Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallelization”

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 181 records · Page 10

Parallel transport of long mean-free-path plasmas along open magnetic field lines: Plasma profile variation

Parallel transport of long mean-free-path plasma along an open magnetic field line is characterized by strong temperature anisotropy, which is driven by two effects. The first is magnetic moment conservation in a non-uniform magnetic field, which can transfer energy between parallel and perpendicular degrees of freedom. The second is decompressional cooling of the parallel temperature due to parallel flow acceleration by conventional presheath electric field which is associated with the sheath condition near the wall surface where the open magnetic field line intercepts the discharge chamber. To the leading order in gyroradius to system gradient length scale expansion, the parallel transport can be understood via the Chew-Goldbeger-Low (CGL) model which retains two components of the parallel heat flux, i.e., q{sub n} associated with the parallel thermal energy and q{sub s} related to perpendicular thermal energy. It is shown that in addition to the effect of magnetic field strength (B) modulation, the two components (q{sub n} and q{sub s}) of the parallel heat flux play decisive roles in the parallel variation of the plasma profile, which includes the plasma density (n), parallel flow (u), parallel and perpendicular temperatures (T{sub Parallel-To} and T{sub Up-Tack }), and the ambipolar potential ({phi}). Both their profile (q{sub n}/B and q{sub s}/B{sup 2}) and the upstream values of the ratio of the conductive and convective thermal flux (q{sub n}/nuT{sub Parallel-To} and q{sub s}/nuT{sub Up-Tack }) provide the controlling physics, in addition to B modulation. The physics described by the CGL model are contrasted with those of the double-adiabatic laws and further elucidated by comparison with the first-principles kinetic simulation for a specific but representative flux expander case.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Exploiting Internal Parallelism for Address Translation in Solid-State Drives

Solid-state Drives (SSDs) have changed the landscape of storage systems and present a promising storage solution for data-intensive applications due to their low latency, high bandwidth, and low power consumption compared to traditional hard disk drives. SSDs achieve these desirable characteristics using internal parallelism—parallel access to multiple internal flash memory chips—and a Flash Translation Layer (FTL) that determines where data are stored on those chips so that they do not wear out prematurely. However, current state-of-the-art cache-based FTLs like the Demand-based Flash Translation Layer (DFTL) do not allow IO schedulers to take full advantage of internal parallelism, because they impose a tight coupling between the logical-to-physical address translation and the data access. In this study to address this limitation, we introduce a new FTL design called Parallel-DFTL that works with the DFTL to decouple address translation operations from data accesses. Parallel-DFTL separates address translation and data access operations into different queues, allowing the SSD to use concurrent flash accesses for both types of operations. We also present a Parallel-LRU cache replacement algorithm to improve the concurrency of address translation operations. To compare Parallel-DFTL against existing FTL approaches, we present a Parallel-DFTL performance model and compare its predictions against those for DFTL and an ideal page-mapping approach. We also implemented the Parallel-DFTL approach in an SSD simulator using real device parameters, and used trace-driven simulation to evaluate Parallel-DFTL’s efficacy. Our evaluation results show that Parallel-DFTL improved the overall performance by up to 32% for the real IO workloads we tested, and by up to two orders of magnitude with synthetic test workloads. Finally, we also found that Parallel-DFTL is able to achieve reasonable performance with a very small cache size and that it provides the best benefit for those workloads with large request size or with high write ratio.

97 MATHEMATICS AND COMPUTING↗

Integration experiments and performance studies of a COTS parallel archive system

