Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel computer architecture”

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 1,351 records · Page 75

Evaluating Multi-core Architectures through Accelerating the Three-Dimensional Lax–Wendroff Correction

Wave propagation forward modeling is a widely used computational method in oil and gas exploration. The iterative stencil loops in such problems have broad applications in scientific computing. However, executing such loops can be highly time time-consuming, which greatly limits application’s performance and power efficiency. In this paper, we accelerate the forward modeling technique on the latest multi-core and many-core architectures such as Intel Sandy Bridge CPUs, NVIDIA Fermi C2070 GPU, NVIDIA Kepler K20x GPU, and the Intel Xeon Phi Co-processor. For the GPU platforms, we propose two parallel strategies to explore the performance optimization opportunities for our stencil kernels. For Sandy Bridge CPUs and MIC, we also employ various optimization techniques in order to achieve the best.

3D Wave Forward Modeling↗

Supercomputing systems - A projection to 2000

Advances in computer architecture, computer science, computational methods, and constituent technologies are expected to lead to significant advances in the performance of scientific supercomputing system capabilities over the next decade. By the year 2000, single 1-in-sq dies are projected to incorporate four processors, each of which would be operating faster than 750 million instructions per second (MIPS) for a total on-chip processing performance in excess of 2000 MIPS. Scalable parallel processors can be expected to contain thousands of such multiple processor chips. In general, semiconductor performance advances appear to change about one order of magnitude every five years. Rotating magnetic memory and communications technology are not advancing as rapidly, with the result that the allocation of functions within the system configurations fo future supercomputer systems will require important changes. Availability of massively parallel heterogeneous processing capabilities should be a catalyst leading to new approaches for applications.

Lundstrom, S. F.↗

Very fast motion planning for highly dexterous-articulated robots

Due to the inherent danger of space exploration, the need for greater use of teleoperated and autonomous robotic systems in space-based applications has long been apparent. Autonomous and semi-autonomous robotic devices have been proposed for carrying out routine functions associated with scientific experiments aboard the shuttle and space station. Finally, research into the use of such devices for planetary exploration continues. To accomplish their assigned tasks, all such autonomous and semi-autonomous devices will require the ability to move themselves through space without hitting themselves or the objects which surround them. In space it is important to execute the necessary motions correctly when they are first attempted because repositioning is expensive in terms of both time and resources (e.g., fuel). Finally, such devices will have to function in a variety of different environments. Given these constraints, a means for fast motion planning to insure the correct movement of robotic devices would be ideal. Unfortunately, motion planning algorithms are rarely used in practice because of their computational complexity. Fast methods have been developed for detecting imminent collisions, but the more general problem of motion planning remains computationally intractable. However, in this paper we show how the use of multicomputers and appropriate parallel algorithms can substantially reduce the time required to synthesize paths for dexterous articulated robots with a large number of joints. We have developed a parallel formulation of the Randomized Path Planner proposed by Barraquand and Latombe. We have shown that our parallel formulation is capable of formulating plans in a few seconds or less on various parallel architectures including: the nCUBE2 multicomputer with up to 1024 processors (nCUBE2 is a registered trademark of the nCUBE corporation), and a network of workstations.

Challou, Daniel J.↗

Progress in Fast, Accurate Multi-scale Climate Simulations

We present a survey of physical and computational techniques that have the potential to con- tribute to the next generation of high-fidelity, multi-scale climate simulations. Examples of the climate science problems that can be investigated with more depth include the capture of remote forcings of localized hydrological extreme events, an accurate representation of cloud features over a range of spatial and temporal scales, and parallel, large ensembles of simulations to more effectively explore model sensitivities and uncertainties. Numerical techniques, such as adaptive mesh refinement, implicit time integration, and separate treatment of fast physical time scales are enabling improved accuracy and fidelity in simulation of dynamics and allow more complete representations of climate features at the global scale. At the same time, part- nerships with computer science teams have focused on taking advantage of evolving computer architectures, such as many-core processors and GPUs, so that these approaches which were previously considered prohibitively costly have become both more efficient and scalable. In combination, progress in these three critical areas is poised to transform climate modeling in the coming decades.

