Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Nodes”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,567 records · Page 87

SU-F-J-123: CT-Based Determination of DIBH Variability and Its Dosimetric Impact On Post-Mastectomy Plus Regional Nodal Radiation Therapy

Purpose: Breast cancer radiotherapy delivered using voluntary deep inspiration breath-hold (DIBH) requires reproducible breath holds, particularly when matching supraclavicular fields to tangential fields. We studied the impact of variation in DIBHs on CTV and OAR dose metrics by comparing the dose distribution computed on two DIBH CT scans taken at the time of simulation. Methods: Ten patients receiving 50Gy in 25 fractions to the left chestwall and regional lymph nodes were studied. Two simulation CT scans were taken during separate DIBHs along with a free-breathing (FB) scan. The treatment was planned using one DIBH CT. The dose was recomputed on the other two scans using adaptive planning (Pinnacle 9.10) in which the scans are registered using a cross-correlation algorithm. The chestwall, lymph nodes and OARs were contoured on the scans following the RTOG consensus guidelines. The overall translational and rotational variation between the DIBH scans was used to estimate positional variation between breath-holds. Dose metrics between plans were compared using paired t-tests (p < 0.05) and means and standard deviations were reported. Results: The registration parameters were sub-millimeter and sub-degree. Although DIBH significantly reduced mean heart dose by 2.4Gy compared to FB (p < 0.01), no significant changes in dose were observed for targets or OARs between the two DIBH scans. Nodal coverage as assessed by V90% was 90%±8% and 89%±8% for supraclavicular and 99%±2% and 97%±22% for IM nodes. Though a significant decrease (10.5%±12.4%) in lung volume in the second DIBH CT was observed, the lung V20Gy was unchanged (14±2% and 14±3%) between the two DIBH scans. Conclusion: While the lung volume often varied between DIBHs, the CTV and OAR dose metrics were largely unchanged. This indicates that manual DIBH has the potential to provide consistent dose delivery to the chestwall and regional nodes targets when using matched fields.

60 APPLIED LIFE SCIENCES↗

Genetic algorithm based task reordering to improve the performance of batch scheduled massively parallel scientific applications

The growth in size of networked high performance computers along with novel accelerator-based node architectures has further emphasized the importance of communication efficiency in high performance computing. The world's largest high performance computers are usually operated as shared user facilities due to the costs of acquisition and operation. Applications are scheduled for execution in a shared environment and are placed on nodes that are not necessarily contiguous on the interconnect. Furthermore, the placement of tasks on the nodes allocated by the scheduler is sub-optimal, leading to performance loss and variability. Here, we investigate the impact of task placement on the performance of two massively parallel application codes on the Titan supercomputer, a turbulent combustion flow solver (S3D) and a molecular dynamics code (LAMMPS). Benchmark studies show a significant deviation from ideal weak scaling and variability in performance. The inter-task communication distance was determined to be one of the significant contributors to the performance degradation and variability. A genetic algorithm-based parallel optimization technique was used to optimize the task ordering. This technique provides an improved placement of the tasks on the nodes, taking into account the application's communication topology and the system interconnect topology. As a result, application benchmarks after task reordering through genetic algorithm show a significant improvement in performance and reduction in variability, therefore enabling the applications to achieve better time to solution and scalability on Titan during production.

97 MATHEMATICS AND COMPUTING↗

Global‐view coefficients: a data management solution for parallel quantum Monte Carlo applications

Summary Quantum Monte Carlo (QMC) applications perform simulation with respect to an initial state of the quantum mechanical system, which is often captured by using a cubic B‐spline basis. This representation is stored as a read‐only table of coefficients and accesses to the table are generated at random as part of the Monte Carlo simulation. Current QMC applications, such as QWalk and QMCPACK, replicate this table at every process or node, which limits scalability because increasing the number of processors does not enable larger systems to be run. We present a partitioned global address space approach to transparently managing this data using Global Arrays in a manner that allows the memory of multiple nodes to be aggregated. We develop an automated data management system that significantly reduces communication overheads, enabling new capabilities for QMC codes. Experimental results with QWalk and QMCPACK demonstrate the effectiveness of the data management system. Copyright © 2016 John Wiley & Sons, Ltd.