Current and future Archive Storage Systems have been asked to (a) scale to very high bandwidths, (b) scale in metadata performance, (c) support policy-based hierarchical storage management capability, (d) scale in supporting changing needs of very large data sets, (e) support standard interface, and (f) utilize commercial-off-the-shelf (COTS) hardware. Parallel file systems have been asked to do the same thing but at one or more orders of magnitude faster in performance. Archive systems continue to move closer to file systems in their design due to the need for speed and bandwidth, especially metadata searching speeds such as more caching and less robust semantics. Currently the number of extreme highly scalable parallel archive solutions is very small especially those that will move a single large striped parallel disk file onto many tapes in parallel. We believe that a hybrid storage approach of using COTS components and innovative software technology can bring new capabilities into a production environment for the HPC community much faster than the approach of creating and maintaining a complete end-to-end unique parallel archive software solution. In this paper, we relay our experience of integrating a global parallel file system and a standard backup/archive product with a very small amount of additional code to provide a scalable, parallel archive. Our solution has a high degree of overlap with current parallel archive products including (a) doing parallel movement to/from tape for a single large parallel file, (b) hierarchical storage management, (c) ILM features, (d) high volume (non-single parallel file) archives for backup/archive/content management, and (e) leveraging all free file movement tools in Linux such as copy, move, Is, tar, etc. We have successfully applied our working COTS Parallel Archive System to the current world's first petafiop/s computing system, LANL's Roadrunner machine, and demonstrated its capability to address requirements of future archival storage systems.

99 GENERAL AND MISCELLANEOUS↗

Integration experiences and performance studies of A COTS parallel archive systems

Current and future Archive Storage Systems have been asked to (a) scale to very high bandwidths, (b) scale in metadata performance, (c) support policy-based hierarchical storage management capability, (d) scale in supporting changing needs of very large data sets, (e) support standard interface, and (f) utilize commercial-off-the-shelf(COTS) hardware. Parallel file systems have been asked to do the same thing but at one or more orders of magnitude faster in performance. Archive systems continue to move closer to file systems in their design due to the need for speed and bandwidth, especially metadata searching speeds such as more caching and less robust semantics. Currently the number of extreme highly scalable parallel archive solutions is very small especially those that will move a single large striped parallel disk file onto many tapes in parallel. We believe that a hybrid storage approach of using COTS components and innovative software technology can bring new capabilities into a production environment for the HPC community much faster than the approach of creating and maintaining a complete end-to-end unique parallel archive software solution. In this paper, we relay our experience of integrating a global parallel file system and a standard backup/archive product with a very small amount of additional code to provide a scalable, parallel archive. Our solution has a high degree of overlap with current parallel archive products including (a) doing parallel movement to/from tape for a single large parallel file, (b) hierarchical storage management, (c) ILM features, (d) high volume (non-single parallel file) archives for backup/archive/content management, and (e) leveraging all free file movement tools in Linux such as copy, move, ls, tar, etc. We have successfully applied our working COTS Parallel Archive System to the current world's first petaflop/s computing system, LANL's Roadrunner, and demonstrated its capability to address requirements of future archival storage systems.

99 GENERAL AND MISCELLANEOUS↗

Accelerating Parallel Applications in Cloud Platforms via Adaptive Time-Slice Control

Cloud platforms can provide flexible and cost-effective environments for parallel applications. However, the resource over-commitment issues, i.e., cloud providers often provide much more executable virtual CPUs than available physical CPUs, still impede the synchronization operations of parallel applications, causing severe performance degradation. Existing methods optimize parallel applications by promoting the priorities of involved VMs. They cannot fully explore the performance of parallel applications, because they ignore the time-slice requirements of different phases of parallel applications. Furthermore, non-parallel applications experience unsatisfied performance because of low scheduling priorities. Given empirical analysis on time-slices of virtual machines (VMs), we find that shortening time-slices can mitigate synchronization overhead which incurs during communication phases, while over-short time-slices cause frequent cache misses in computation phases. Accordingly, we propose an Adaptive Time-slice Control (ATC) mechanism. ATC first detects the phases of parallel applications based on lock latency or cache misses. Then, ATC shortens time-slices during communication phases and prolongs time-slices during computation phases for parallel applications, and sets a uniform time-slice for non-parallel applications. Finally, we evaluate ATC using seven well-known benchmarks with 25+ applications. Experiments show that ATC obtains 1.5-75x performance gain for running parallel applications than state-of-the-art solutions, with nearly unaffected impact on non-parallel applications.

