Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Parallel Supercomputers”

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 685 records · Page 38

A matrix-algebraic formulation of distributed-memory maximal cardinality matching algorithms in bipartite graphs

We describe parallel algorithms for computing maximal cardinality matching in a bipartite graph on distributed-memory systems. Unlike traditional algorithms that match one vertex at a time, our algorithms process many unmatched vertices simultaneously using a matrix-algebraic formulation of maximal matching. This generic matrix-algebraic framework is used to develop three efficient maximal matching algorithms with minimal changes. The newly developed algorithms have two benefits over existing graph-based algorithms. First, unlike existing parallel algorithms, cardinality of matching obtained by the new algorithms stays constant with increasing processor counts, which is important for predictable and reproducible performance. Second, relying on bulk-synchronous matrix operations, these algorithms expose a higher degree of parallelism on distributed-memory platforms than existing graph-based algorithms. We report high-performance implementations of three maximal matching algorithms using hybrid OpenMP-MPI and evaluate the performance of these algorithm using more than 35 real and randomly generated graphs. On real instances, our algorithms achieve up to 200 × speedup on 2048 cores of a Cray XC30 supercomputer. Even higher speedups are obtained on larger synthetically generated graphs where our algorithms show good scaling on up to 16,384 cores.

97 MATHEMATICS AND COMPUTING↗

Discrete Event Execution with One-Sided and Two-Sided GVT Algorithms on 216,000 Processor Cores

Global virtual time (GVT) computation is a key determinant of the efficiency and runtime dynamics of parallel discrete event simulations (PDES), especially on large-scale parallel platforms. Here, three execution modes of a generalized GVT computation algorithm are studied on high-performance parallel computing systems: (1) a synchronous GVT algorithm that affords ease of implementation, (2) an asynchronous GVT algorithm that is more complex to implement but can relieve blocking latencies, and (3) a variant of the asynchronous GVT algorithm to exploit one-sided communication in extant supercomputing platforms. Performance results are presented of implementations of these algorithms on up to 216,000 cores of a Cray XT5 system, exercised on a range of parameters: optimistic and conservative synchronization, fine- to medium-grained event computation, synthetic and non-synthetic applications, and different lookahead values. Performance of up to 54 billion events executed per second is registered. Detailed PDES-specific runtime metrics are presented to further the understanding of tightly-coupled discrete event dynamics on massively parallel platforms.

Perumalla, Kalyan S↗

Collaborating CPU and GPU for large-scale high-order CFD simulations with complex grids on the TianHe-1A supercomputer

Programming and optimizing complex, real-world CFD codes on current many-core accelerated HPC systems is very challenging, especially when collaborating CPUs and accelerators to fully tap the potential of heterogeneous systems. In this paper, with a tri-level hybrid and heterogeneous programming model using MPI + OpenMP + CUDA, we port and optimize our high-order multi-block structured CFD software HOSTA on the GPU-accelerated TianHe-1A supercomputer. HOSTA adopts two self-developed high-order compact definite difference schemes WCNS and HDCS that can simulate flows with complex geometries. We present a dual-level parallelization scheme for efficient multi-block computation on GPUs and perform particular kernel optimizations for high-order CFD schemes. The GPU-only approach achieves a speedup of about 1.3 when comparing one Tesla M2050 GPU with two Xeon X5670 CPUs. To achieve a greater speedup, we collaborate CPU and GPU for HOSTA instead of using a naive GPU-only approach. We present a novel scheme to balance the loads between the store-poor GPU and the store-rich CPU. Taking CPU and GPU load balance into account, we improve the maximum simulation problem size per TianHe-1A node for HOSTA by 2.3×, meanwhile the collaborative approach can improve the performance by around 45% compared to the GPU-only approach. Further, to scale HOSTA on TianHe-1A, we propose a gather/scatter optimization to minimize PCI-e data transfer times for ghost and singularity data of 3D grid blocks, and overlap the collaborative computation and communication as far as possible using some advanced CUDA and MPI features. Scalability tests show that HOSTA can achieve a parallel efficiency of above 60% on 1024 TianHe-1A nodes. With our method, we have successfully simulated an EET high-lift airfoil configuration containing 800M cells and China's large civil airplane configuration containing 150M cells. To our best knowledge, those are the largest-scale CPU–GPU collaborative simulations that solve realistic CFD problems with both complex configurations and high-order schemes.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Towards an Open (Data) Science Analytics-Hub for Reproducible Multi-Model Climate Analysis at Scale