Niu, Qingpeng↗

Coupled‐cluster singles, doubles and perturbative triples with density fitting approximation for massively parallel heterogeneous platforms

Abstract A high‐performance implementation of the coupled‐cluster singles, doubles, and perturbative triples [CCSD(T)] is developed in the Massively Parallel Quantum Chemistry program. Novel features include: (1) reduced memory requirements via a density‐fitting (DF) CCSD implementation utilizing distributed lazy evaluation for tensors with more than two unoccupied indices and (2) the ability to utilize efficiently many‐core nodes (Intel Xeon Phi) and heterogeneous nodes with multiple NVIDIA GPUs on each node. All data that are greater than quadratic in the system size are distributed among processes. Excellent strong scaling is observed on distributed‐memory computers equipped with conventional CPUs, Intel Xeon Phi processors, and heterogeneous nodes with multiple NVIDIA GPUs Canonical CCSD(T) energies can be evaluated for systems containing 200 electrons and 1000 basis functions in a few days using a small size commodity cluster, with even larger computations possible on leadership‐class computing resources.

Peng, Chong↗

Wind turbine sensor array for monitoring avian and bat collisions

Assessment of avian and bat collisions with wind turbines is necessary to ensure that the benefits of renewable wind power generation are not outweighed by mortality of protected species. An onboard, integrated multisensor system capable of providing detection of turbine collision events, including taxonomic information, was developed. The conceptual design of a multisensor system including a vibration sensing node, an optics node, and an bioacoustic node with an event–driven trigger architecture was field–tested on utility–scale wind turbines. A pixel density computational model was built to estimate the spatial coverage and target resolution to the optimized configuration for camera placement. Field test results of the vibration node showed that nearly half of the recorded impact events were successfully identified by visual inspection and running short–time Fourier transform on recorded vibration signals. The remaining undetected impact events were masked under background noise due to low impact energy and high background noise of the operating turbine, which result in subsequent low signal–to–noise ratio. Furthermore, our results demonstrate the feasibility of triggering the system through single impact event sensed by vibration sensors.

17 WIND ENERGY↗

Porting DMRG++ Scientific Application to OpenPOWER

With the rapidly changing microprocessor designs and architectural diversity (multi-cores, many-cores, accelerators) for the next generation HPC systems, scientific applications must adapt to the hardware, to exploit the different types of parallelism and resources available in the architecture. To get the benefit of all the in-node hardware threads, it is important to use a single programming model to map and coordinate the available work to the different heterogeneous execution units in the node (e.g., multi-core hardware threads (latency optimized), accelerators (bandwidth optimized), etc.).Our goal is to show that we can manage the node complexity of these systems by using OpenMP for in-node parallelization by exploiting different “programming styles” supported by OpenMP 4.5 to program CPU cores and accelerators. Finding out the suitable programming-style (e.g., SPMD style, multi-level tasks, accelerator programming, nested parallelism, or a combination of these) using the latest features of OpenMP to maximize performance and achieve performance portability across heterogeneous and homogeneous systems is still an open research problem.We developed a mini-application, Kronecker Product (KP), from the original DMRG++ application (sparse matrix algebra) computational motif to experiment with different OpenMP programming styles on an OpenPOWER architecture and present their results in this paper.

Chatterjee, Ronnie↗

$\mathrm{RADICAL}$-Pilot and $\mathrm{PMIx}$/$\mathrm{PRRTE}$: Executing Heterogeneous Workloads at Large Scale on Partitioned $\mathrm{HPC}$ Resources

Execution of heterogeneous workflows on high-performance computing (HPC) platforms present unprecedented resource management and execution coordination challenges for runtime systems. Task heterogeneity increases the complexity of resource and execution management, limiting the scalability and efficiency of workflow execution. Re-source partitioning and distribution of tasks execution over portioned re-sources promises to address those problems but we lack an experimental evaluation of its performance at scale. Here this paper provides a performance evaluation of the Process Management Interface for Exascale (PMIx) and its reference implementation PRRTE on the leadership-class HPC plat-form Summit, when integrated into a pilot-based runtime system called RADICAL-Pilot. We partition resources across multiple PRRTE Distributed Virtual Machine (DVM) environments, responsible for launching tasks via the PMIx interface. We experimentally measure the work-load execution performance in terms of task scheduling/launching rate and distribution of DVM task placement times, DVM startup and termination overheads on the Summit leadership-class HPC platform. Integrated solution with PMIx/PRRTE enables using an abstracted, standardized set of interfaces for orchestrating the launch process, dynamic process management and monitoring capabilities. It extends scaling capabilities allowing to overcome a limitation of other launching mechanisms (e.g., JSM/LSF). Explored different DVM setup configurations provide insights on DVM performance and a layout to leverage it. Our experimental results show that heterogeneous workload of 65,500 tasks on 2048 nodes, and partitioned across 32 DVMs, runs steady with resource utilization not lower than 52%. While having less concurrently executed tasks resource utilization is able to reach up to 85%, based on results of heterogeneous workload of 8200 tasks on 256 nodes and 2 DVMs.