97 MATHEMATICS AND COMPUTING↗

Race conditions and data partitioning: risks posed by common errors to reproducible parallel simulations

When parallel algorithms for simulation were introduced in the 1970s, their development and use interested only experts in parallel computation. This circumstance changed as multi-core processors became commonplace, putting a parallel computer into the hands of every modeler. A natural outcome is growing interest in parallel simulation among persons not intimately familiar with parallel computing. At the same time, parallel simulation tools continue to be developed with the implicit assumption that the modeler is knowledgeable about parallel programming. The unintended consequence is a rapidly growing number of users of parallel simulation tools that are unlikely to recognize when the interaction of race conditions, partitioning strategies, and simultaneous action in their simulation models make results non-reproducible, thereby calling into question the validity of conclusions drawn from the simulation data. Here, we illustrate the potential dangers of exposing parallel algorithms to users who are not experts in parallel computation with example models constructed using existing parallel simulation tools. By doing so, we hope to refocus tool developers on usability, even if this new focus incurs loss of some performance.

97 MATHEMATICS AND COMPUTING↗

Symmetry breaking effects of density gradient on parallel momentum transport: A new {rho}{sub s}{sup *} effect

Symmetry breaking effects of density gradient on parallel momentum transport is studied via quasilinear theory. It is shown that finite {rho}{sub s}*({identical_to}{rho}{sub s}/L{sub n}), where {rho}{sub s} is ion sound radius and L{sub n} is density scale length, leads to symmetry breaking of the ion temperature gradient (ITG) eigenfunction. This broken symmetry persists even in the absence of mean poloidal (from radial electric field shear) and toroidal flows. This effect, as explained in the text, originates from the divergence of polarization particle current in the ion continuity equation. The form of the eigenfunction allows the microturbulence to generate parallel residual stress via symmetry breaking. Comparison with the E-vector x B-vector shear driven parallel residual stress, parallel polarization stress and turbulence intensity gradient driven parallel residual stress are discussed. It is shown that this {rho}{sub s}* driven parallel residual stress may become comparable to E-vector x B-vector shear driven parallel residual stress in small L{sub n} region. In the regular drift wave ordering, where {rho}{sub s}*<<1, this effect is found to be of the same order as the parallel polarization stress. This {rho}{sub s}* driven parallel residual stress can also overtake the turbulence intensity gradient driven parallel residual stress in strong density gradient region whereas the later one is dominant in the strong profile curvature region. The parallel momentum diffusivity is found to remain undisturbed by this {rho}{sub s}* effect as long as the turbulence intensity inhomogenity is not important.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Current parallel I/O limitations to scalable data analysis.

This report describes the limitations to parallel scalability which we have encountered when applying our otherwise optimally scalable parallel statistical analysis tool kit to large data sets distributed across the parallel file system of the current premier DOE computational facility. This report describes our study to evaluate the effect of parallel I/O on the overall scalability of a parallel data analysis pipeline using our scalable parallel statistics tool kit [PTBM11]. In this goal, we tested it using the Jaguar-pf DOE/ORNL peta-scale platform on a large combustion simulation data under a variety of process counts and domain decompositions scenarios. In this report we have recalled the foundations of the parallel statistical analysis tool kit which we have designed and implemented, with the specific double intent of reproducing typical data analysis workflows, and achieving optimal design for scalable parallel implementations. We have briefly reviewed those earlier results and publications which allow us to conclude that we have achieved both goals. However, in this report we have further established that, when used in conjuction with a state-of-the-art parallel I/O system, as can be found on the premier DOE peta-scale platform, the scaling properties of the overall analysis pipeline comprising parallel data access routines degrade rapidly. This finding is problematic and must be addressed if peta-scale data analysis is to be made scalable, or even possible. In order to attempt to address these parallel I/O limitations, we will investigate the use the Adaptable IO System (ADIOS) [LZL+10] to improve I/O performance, while maintaining flexibility for a variety of IO options, such MPI IO, POSIX IO. This system is developed at ORNL and other collaborating institutions, and is being tested extensively on Jaguar-pf. Simulation code being developed on these systems will also use ADIOS to output the data thereby making it easier for other systems, such as ours, to process that data.

99 GENERAL AND MISCELLANEOUS↗

Parallel sorting algorithm classification: is manual instrumentation necessary?