climate modeling↗

Multi-node and Multi-core Performance Studies of a Monte Carlo code RMC

Over the past 15 years the architecture of high-performance computing systems has become increasingly heterogeneous with individual computing cores being subdivided into nodes and even auxiliary co-processing units such as graphics processing units or the Intel many integrated core co-processors. In fact, future computer platforms is moving toward a larger number of nodes and cores with lower memory available per node. Today, several production Monte Carlo codes use some types of hybrid parallelism strategy that makes use of shared-memory parallelism within a computing node and message-passing parallelism between nodes, which enables the codes to utilize available computing resources as much as possible, both among and within computing nodes on a cluster. The so-called hybrid parallelization is also possible with Reactor Monte Carlo Code RMC by activating both MPI and OpenMP parallel algorithms together. On HPC' which typically consists of several or several-ten thousand nodes equipping 16-32 computing intra-cores per node, simulation of RMC can be carried out in the hybrid parallelization with the MPI inter-node parallelization and the OpenMP intra-node parallelization. RMC has adopted a hybrid message-passing and shared-memory approach to parallelism, which enables the code to utilize available computing resources, both among and within computing nodes on a cluster. Shared-memory parallelism is especially valuable for large-model simulations because it enables multiple compute cores within a node to share common model information such as the geometry description, material, cross section and tally data, thus reducing the memory usage per processor. The calculation on the Tianhe-2 supercomputer took 6144 cores, resulting in a rate of 173,314,000 neutrons/minute for BEAVRS benchmark. While pin-wise full core H-M benchmark burnup calculations with millions of depletion regions was selected using MPI/OpenMP parallelism, the speed of 6,144 cores reach 1,843,200 burnable regions per minute. Keeping up with the continuing trend for increases in the number of computing resources within modern HPC systems is a significant challenge for teams that develop and maintain high-performance MC radiation transport solvers. order to address this challenge, RMC has adopted a hybrid message-passing and shared-memory approach to parallelism, which enables the code to utilize available computing resources, both among and within computing nodes on a cluster. Shared-memory parallelism is especially valuable for large-model simulations because it enables multiple compute cores within a node to share common model information such as the geometry description, material and composition definitions, and cross section data, thus reducing the memory usage per processor. The parallel performance has been studied by using two complex and representative benchmarks, H-M and BEAVRS benchmark. All simulations were run on Intel Xeon E5-2692 12 core, 2.2 GHz processors at MilkyWay-2 Super Computer with up to 12,288 available computing cores. While employing MPI parallelism, excellent parallel efficiency is achieved even above 12,000 processors and RMC reaches the parallel efficiency of 96% on 12,288 cores for the pin-by-pin computation of the BEAVRS benchmark. While BEAVRS criticality benchmark was selected using MPI/OpenMP parallelism, the speed of 6,144 cores reach 173,314,000 particles per minute. While pin-wise full core H-M benchmark burnup calculations with millions of depletion regions was selected using MPI/OpenMP parallelism, the speed of 6,144 cores reach 1,843,200 burnable regions per minute. Results from these studies demonstrate that RMC, using the hybrid parallel model described in this paper, is able to handle pin-wise full core burnup calculations with millions of depletion regions and scales well through thousands of processors. (authors)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

On-demand data analytics in HPC environments at leadership computing facilities: Challenges and experiences

Scaling-up scientific data analysis and machine learning algorithms for data-driven discovery is a grand challenge that we face today. Despite the growing need for analysis from science domains that are generating `Big Data' from instruments and simulations, building high-performance analytical workflows of data-intensive algorithms have been daunting because: (i) the `Big Data' hardware and software architecture landscape is constantly evolving, (ii) newer architectures impose new programming models, and (iii) data-parallel kernels of analysis algorithms and their performance facets on different architectures are poorly understood. To address these problems, we have: (i) identified scalable data-parallel kernels of popular data analysis algorithms, (ii) implemented `Mini-Apps' of those kernels using different programming models (e.g. Map Reduce, MPI, etc.), (iii) benchmarked and validated the performance of the kernels in diverse architectures. In this paper, we discuss two of those Mini-Apps and show the execution of principal component analysis built as a workflow of the Mini-Apps. We show that Mini-Apps enable scientists to (i) write domain-specific data analysis code that scales on most HPC hardware and (ii) and offers the ability (most times with over a 10x speed-up) to analyze data sizes 100 times the size of what off-the-shelf desktop/workstations of today can handle.