97 MATHEMATICS AND COMPUTING↗

Integrating Asynchronous Task Parallelism with OpenSHMEM, In: OpenSHMEM 2016: OpenSHMEM and Related Technologies. Enhancing OpenSHMEM for Hybrid Environments

Partitioned Global Address Space (PGAS) programming models combine shared and distributed memory features, and provide a foundation for high-productivity parallel programming using lightweight one-sided communications. The OpenSHMEM programming interface has recently begun gaining popularity as a lightweight library-based approach for developing PGAS applications, in part through its use of a symmetric heap to realize more efficient implementations of global pointers than in other PGAS systems. However, current approaches to hybrid inter-node and intra-node parallel programming in OpenSHMEM rely on the use of multithreaded programming models (e.g., pthreads, OpenMP) that harness intra-node parallelism but are opaque to the OpenSHMEM runtime. This OpenSHMEM+X approach can encounter performance challenges such as bottlenecks on shared resources, long pause times due to load imbalances, and poor data locality. Furthermore, OpenSHMEM+X requires the expertise of hero-level programmers, compared to the use of just OpenSHMEM. All of these are hard challenges to mitigate with incremental changes. This situation will worsen as computing nodes increase their use of accelerators and heterogeneous memories.

Computer Science↗

Optimizing fusion PIC code performance at scale on Cori Phase 2

In this paper we present the results of optimizing the performance of the gyrokinetic full-f fusion PIC code XGC1 on the Cori Phase Two Knights Landing system. The code has undergone substantial development to enable the use of vector instructions in its most expensive kernels within the NERSC Exascale Science Applications Program. We study the single-node performance of the code on an absolute scale using the roofline methodology to guide optimization efforts. We have obtained 2x speedups in single node performance due to enabling vectorization and performing memory layout optimizations. On multiple nodes, the code is shown to scale well up to 4000 nodes, near half the size of the machine. We discuss some communication bottlenecks that were identified and resolved during the work.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Experiences with High-Level Programming Directives for Porting Applications to GPUs. In: Facing the Multicore--Challenge II, Lecture Notes in Computer Science.

HPC systems now exploit GPUs within their compute nodes to accelerate program performance. As a result, high-end application development has become extremely complex at the node level. In addition to restructuring the node code to exploit the cores and specialized devices, the programmer may need to choose a programming model such as OpenMP or CPU threads in conjunction with an accelerator programming model to share and manage the different node resources. This comes at a time when programmer productivity and the ability to produce portable code has been recognized as a major concern. In order to offset the high development cost of creating CUDA or OpenCL kernels, directives have been proposed for programming accelerator devices, but their implications are not well known. In this paper, we evaluate the state of the art accelerator directives to program several applications kernels, explore transformations to achieve good performance, and examine the expressivity and performance penalty of using high-level directives versus CUDA. We also compare our results to OpenMP implementations to understand the benefits of running the kernels in the accelerator versus CPU cores.

CUDA Implementation↗

Instanton operators and symmetry enhancement in 5d supersymmetric quiver gauge theories

We consider general 5d SU(N ) quiver gauge theories whose nodes form an ADE Dynkin diagram of type G. Each node has SU(N i ) gauge group of general rank, Chern-Simons level κ i and additional w i fundamentals. When the total flavor number at each node is less than or equal to 2N i - 2|κ i |, we give general rules under which the symmetries associated to instanton currents are enhanced to G × G or a subgroup of it in the UV 5d superconformal theory. When the total flavor number violates that condition at some of the nodes, further enhancement of flavor symmetries occurs. In particular we find a large class of gauge theories interpreted as S1 compactification of 6d superconformal theories which are waiting for string/F-theory realization. We also consider hypermultiplets in (anti-)symmetric representation.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Improved MPI collectives for MPI processes in shared address spaces

As the number of cores per node keeps increasing, it becomes increasingly important for MPI to leverage shared memory for intranode communication. This paper investigates the design and optimization of MPI collectives for clusters of NUMA nodes. We develop performance models for collective communication using shared memory and we demonstrate several algorithms for various collectives. Experiments are conducted on both Xeon X5650 and Opteron 6100 InfiniBand clusters. The measurements agree with the model and indicate that different algorithms dominate for short vectors and long vectors. We compare our shared-memory allreduce with several MPI implementations-Open MPI, MPICH2, and MVAPICH2-that utilize system shared memory to facilitate interprocess communication. On a 16-node Xeon cluster and 8-node Opteron cluster, our implementation achieves on geometric average 2.3X and 2.1X speedup over the best MPI implementation, respectively. Our techniques enable an efficient implementation of collective operations on future multi- and manycore systems.