Understanding parallel algorithms is crucial for accelerating scientific simulations on complex, distributed memory, high-performance computers. Modern algorithm classification approaches learn semantics directly from source code to differentiate between algorithms, however, accessing source code is not always possible. We can learn about parallel algorithms from observing their performance, as programs running the same algorithms and using the same hardware should exhibit similar performance characteristics. We present an approach to learn algorithm classes from parallel performance data directly in order to classify algorithms without access to the source code. We extend previous work to enable classifying parallel sorting algorithms using automatic instrumentation instead of requiring manual region annotations in the source code. In this work, we design and demonstrate a study for classification of parallel sorting algorithms using parallel performance data collected from automatic instrumentation, and evaluate the performance of our new methodology on classification. We leverage Caliper to collect the performance data, Thicket for our exploratory data analysis (EDA), and PyTorch and Scikit-learn to evaluate the effectiveness of random forests, support vector machines (SVMs), decision trees, neural networks, and logistic regressions on parallel performance data. Additionally, we study noise in parallel performance data, whether the removal of noise and pre-processing of the data is necessary to accurately classify parallel sorting algorithms, and determine the effectiveness of features created from performance data. In conclusion, we demonstrate classification accuracy for these five different models of up to 97.7% across four different parallel algorithm classes.

Algorithm Classification↗

TorchBraid: High-Performance Layer-Parallel Training of Deep Neural Networks with MPI and GPU Acceleration

TorchBraid is a high-performance implementation of layer-parallel training for deep neural networks (DNNs) supporting MPI-based parallelism and GPU acceleration. Layer-parallel training has been developed to overcome the serialization inherent in forward and backward propagation of DNNs that limits utilization of computational resources in the strong scaling limit. To achieve this, TorchBraid integrates the PyTorch neural network framework with the state-of-the-art XBraid time-parallel library. Furthermore, this article presents the use and performance of TorchBraid, in addition to solutions for overcoming the algorithmic challenges inherent in combining automatic differentiation with layer-parallel. Results are presented with and without GPU acceleration for the Tiny ImageNet and MNIST image classification data sets, as well as recurrent neural networks. Overall, TorchBraid enables fast training of DNNs, both in a strong and weak scaling context. In addition to the TorchBraid software, several new advances in applying layer-parallel algorithms are detailed. Integration of layer-parallel with data-parallel algorithms is presented for the first time, showing the computational advantages of the combination. Standard deep learning techniques, like batch-normalization, are developed for layer-parallel training. Finally, a new approach combining layer-parallel with spatial coarsening in order to accelerate training for 3D image classification shows roughly a 10× speedup over serial execution.

Layer-parallel↗

Improving Parallel I/O Performance with Data Layout Awareness

Parallel applications can benefit greatly from massive computational capability, but their performance suffers from large latency of I/O accesses. The poor I/O performance has been attributed as a critical cause of the low sustained performance of parallel computing systems. In this study, we propose a data layout-aware optimization strategy to promote a better integration of the parallel I/O middleware and parallel file systems, two major components of the current parallel I/O systems, and to improve the data access performance. We explore the layout-aware optimization in both independent I/O and collective I/O, two primary forms of I/O in parallel applications. We illustrate that the layout-aware I/O optimization could improve the performance of current parallel I/O strategy effectively. The experimental results verify that the proposed strategy could improve parallel I/O performance by nearly 40% on average. The proposed layout-aware parallel I/O has a promising potential in improving the I/O performance of parallel systems.

99 GENERAL AND MISCELLANEOUS↗

Complete classification of parallel Lorentz surfaces in four-dimensional neutral pseudosphere