Harney, John↗

Automatic Energy Schemes for High Performance Applications

Although high-performance computing traditionally focuses on the efficient execution of large-scale applications, both energy and power have become critical concerns when approaching exascale. Drastic increases in the power consumption of supercomputers affect significantly their operating costs and failure rates. In modern microprocessor architectures, equipped with dynamic voltage and frequency scaling (DVFS) and CPU clock modulation (throttling), the power consumption may be controlled in software. Additionally, network interconnect, such as Infiniband, may be exploited to maximize energy savings while the application performance loss and frequency switching overheads must be carefully balanced. This work first studies two important collective communication operations, all-to-all and allgather and proposes energy saving strategies on the per-call basis. Next, it targets point-to-point communications to group them into phases and apply frequency scaling to them to save energy by exploiting the architectural and communication stalls. Finally, it proposes an automatic runtime system which combines both collective and point-to-point communications into phases, and applies throttling to them apart from DVFS to maximize energy savings. The experimental results are presented for NAS parallel benchmark problems as well as for the realistic parallel electronic structure calculations performed by the widely used quantum chemistry package GAMESS. Close to the maximum energy savings were obtained with a substantially low performance loss on the given platform.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Bayesian Optimized Deep Ensemble for Uncertainty Quantification of Deep Neural Networks: a System Safety Case Study on Sodium Fast Reactor Thermal Stratification Modeling

Deep neural networks (DNNs) are increasingly important to scientific computing and engineering system simulations. Accurate uncertainty quantification (UQ) for DNNs is critical in safety-sensitive engineering domains. Traditional Deep Ensemble (DE) methods, while easy to implement, frequently suffer from poorly calibrated uncertainty estimates and limited predictive accuracy due to reliance on fixed architectures with varied weight initializations. To address these issues, we introduce a workflow that combines Bayesian Optimization (BO) and DE. The workflow is modular, scalable, and integrates parallel BO initialized with Sobol sequences to individually optimize the hyperparameters of each ensemble member. This method enhances ensemble diversity, improves predictive accuracy, and provides reliable uncertainty estimates. We evaluate the proposed BODE approach in a sodium fast reactor thermal stratification modeling case study, where we used a densely connected convolutional neural network to predict turbulent viscosity during the reactor transient with consideration of data noise. We benchmark its performance against several optimization approaches, including baseline deep ensemble, evolutionary algorithm-optimized ensemble, ensemble formed via random search combined with greedy selection, and a BO ensemble using random initialization. Here, our results demonstrate superior performance of the developed BODE approach. In noise-free scenarios, BODE notably reduces incorrect aleatoric uncertainty and significantly enhances predictive accuracy. Under conditions of 5% and 10% Gaussian noise, BODE adaptively quantifies uncertainty proportional to data noise, achieving up to an 80% reduction in root mean square error compared to baseline methods and producing well-calibrated prediction intervals.

Bayesian optimization↗

Parallel replica dynamics method for bistable stochastic reaction networks: Simulation and sensitivity analysis

Stochastic reaction networks that exhibit bistable behavior are common in systems biology, materials science, and catalysis. Sampling of stationary distributions is crucial for understanding and characterizing the long-time dynamics of bistable stochastic dynamical systems. However, simulations are often hindered by the insufficient sampling of rare transitions between the two metastable regions. In this paper, we apply the parallel replica method for a continuous time Markov chain in order to improve sampling of the stationary distribution in bistable stochastic reaction networks. The proposed method uses parallel computing to accelerate the sampling of rare transitions. Furthermore, it can be combined with the path-space information bounds for parametric sensitivity analysis. With the proposed methodology, we study three bistable biological networks: the Schlögl model, the genetic switch network, and the enzymatic futile cycle network. We demonstrate the algorithmic speedup achieved in these numerical benchmarks. More significant acceleration is expected when multi-core or graphics processing unit computer architectures and programming tools such as CUDA are employed