Collective communication↗

Distributed out-of-memory NMF on CPU/GPU architectures

We propose an efficient distributed out-of-memory implementation of the non-negative matrix factorization (NMF) algorithm for heterogeneous high-performance-computing systems. The proposed implementation is based on prior work on NMFk, which can perform automatic model selection and extract latent variables and patterns from data. In this work, we extend NMFk by adding support for dense and sparse matrix operation on multi-node, multi-GPU systems. The resulting algorithm is optimized for out-of-memory problems where the memory required to factorize a given matrix is greater than the available GPU memory. Memory complexity is reduced by batching/tiling strategies, and sparse and dense matrix operations are significantly accelerated with GPU cores (or tensor cores when available). Input/output latency associated with batch copies between host and device is hidden using CUDA streams to overlap data transfers and compute asynchronously, and latency associated with collective communications (both intra-node and inter-node) is reduced using optimized NVIDIA Collective Communication Library (NCCL) based communicators. Benchmark results show significant improvement, from 32X to 76x speedup, with the new implementation using GPUs over the CPU-based NMFk. Good weak scaling was demonstrated on up to 4096 multi-GPU cluster nodes with approximately 25,000 GPUs when decomposing a dense 340 Terabyte-size matrix and an 11 Exabyte-size sparse matrix of density 10 -6 .

97 MATHEMATICS AND COMPUTING↗

Retracted Article: Distributed energy management for networked microgrids in a three-phase unbalanced distribution network

Owing to increased penetration of three-phase and single-phase microgrids, distributed energy resources (DERs), and responsive loads, the maintenance of a three-phase balance by distribution networks is a significant challenge. Existing literature on distributed energy management for networked microgrids generally neglects the distribution network or employs a simplified phase balanced distribution network; thus, these evaluations are not applicable. Further, the underlying mutual coupling between the different phases of distribution feeders results in a more challenging situation. Here, to solve this issue, this study sought to propose distributed energy management based on a three-phase unbalanced distribution network. Various three-phase or single-phase microgrids, utility-owned DERs, and responsive loads were coordinated through iteratively adjusted price signals. Based on the price signals received, the microgrid controllers (MCs) and distribution management system (DMS) updated the schedules of the DERs and responsive loads under their jurisdiction separately. The price signals were then updated according to the generation-load mismatch at each node and distributed to the corresponding MCs and DMS for the next iteration. The iteration continued until a sufficiently small generation-load mismatch was achieved at all nodes, that is, a balanced generation and load at all nodes under the agreed price signals. Considering a three-phase unbalanced distribution network, the price signals were determined per phase per node. Overall, the proposed distributed energy management coordinates microgrids, utility-owned DERs, responsive loads with guaranteed network constraints, and preserves the privacy of microgrid customers. This distributed energy management method was further demonstrated through various case studies on a three-phase networked microgrid test system.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Developing an integrated hub location and revenue management model considering multi-classes of customers in the airline industry

The hub location and revenue management problems are two interesting research fields in the transportation and network design studies. The hub location model designs the transportation network structure and the revenue management model allocates the network’s capacity for customer classes based on their sensitivity to prices. In this paper, we consider the integrated hub location and revenue management problem in the airline industry to maximize the revenue of transportation network and minimize hub installation costs. p hubs have been chosen from a set of n nodes and connected to a central node based on star–star network design. The capacity of the installed links between the hubs and the central node and the hubs and non-hub nodes is limited. This limited capacity has been allocated to the stochastic demands of customer classes using the revenue management approach. A two-stage stochastic model has been derived to determine the locations of hubs and protection levels in the ticket sale process, and an efficient modification of genetic algorithm has been proposed for the large-scale problems. Numerical experiments have been carried out to assess the effectiveness of the proposed algorithm.