A Lorentz surface of an indefinite space form is called parallel if its second fundamental form is parallel with respect to the Van der Waerden-Bortolotti connection. Such surfaces are locally invariant under the reflection with respect to the normal space at each point. Parallel surfaces are important in geometry as well as in general relativity since extrinsic invariants of such surfaces do not change from point to point. Parallel Lorentz surfaces in four-dimensional (4D) Lorentzian space forms are classified by Chen and Van der Veken [''Complete classification of parallel surfaces in 4-dimensional Lorentz space forms,'' Tohoku Math. J. 61, 1 (2009)]. Recently, explicit classification of parallel Lorentz surfaces in the pseudo-Euclidean 4-space E{sub 2}{sup 4} and in the pseudohyperbolic 4-space H{sub 2}{sup 4}(-1) are obtained recently by Chen et al. [''Complete classification of parallel Lorentzian surfaces in Lorentzian complex space forms,'' Int. J. Math. 21, 665 (2010); ''Complete classification of parallel Lorentz surfaces in neutral pseudo hyperbolic 4-space,'' Cent. Eur. J. Math. 8, 706 (2010)], respectively. In this article, we completely classify the remaining case; namely, parallel Lorentz surfaces in 4D neutral pseudosphere S{sub 2}{sup 4}(1). Our result states that there are 24 families of such surfaces in S{sub 2}{sup 4}(1). Conversely, every parallel Lorentz surface in S{sub 2}{sup 4}(1) is obtained from one of the 24 families. The main result indicates that there are major differences between Lorentz surfaces in the de Sitter 4-space dS{sub 4} and in the neutral pseudo 4-sphere S{sub 2}{sup 4}.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Modeling Cooperative Threads to Project GPU Performance for Adaptive Parallelism

Most accelerators, such as graphics processing units (GPUs) and vector processors, are particularly suitable for accelerating massively parallel workloads. On the other hand, conventional workloads are developed for multi-core parallelism, which often scale to only a few dozen OpenMP threads. When hardware threads significantly outnumber the degree of parallelism in the outer loop, programmers are challenged with efficient hardware utilization. A common solution is to further exploit the parallelism hidden deep in the code structure. Such parallelism is less structured: parallel and sequential loops may be imperfectly nested within each other, neigh boring inner loops may exhibit different concurrency patterns (e.g. Reduction vs. Forall), yet have to be parallelized in the same parallel section. Many input-dependent transformations have to be explored. A programmer often employs a larger group of hardware threads to cooperatively walk through a smaller outer loop partition and adaptively exploit any encountered parallelism. This process is time-consuming and error-prone, yet the risk of gaining little or no performance remains high for such workloads. To reduce risk and guide implementation, we propose a technique to model workloads with limited parallelism that can automatically explore and evaluate transformations involving cooperative threads. Eventually, our framework projects the best achievable performance and the most promising transformations without implementing GPU code or using physical hardware. We envision our technique to be integrated into future compilers or optimization frameworks for autotuning.

Meng, Jiayuan↗

Enabling Low-Overhead HT-HPC Workflows at Extreme Scale using GNU Parallel

GNU Parallel is a versatile and powerful tool for process parallelization widely used in scientific computing. This paper demonstrates its effective application in high-performance computing (HPC) environments, particularly focusing on its scalability and efficiency in executing large-scale high-throughput high-performance computing (HT-HPC) workflows. Through real-world examples, we highlight GNU Parallel’s performance across various HPC workloads, including GPU computing, container-based workloads, and node-local NVMe storage. Our results on two leading supercomputers, OLCF’s Frontier and NERSC’s Perlmutter, showcase GNU Parallel’s rapid process dispatching ability and its capacity to maintain low overhead even at extreme scales. We explore GNU Parallel’s application in massive parallel file transfers using a scheduled Data Transfer Node (DTN) cluster, emphasizing its broad utility in diverse scientific workflows. Beyond its direct application as a viable workflow manager, GNU Parallel can be employed in conjunction with other workflow systems as a "last-mile" parallelizing driver and as a quick prototyping tool to design and extract parallel profiles from application executions. We then argue that the potential for GNU Parallel to transform workflow management at extreme scales is substantial, paving the way for more efficient and effective scientific discoveries.

Maheshwari, Ketan↗

Exploiting Vector and Multicore Parallelsim for Recursive, Data- and Task-Parallel Programs

Modern hardware contains parallel execution resources that are well-suited for data-parallelism-vector units-and task parallelism-multicores. However, most work on parallel scheduling focuses on one type of hardware or the other. In this work, we present a scheduling framework that allows for a unified treatment of task- and data-parallelism. Our key insight is an abstraction, task blocks, that uniformly handles data-parallel iterations and task-parallel tasks, allowing them to be scheduled on vector units or executed independently as multicores. Our framework allows us to define schedulers that can dynamically select between executing task- blocks on vector units or multicores. We show that these schedulers are asymptotically optimal, and deliver the maximum amount of parallelism available in computation trees. To evaluate our schedulers, we develop program transformations that can convert mixed data- and task-parallel pro- grams into task block-based programs. Using a prototype instantiation of our scheduling framework, we show that, on an 8-core system, we can simultaneously exploit vector and multicore parallelism to achieve 14×-108× speedup over sequential baselines.