Open Science is key to future scientific research and promotes a deep transformation in the whole scientific research process encouraging the adoption of transparent and collaborative scientific approaches aimed at knowledge sharing. Open Science is increasingly gaining attention in the current and future research agenda worldwide. To effectively address Open Science goals, besides Open Access to results and data, it is also paramount to provide tools or environments to support the whole research process, in particular the design, execution and sharing of transparent and reproducible experiments, including data provenance (or lineage) tracking. This work introduces the Climate Analytics-Hub, a new component on top of the Earth System Grid Federation (ESGF), which joins big data approaches and parallel computing paradigms to provide an Open Science environment for reproducible multi-model climate change data analytics experiments at scale. An operational implementation has been set up at the SuperComputing Centre of the Euro- Mediterranean Center on Climate Change, with the main goal of becoming a reference Open Science hub in the climate community regarding the multi-model analysis based on the Coupled Model Intercomparison Project (CMIP).

analytics-hub↗

End-to-end I/O portfolio for the summit supercomputing ecosystem

The I/O subsystem for the Summit supercomputer, No. 1 on the Top500 list, and its ecosystem of analysis platforms is composed of two distinct layers, namely the in-system layer and the center-wide parallel file system layer (PFS), Spider 3. The in-system layer uses node-local SSDs and provides 26.7 TB/s for reads, 9.7 TB/s for writes, and 4.6 billion IOPS to Summit. The Spider 3 PFS layer uses IBM's Spectrum Scale™ and provides 2.5 TB/s and 2.6 million IOPS to Summit and other systems. While deploying them as two distinct layers was operationally efficient, it also presented usability challenges in terms of multiple mount points and lack of transparency in data movement. To address these challenges, we have developed novel end-to-end I/O solutions for the concerted use of the two storage layers. We present the I/O subsystem architecture, the end-to-end I/O solution space, their design considerations and our deployment experience.

Oral, Sarp↗

High speed civil transport: Sonic boom softening and aerodynamic optimization

An improvement in sonic boom extrapolation techniques has been the desire of aerospace designers for years. This is because the linear acoustic theory developed in the 60's is incapable of predicting the nonlinear phenomenon of shock wave propagation. On the other hand, CFD techniques are too computationally expensive to employ on sonic boom problems. Therefore, this research focused on the development of a fast and accurate sonic boom extrapolation method that solves the Euler equations for axisymmetric flow. This new technique has brought the sonic boom extrapolation techniques up to the standards of the 90's. Parallel computing is a fast growing subject in the field of computer science because of its promising speed. A new optimizer (IIOWA) for the parallel computing environment has been developed and tested for aerodynamic drag minimization. This is a promising method for CFD optimization making use of the computational resources of workstations, which unlike supercomputers can spend most of their time idle. Finally, the OAW concept is attractive because of its overall theoretical performance. In order to fully understand the concept, a wind-tunnel model was built and is currently being tested at NASA Ames Research Center. The CFD calculations performed under this cooperative agreement helped to identify the problem of the flow separation, and also aided the design by optimizing the wing deflection for roll trim.

Cheung, Samson↗

Comparing High Performance Computing Accelerator Programming Models

Accelerator devices are becoming a norm in High Performance Computing (HPC). With more systems opting for heterogeneous architectures, portable programming models like OpenMP and OpenACC are becoming increasingly important. The SPEC ACCEL 1.2 benchmark suite consists of comparable benchmarks in OpenCL, OpenMP 4.5, and OpenACC 2.5 that can be used to evaluate the performance and support for programming models and frameworks on heterogeneous platforms. In this paper we go beneath the normative metric of performance times and look at the individual kernels to study the usage, strengths, and weaknesses of the two prevalent portable heterogeneous programming models, OpenMP and OpenACC. From our analysis we identify that benchmarks like MRI-Q, SP and BT have better performance using OpenACC, while benchmarks like MiniGhost, LBM and LBDC do consistently better with the OpenMP programming model across supercomputers like Titan, and Summit.We deep dive into the kernels of select four benchmarks to answer questions like: Where does the benchmark spend most of its cycles? What is the parallelization strategy used? Why is one programming model more performant than the other? By identifying the similarities and dierences we want to contrast between the benchmark implementation strategies in the SPEC ACCEL 1.2 benchmarks and provide more insights into the OpenMP and OpenACC programming models.

