 0.2 ) 3 O 4 oxide on the internal surface of porous LSM/YSZ cathode from SOFCs, has resulted in the large reduction of the cell polarizations resistance by up to 53%, and enormous increase of power density over 370%. For the cells with LSCF/Sm 2 O 3 doped CeO 2 (SDC) cathode, the conformal layer of nano-composite consisting of superjacent CoOx and subjacent minimum amount of Pt nano-grains has resulted in the power density enhancement by 126% for the large scale industry tubular cells at 750°C, and both the performance enhancement and nanostructure of the ALD layer are stable over ~ 2000 h continuous operation performed at industry test station. In the meanwhile, those ALD coating layer developed by PI's work is also inherently Cr-tolerant, and could act as physical barrier for preventing Cr diffusion into the cathode backbone, so as to mitigate the Cr poisoning effect on the cathode. In this project, the impact of Cr on the performance of those ALD coated commercial cells has been evaluated. Based on evolution of the cell performance, the ALD coating layer chemistry and ALD coating layer thickness has been optimized to maximize the overall Cr tolerance, cell power density and cell longevity. Different ALD coating has been applied onto the internal surface of LSM/YSZ and LSCF/SDC backbone respectively. The architecture/scaffold structures on the internal surface of different cathode, designed by this project, was catalogued and analyzed using High Resolution Transmission Electron Microscopy (HRTEM), and cell power/durability performance are assured via comprehensive electrochemical performance testing in the industry operation relevant conditions. The impact of the electrochemical operation current density, the water humidity, the cell operation temperature, and cell operation duration on the Cr tolerance of ALD coated cells has been systematically investigated. There is completely different nanostructure degradation mechanisms between LSM and LSCF cells induced by Cr contamination. For the LSCF/SDC baseline cell, With the Cr source, there is no apparent Sr surface segregation phase even for the baseline cell operated for 3000 h at 750 °C. With the Cr source, there is significant amorphous (SrCr)Ox phase accumulated in the original pore region. For the commercial baseline cells, Cr contaminants on the LSM electrode severely impacted the entire cell's electrochemical performance and nanostructure degradation. Those degradations include (1). Peak power density loss of 64 % after 109 h of operation. The dramatic increase in Rp (2). They are cracking at LSM/SSZ interface, LSM grains. SSZ remains intact but with (CrMn)Ox. By contrast, ALD coating (MnCo)Ox/Pt dramatically improves the Cr resistance, as follows (1). ALD-coated cell with a power density is 280-380 % of the baseline cell, depending on the ALD layer thickness. (2). For a cell with a 20 nm thick ALD layer, there is a large performance enhancement (> 200 % power density) induced by ALD coating of Cr-tolerant Mn 0.8 Co 0.2 Ox. (3). For a cell with a 20 nm thick ALD layer, after 168 h at 750 °C power density of the ALD-coated cell is ~ 600% of that baseline cell upon operation with Cr contamination for 109 h. The ALD coating on the internal surface of cathode developed by this project integrated multi-functions. Those multi-functions include (1). Dramatically improving the cell power density for the commercial cells; (2). Dramatically improving contamination resistance of the cathode, for being an excellent protection coating layer sealing off Cr contamination. (3). Dramatically increasing the cell longevity by potentially preventing the microstructure evolution and grain coarsening of the cathode. Overall, this project will provide a simple solution to simultaneously enhance power density and increase the reliability, robustness, and endurance of commercial SOFCs, over the entire operating temperature range of 650-800 °C. For the inherently functional SOFC, the ALD coating of LSM based cathode mitigate the Cr-contamination. Power density of ALD-coated cell is ~ 600% of that baseline cell upon operation with Cr contamination. In addition to SOFCs, the novel on-demand design approach and creation of multifunctional heterogeneous architecture on the electrode surface presented in this work opens further research for their application in other types of fuel cells, batteries, and sensors for which electrochemical reactions on the surface are similarly critical.

36 MATERIALS SCIENCE↗