97 MATHEMATICS AND COMPUTING↗

Algorithms for Balanced Graph Colorings with Applications in Parallel Computing

Graph coloring is used to identify subsets of independent tasks in parallel scientific computing applications. Traditional coloring heuristics aim to reduce the number of colors used as that number also corresponds to the number of parallel steps in the application. However, if the color classes produced have a skew in their sizes, utilization of hardware resources becomes inefficient, especially for the smaller color classes. Equitable coloring is a theoretical formulation of coloring that guarantees a perfect balance among color classes, and its practical relaxation is referred to as balanced coloring. In this paper, we revisit the problem of balanced coloring in the context of parallel computing. The goal is to achieve a balanced coloring of an input graph without increasing the number of colors that an algorithm oblivious to balance would have used. We propose and study multiple heuristics that aim to achieve such a balanced coloring for two variants of coloring problem, distance-1 coloring and partial distance-2 coloring (defined on a bipartite graph). We present parallelization approaches for multi-core and manycore architectures, and cross-evaluate their effectiveness with respect to the quality of balance achieved and performance. Furthermore, we study the impact of the proposed balanced coloring heuristics on a concrete application - viz. parallel community detection, which is an example of an irregular application. In addition, we propose several extensions to our balancing schemes and evaluate their balancing efficacy and performance characteristics. The thorough treatment of balanced coloring presented in this paper from algorithms to application is expected to serve as a valuable resource to parallel application developers who seek to improve parallel performance of their applications using coloring.

graph coloring, multithreaded algorithms↗

Biological neural networks as model systems for designing future parallel processing computers

One of the more interesting debates of the present day centers on whether human intelligence can be simulated by computer. The author works under the premise that neurons individually are not smart at all. Rather, they are physical units which are impinged upon continuously by other matter that influences the direction of voltage shifts across the units membranes. It is only the action of a great many neurons, billions in the case of the human nervous system, that intelligent behavior emerges. What is required to understand even the simplest neural system is painstaking analysis, bit by bit, of the architecture and the physiological functioning of its various parts. The biological neural network studied, the vestibular utricular and saccular maculas of the inner ear, are among the most simple of the mammalian neural networks to understand and model. While there is still a long way to go to understand even this most simple neural network in sufficient detail for extrapolation to computers and robots, a start was made. Moreover, the insights obtained and the technologies developed help advance the understanding of the more complex neural networks that underlie human intelligence.

Ross, Muriel D.↗

Reconfigurable Hardware for Compressing Hyperspectral Image Data

High-speed, low-power, reconfigurable electronic hardware has been developed to implement ICER-3D, an algorithm for compressing hyperspectral-image data. The algorithm and parts thereof have been the topics of several NASA Tech Briefs articles, including Context Modeler for Wavelet Compression of Hyperspectral Images (NPO-43239) and ICER-3D Hyperspectral Image Compression Software (NPO-43238), which appear elsewhere in this issue of NASA Tech Briefs. As described in more detail in those articles, the algorithm includes three main subalgorithms: one for computing wavelet transforms, one for context modeling, and one for entropy encoding. For the purpose of designing the hardware, these subalgorithms are treated as modules to be implemented efficiently in field-programmable gate arrays (FPGAs). The design takes advantage of industry- standard, commercially available FPGAs. The implementation targets the Xilinx Virtex II pro architecture, which has embedded PowerPC processor cores with flexible on-chip bus architecture. It incorporates an efficient parallel and pipelined architecture to compress the three-dimensional image data. The design provides for internal buffering to minimize intensive input/output operations while making efficient use of offchip memory. The design is scalable in that the subalgorithms are implemented as independent hardware modules that can be combined in parallel to increase throughput. The on-chip processor manages the overall operation of the compression system, including execution of the top-level control functions as well as scheduling, initiating, and monitoring processes. The design prototype has been demonstrated to be capable of compressing hyperspectral data at a rate of 4.5 megasamples per second at a conservative clock frequency of 50 MHz, with a potential for substantially greater throughput at a higher clock frequency. The power consumption of the prototype is less than 6.5 W. The reconfigurability (by means of reprogramming) of the FPGAs makes it possible to effectively alter the design to some extent to satisfy different requirements without adding hardware. The implementation could be easily propagated to future FPGA generations and/or to custom application-specific integrated circuits.