97 MATHEMATICS AND COMPUTING↗

Expression of the hypoxia-inducible monocarboxylate transporter MCT4 is increased in triple negative breast cancer and correlates independently with clinical outcome

Highlights: • Glycolytic markers are highly expressed in triple negative breast cancers. • Lactate/H{sup +} symporter MCT4 demonstrated the strongest deleterious impact on survival. • MCT4 should serve as a new prognostic factor in node-negative breast cancers. - Abstract: Background: {sup 18}Fluor-deoxy-glucose PET-scanning of glycolytic metabolism is being used for staging in many tumors however its impact on prognosis has never been studied in breast cancer. Methods: Glycolytic and hypoxic markers: glucose transporter (GLUT1), carbonic anhydrase IX (CAIX), monocarboxylate transporter 1 and 4 (MCT1, 4), MCT accessory protein basigin and lactate-dehydrogenase A (LDH-A) were assessed by immunohistochemistry in two cohorts of breast cancer comprising 643 node-negative and 127 triple negative breast cancers (TNBC) respectively. Results: In the 643 node-negative breast tumor cohort with a median follow-up of 124 months, TNBC were the most glycolytic (≈70%), followed by Her-2 (≈50%) and RH-positive cancers (≈30%). Tumoral MCT4 staining (without stromal staining) was a strong independent prognostic factor for metastasis-free survival (HR = 0.47, P = 0.02) and overall-survival (HR = 0.38, P = 0.002). These results were confirmed in the independent cohort of 127 cancer patients. Conclusion: Glycolytic markers are expressed in all breast tumors with highest expression occurring in TNBC. MCT4, the hypoxia-inducible lactate/H{sup +} symporter demonstrated the strongest deleterious impact on survival. We propose that MCT4 serves as a new prognostic factor in node-negative breast cancer and can perhaps act soon as a theranostic factor considering the current pharmacological development of MCT4 inhibitors.

60 APPLIED LIFE SCIENCES↗

Tangential Radiotherapy Without Axillary Surgery in Early-Stage Breast Cancer: Results of a Prospective Trial

Purpose: To determine the risk of regional-nodal recurrence in patients with early-stage, invasive breast cancer, with clinically negative axillary nodes, who were treated with breast-conserving surgery, 'high tangential' breast radiotherapy, and hormonal therapy, without axillary surgery or the use of a separate nodal radiation field. Methods and Materials: Between September 1998 and November 2003, 74 patients who were {>=}55 years of age with Stage I-II clinically node-negative, hormone-receptor-positive breast cancer underwent tumor excision to negative margins without axillary surgery as a part of a multi-institutional prospective study. Postoperatively, all underwent high-tangential, whole-breast radiotherapy with a boost to the tumor bed, followed by 5 years of hormonal therapy. Results: For the 74 patients enrolled, the median age was 74.5 years, and the median pathologic tumor size was 1.2 cm. Lymphatic vessel invasion was present in 5 patients (7%). At a median follow-up of 52 months, no regional-nodal failures or ipsilateral breast recurrences had been identified (95% confidence interval, 0-4%). Eight patients died, one of metastatic disease and seven of other causes. Conclusion: In this select group of mainly older patients with early-stage hormone-responsive breast cancer and clinically negative axillary nodes, treatment with high-tangential breast radiotherapy and hormonal therapy, without axillary surgery, yielded a low regional recurrence rate. Such patients might be spared more extensive axillary treatment (axillary surgery, including sentinel node biopsy, or a separate nodal radiation field), with its associated time, expense, and morbidity.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Patterns of Regional Recurrence After Definitive Radiotherapy for Cervical Cancer

Purpose: To determine the patterns of regional recurrence in patients treated with definitive radiotherapy (RT) for cervical cancer. Methods and Materials: The records of 198 patients treated with definitive RT for cervical cancer between 1980 and 2000 who experienced a regional recurrence without a central or distal vaginal recurrence were reviewed. All patients received a combination of external-beam RT and intracavitary brachytherapy. In the 180 patients with a documented location of regional recurrence, the relationship between the recurrence and the radiation fields was determined. Results: The median time to regional recurrence was 13 months (range, 2-85 months). Of the 180 patients who had an evaluable regional recurrence, 119 (66%) had a component of marginal failure; 71 patients recurred above-the-field, 2 patients occurred in the inguinal nodes, and 2 patients recurred above-the-field and in the inguinal nodes. In addition, 105 patients (58%) had a component of in-field failure; 59 patients recurred in-field only, 39 patients recurred in-field and above-the-field, 2 patients recurred in-field, above-the-field, and in the inguinal nodes, and 5 patients recurred in-field and in the inguinal nodes. The median survival after regional recurrence was 8 months (range, 0-194 months). Conclusions: Most regional recurrences after definitive RT for cervical cancer include a component of marginal failure, usually immediately superior to the radiation field. These recurrences suggest a deficiency in target volume. Recurrences also occur in-field, suggesting a deficiency in dose. Developments in pretreatment staging, field delineation, dose escalation, and posttreatment surveillance may help to improve outcome in these patients.

62 RADIOLOGY AND NUCLEAR MEDICINE↗