Pophale, Swaroop S.↗

Stable Parallel Training of Wasserstein Conditional Generative Adversarial Neural Networks : *Full/Regular Research Paper submission for the symposium CSCI-ISAI: Artificial Intelligence

We use a stable parallel approach to train Wasserstein Conditional Generative Adversarial Neural Networks (W-CGANs). The parallel training reduces the risk of mode collapse and enhances scalability by using multiple generators that are concurrently trained, each one of them focusing on a single data label. The use of the Wasserstein metric reduces the risk of cycling by stabilizing the training of each generator. We apply the approach on the CIFAR10 and the CIFAR100 datasets, two standard benchmark datasets with images of the same resolution, but different number of classes. Performance is assessed using the inception score, the Fréchet inception distance, and image quality. An improvement in inception score and Fréchet inception distance is shown in comparison to previous results obtained by performing the parallel approach on deep convolutional conditional generative adversarial neural networks (DC-CGANs). Weak scaling is attained on both datasets using up to 100 NVIDIA V100 GPUs on the OLCF supercomputer Summit.

Lupo Pasini, Massimiliano↗

Parallel performance optimizations on unstructured mesh-based simulations

This paper addresses two key parallelization challenges the unstructured mesh-based ocean modeling code, MPAS-Ocean, which uses a mesh based on Voronoi tessellations: (1) load imbalance across processes, and (2) unstructured data access patterns, that inhibit intra- and inter-node performance. Our work analyzes the load imbalance due to naive partitioning of the mesh, and develops methods to generate mesh partitioning with better load balance and reduced communication. Furthermore, we present methods that minimize both inter- and intranode data movement and maximize data reuse. Our techniques include predictive ordering of data elements for higher cache efficiency, as well as communication reduction approaches. We present detailed performance data when running on thousands of cores using the Cray XC30 supercomputer and show that our optimization strategies can exceed the original performance by over 2×. Additionally, many of these solutions can be broadly applied to a wide variety of unstructured grid-based computations.

97 MATHEMATICS AND COMPUTING↗

Computational Approaches to Simulation and Optimization of Global Aircraft Trajectories

This study examines three possible approaches to improving the speed in generating wind-optimal routes for air traffic at the national or global level. They are: (a) using the resources of a supercomputer, (b) running the computations on multiple commercially available computers and (c) implementing those same algorithms into NASAs Future ATM Concepts Evaluation Tool (FACET) and compares those to a standard implementation run on a single CPU. Wind-optimal aircraft trajectories are computed using global air traffic schedules. The run time and wait time on the supercomputer for trajectory optimization using various numbers of CPUs ranging from 80 to 10,240 units are compared with the total computational time for running the same computation on a single desktop computer and on multiple commercially available computers for potential computational enhancement through parallel processing on the computer clusters. This study also re-implements the trajectory optimization algorithm for further reduction of computational time through algorithm modifications and integrates that with FACET to facilitate the use of the new features which calculate time-optimal routes between worldwide airport pairs in a wind field for use with existing FACET applications. The implementations of trajectory optimization algorithms use MATLAB, Python, and Java programming languages. The performance evaluations are done by comparing their computational efficiencies and based on the potential application of optimized trajectories. The paper shows that in the absence of special privileges on a supercomputer, a cluster of commercially available computers provides a feasible approach for national and global air traffic system studies.

global air traffic optimization↗

Computational Approaches to Simulation and Optimization of Global Aircraft Trajectories

This study examines three possible approaches to improving the speed in generating wind-optimal routes for air traffic at the national or global level. They are: (a) using the resources of a supercomputer, (b) running the computations on multiple commercially available computers and (c) implementing those same algorithms into NASA’s Future ATM Concepts Evaluation Tool (FACET) and compares those to a standard implementation run on a single CPU. Wind-optimal aircraft trajectories are computed using global air traffic schedules. The run time and wait time on the supercomputer for trajectory optimization using various numbers of CPUs ranging from 80 to 10,240 units are compared with the total computational time for running the same computation on a single desktop computer and on multiple commercially available computers for potential computational enhancement through parallel processing on the computer clusters. This study also re-implements the trajectory optimization algorithm for further reduction of computational time through algorithm modifications and integrates that with FACET to facilitate the use of the new features which calculate time-optimal routes between worldwide airport pairs in a wind field for use with existing FACET applications. The implementations of trajectory optimization algorithms use MATLAB, Python, and Java programming languages. The performance evaluations are done by comparing their computational efficiencies and based on the potential application of optimized trajectories. The paper shows that in the absence of special privileges on a supercomputer, a cluster of commercially available computers provides a good option for computing wind-optimal trajectories for national and global air traffic system studies.