Ren, Bin↗

A space-time parallel algorithm with adaptive mesh refinement for computational fluid dynamics

This work describes a space-time parallel algorithm with space-time adaptive mesh refinement (AMR). AMR with subcycling is added to multigrid reduction-in-time (MGRIT) in order to provide solution efficient adaptive grids with a reduction in work performed on coarser grids. This algorithm is achieved by integrating two software libraries: XBraid (Parallel time integration with multigrid. https://computation.llnl.gov/projects/parallel-timeintegration-multigrid) and Chombo (Chombo software package for AMR applications—design document, 2014). The former is a parallel time integration library using multigrid and the latter is a massively parallel structured AMR library. Employing this adaptive space-time parallel algorithm is Chord (Comput Fluids 123:202–217, 2015), a computational fluid dynamics (CFD) application code for solving compressible fluid dynamics problems. For the same solution accuracy, speedups are demonstrated from the use of space-time parallelization over the time-sequential integration on Couette flow and Stokes’ second problem. On a transient Couette flow case, at least a 1.5× speedup is achieved, and with a time periodic problem, a speedup of up to 13.7× over the time-sequential case is obtained. In both cases, the speedup is achieved by adding processors and exploring additional parallelization in time. The numerical experiments show the algorithm is promising for CFD applications that can take advantage of the time parallelism. Future work will focus on improving the parallel performance and providing more tests with complex fluid dynamics to demonstrate the full potential of the algorithm.

97 MATHEMATICS AND COMPUTING↗

A Parallel Alternative for Energy-Efficient Neural Network Training and Inferencing

Energy efficiency of training and inferencing with large neural network models is a critical challenge facing the future of sustainable large-scale machine learning workloads. This paper introduces an alternative strategy, called phantom parallelism, to minimize the net energy consumption of traditional tensor (model) parallelism, the most energy-inefficient component of large neural network training. The approach is presented in the context of feed-forward network architectures as a preliminary, but comprehensive, proof-of-principle study of the proposed methodology. We derive new forward and backward propagation operators for phantom parallelism, implement them as custom autograd operations within an end-to-end phantom parallel training pipeline and compare its parallel performance and energy-efficiency against those of conventional tensor parallel training pipelines. Formal analyses that predict lower bandwidth and FLOP counts are presented with supporting empirical results on up to 256 GPUs that corroborate these gains. Experiments are shown to deliver ∼50% reduction in the energy consumed to train FFNs using the proposed phantom parallel approach when compared with conventional tensor parallel methods. Additionally, the proposed approach is shown to train smaller phantom models to the same model loss on smaller GPU counts as larger tensor parallel models on larger GPU counts offering the possibility for even greater energy savings.

Seal, Sudip [ORNL] (ORCID:0000000332330656)↗

Integrated Model, Batch and Domain Parallelism in Training Neural Networks

We propose a new integrated method of exploiting model, batch and domain parallelism for the training of deep neural networks (DNNs) on large distributed-memory computers using minibatch stochastic gradient descent (SGD). Our goal is to find an efficient parallelization strategy for a fixed batch size using $P$ processes. Our method is inspired by the communication-avoiding algorithms in numerical linear algebra. We see $P$ processes as logically divided into a $P_r \times P_c$ grid where the $P_r$ dimension is implicitly responsible for model/domain parallelism and the $P_c$ dimension is implicitly responsible for batch parallelism. In practice, the integrated matrix-based parallel algorithm encapsulates these types of parallelism automatically. We analyze the communication complexity and analytically demonstrate that the lowest communication costs are often achieved neither with pure model nor with pure data parallelism. We also show how the domain parallel approach can help in extending the theoretical scaling limit of the typical batch parallel method.

97 MATHEMATICS AND COMPUTING↗