Ng, Hok K.↗

Evaluating the Potential of Cray Gemini Interconnect for PGAS Communication Runtime Systems

The Cray Gemini Interconnect has been recently introduced as the next generation network for building scalable multi-petascale supercomputers. The Cray XE6 systems, which use the Gemini Interconnect are becoming available with Message Passing Interface (MPI) and Partitioned Global Address Space (PGAS) Models such as as Global Arrays, Unified Parallel C, Co-Array Fortran and Cascade High Performance Language. These PGAS models use one-sided communication runtime systems such as MPI-Remote Memory Access, Aggregate Remote Memory Copy Interface and proprietary communication runtime systems. The primary objective of our work is to study the potential of Cray Gemini Interconnect by designing application specific micro-benchmarks using the DMAPP userspace library provided by the Gemini Interconnect. To meet this objective, we design micro-benchmarks to study the performance of simple communication primitives and application specific micro-benchmarks to understand the behavior of Gemini Interconnect at scale. In our experiments, the Gemini Interconnect can achieve a peak bandwidth of 6911 MB/s and a latency of 1us for get communication primitive. Scalability tests for atomic memory operations and shift communication operation up to 65536 processes shows the efficacy of the Gemini Interconnect.

Vishnu, Abhinav↗

Distributed Cross-Channel Hierarchical Aggregation for Foundation Models

Vision-based scientific foundation models hold significant promise for advancing scientific discovery and innovation. This potential stems from their ability to aggregate images from diverse sources—such as varying physical groundings or data acquisition systems—and to learn spatio-temporal correlations using transformer architectures. However, tokenizing and aggregating images can be compute-intensive, a challenge not fully addressed by current distributed methods. In this work, we introduce the Distributed Cross-Channel Hierarchical Aggregation (D-CHAG) approach designed for datasets with a large number of channels across image modalities. Our method is compatible with any model-parallel strategy and any type of vision transformer architecture, significantly improving computational efficiency. We evaluated D-CHAG on hyperspectral imaging and weather forecasting tasks. When integrated with tensor parallelism and model sharding, our approach achieved up to a 75% reduction in memory usage and more than doubled sustained throughput on up to 1,024 AMD GPUs on the Frontier Supercomputer.

Tsaris, Aristeidis (aris) [ORNL] (ORCID:0000000277↗

Machine Learning for Distributed Acoustic Sensing data (MLDAS) v1.0.1

MLDAS is a Python-written package for exploratory data analysis and deep learning training on Distributed Acoustic Sensing data. The machine learning tools are powered by the PyTorch library and designed to work efficiently on large scale datasets using parallel computing. Various SLURM scripts as well as a tutorial have also been made available to allow geophysicists to quickly and easily implement the available tools in their analysis workflow on supercomputer facilities.

Dumont, Vincent↗

Summer Student Internship working on the GPU

This summer internship consisted of learning computer programming skills and languages including Thrust, CUDA, C++, C and Linux. The application of this programming is for use in GPU, graphics processing units, that are processors that perform math intensive calculations rapidly in three dimensions typically in advanced video games (“NVIDIA GeForce…”). Our goal is to apply this technology towards exa-scale computing on the LANL supercomputing infrastructure. My mentors use these machines to do computations. The completed algorithms include KD tree and Top K. In order to access the GPU, CUDA was used. CUDA is a parallel computer architecture for general computing on the GPUs. There are different means by which to access CUDA. The first approach uses libraries, like Thrust, Linear Algebra, and Signal and Image Processing. Another approach parallelizes loops in Fortran or C and in the last approach people develop custom parallel algorithms using familiar language (“What is CUDA”). The algorithms I worked on this summer use the first approach of using the thrust library of parallel algorithms and data structures.

97 MATHEMATICS AND COMPUTING↗

Field-scale model for the natural attenuation of uranium at the Hanford 300 area using high performance computing

Three-dimensional reactive flow and transport simulations are carried out to better understand the persistence of uranium [U(VI)] at the Hanford 300 Area bordering the Columbia River. The massively parallel code PFLOTRAN developed under a DOE SciDAC-2 project is employed in the simulations. The calculations were carried out on 4096 processor cores on ORNL's Jaguar XT4 & 5 Cray supercomputers with run times on the order of 6 hours, equivalent to several years if performed on a single processor with sufficient memory. A new conceptual model is presented for understanding present-day and future attenuation rates of U(VI) at the 300 Area site. Unique to the conceptual model is the recognition of three distinct phases in the evolution of the site corresponding to: (I) initial emplacement of waste; (II) present-day conditions of slow leaching of U(VI) from the Hanford sediments; and (III) the complete removal of non-labile U(VI) from the source region. This work focuses on Phase II. Both labile and non-labile forms of U(VI) are included in the model as sorbed and mineralized forms of U(VI), respectively. The non-labile form plays an important role in providing a long-term source of U(VI) as it slowly leaches out of the Hanford sediment. Rapid fluctuations in the Columbia River stage on hourly, weekly and seasonal time scales are found to' playa major role in determining the migration behavior of U(VI). The calculations demonstrate that U(VI) is released into the Columbia River at a highly fluctuating rate in a ratchet-like behavior with nonzero U(VI) flux occurring only during flow from contaminated sediment into the river. The cumulative flux, however, is found to increase approximately linearly with time. The flow rate and U(VI) flux into the Columbia River predicted by the model is highly sensitive to the value used in the conductance boundary condition at the river-sediment interface. By fitting the conductance to the measured piezometric head at well 399-2-1, good agreement was obtained with observation for both the mean flux of water and U(VI) at the river-aquifer boundary with no other adjustable parameters. It was found that a multirate sorption model developed to account for long tails observed in U(VI) breakthrough curves obtained from column experiments with contaminated Hanford sediments, gave similar results for Phase II as an equilibrium surface complexation model for the discharge rate of U(VI) into the Columbia River.

54 ENVIRONMENTAL SCIENCES↗

Refining HPCToolkit for application performance analysis at exascale

As part of the US Department of Energy’s Exascale Computing Project (ECP), Rice University has been refining its HPCToolkit performance tools to better support measurement and analysis of applications executing on exascale supercomputers. To efficiently collect performance measurements of GPU-accelerated applications, HPCToolkit employs novel non-blocking data structures to communicate performance measurements between tool threads and application threads. To attribute performance information in detail to source lines, loop nests, and inlined call chains, HPCToolkit performs parallel analysis of large CPU and GPU binaries involved in the execution of an exascale application to rapidly recover mappings between machine instructions and source code. To analyze terabytes of performance measurements gathered during executions at exascale, HPCToolkit employs distributed-memory parallelism, multithreading, sparse data structures, and out-of-core streaming analysis algorithms. To support interactive exploration of profiles up to terabytes in size, HPCToolkit’s hpcviewer graphical user interface uses out-of-core methods to visualize performance data. The result of these efforts is that HPCToolkit now supports collection, analysis, and presentation of profiles and traces of GPU-accelerated applications at exascale. These improvements have enabled HPCToolkit to efficiently measure, analyze and explore terabytes of performance data for executions using as many as 64K MPI ranks and 64K GPU tiles on ORNL’s Frontier supercomputer. HPCToolkit’s support for measurement and analysis of GPU-accelerated applications has been employed to study a collection of open-science applications developed as part of ECP. This paper reports on these experiences, which provided insight into opportunities for tuning applications, strengths and weaknesses of HPCToolkit itself, as well as unexpected behaviors in executions at exascale.

Adhianto, Laksono↗

By Hand or Not By-Hand: A Case Study of Alternative Approaches to Parallelize CFD Applications

While parallel processing promises to speed up applications by several orders of magnitude, the performance achieved still depends upon several factors, including the multiprocessor architecture, system software, data distribution and alignment, as well as the methods used for partitioning the application and mapping its components onto the architecture. The existence of the Gorden Bell Prize given out at Supercomputing every year suggests that while good performance can be attained for real applications on general purpose multiprocessors, the large investment in man-power and time still has to be repeated for each application-machine combination. As applications and machine architectures become more complex, the cost and time-delays for obtaining performance by hand will become prohibitive. Computer users today can turn to three possible avenues for help: parallel libraries, parallel languages and compilers, interactive parallelization tools. The success of these methodologies, in turn, depends on proper application of data dependency analysis, program structure recognition and transformation, performance prediction as well as exploitation of user supplied knowledge. NASA has been developing multidisciplinary applications on highly parallel architectures under the High Performance Computing and Communications Program. Over the past six years, the transition of underlying hardware and system software have forced the scientists to spend a large effort to migrate and recede their applications. Various attempts to exploit software tools to automate the parallelization process have not produced favorable results. In this paper, we report our most recent experience with CAPTOOL, a package developed at Greenwich University. We have chosen CAPTOOL for three reasons: 1. CAPTOOL accepts a FORTRAN 77 program as input. This suggests its potential applicability to a large collection of legacy codes currently in use. 2. CAPTOOL employs domain decomposition to obtain parallelism. Although the fact that not all kinds of parallelism are handled may seem unappealing, many NASA applications in computational aerosciences as well as earth and space sciences are amenable to domain decomposition. 3. CAPTOOL generates code for a large variety of environments employed across NASA centers: MPI/PVM on network of workstations to the IBS/SP2 and CRAY/T3D.

Yan, Jerry C.↗