Aranki, Nazeeh↗

Biomimetic Fly Eye Sensor for Real-Time Target Tracking

The project objective is to design, build, and demonstrate a biomimetic (inspired by biology) smart sensor based on the common house fly eye for real-time target tracking in three dimensions (3-D). The sensor makes use of revolutionary optical sensor design resulting in significantly improved motion detection capabilities when compared with conventional optical sensors. The design is motivated by the parallel nature of the flys vision system and its demonstrated hyperacuity or precision of visual localization beyond the conventional resolution limit. This means that the sensor is able to accurately detect very fast motion and small changes in motion. The analog architecture of the sensor allows for real-time solution at adjustable bandwidths depending on the application. The low cost, exceptionally lightweight, small form factor, low power requirements, and low computation requirements of the sensor make it ideally suited to many aeronautics applications that have limitations on weight, power, and processing time. The sensor is modular and can be easily adapted in scale and manufactured for a wide variety of sensing applications. Advanced target tracking algorithms will be developed to allow the sensor to track a target in 3-D. These algorithms will be implemented in the smart sensors hardware to enable real-time target tracking.

wing position measurement system↗

A balanced submatrix merging algorithm for multiprocessor architectures

In this article, a parallel algorithm which applies Givens rotations to selectively annihilate k(k + 1)/2 nonzero elements from two k x n(k not more than n) upper trapezoidal submatrices is described. The new algorithm is suitable for implementation on either a pair of directly connected local-memory processors or two clusters of multiple tightly-coupled processors. Analyses show that in both cases the proposed algorithms achieve optimal speed-up by balancing the work load distribution and masking interprocessor or intercluster communication by computation if k is much small than n. In the context of solving large scale least squares problems, this submatrix merging step is repetitively needed during the entire computation and, furthermore, there are usually many pairs of such submatrices to be merged with each submatrix stored in the memory of a processor or a cluster of processors. The proposed algorithm can be applied to each pair of submatrices concurrently, and thus parallelizes an important step in solving the least squares problems.

Chu, Eleanor↗

Who watches the watchers?: preventing fault in a fault tolerance library

The Scalable Checkpoint/Restart library (SCR) was developed and is used by researchers at Lawrence Livermore National Laboratory to provide a fast and efficient method of saving and recovering large applications during runtime on high-performance computing (HPC) systems. Though SCR protects other programs, up until June 2017, nothing was actively protecting SCR. The goal of this project was to automate the building and testing of this library on the varying HPC architectures on which it is used. Our methods centered around the use of a continuous integration tool called Bamboo that allowed for automation agents to be installed on the HPC systems themselves. These agents provided a way for us to establish a new and unique way to automate and customize the allocation of resources and running of tests with CMake’s unit testing framework, CTest, as well as integration testing scripts though an HPC package manager called Spack. These methods provided a parallel environment in which to test the more complex features of SCR. As a result, SCR is now automatically built and tested on several HPC architectures any time changes are made by developers to the library’s source code. The results of these tests are then communicated back to the developers for immediate feedback, allowing them to fix functionality of SCR that may have broken. Hours of developers’ time are now being saved from the tedious process of manually testing and debugging, which saves money and allows the SCR project team to focus their efforts towards development. Thus, HPC system users can use SCR in conjunction with their own applications to efficiently and effectively checkpoint and restart as needed with the assurance that SCR itself is functioning properly.

97 MATHEMATICS AND COMPUTING↗

Multi-core and Many-core Shared-memory Parallel Raycasting Volume Rendering Optimization and Tuning

Given the computing industry trend of increasing processing capacity by adding more cores to a chip, the focus of this work is tuning the performance of a staple visualization algorithm, raycasting volume rendering, for shared-memory parallelism on multi-core CPUs and many-core GPUs. Our approach is to vary tunable algorithmic settings, along with known algorithmic optimizations and two different memory layouts, and measure performance in terms of absolute runtime and L2 memory cache misses. Our results indicate there is a wide variation in runtime performance on all platforms, as much as 254% for the tunable parameters we test on multi-core CPUs and 265% on many-core GPUs, and the optimal configurations vary across platforms, often in a non-obvious way. For example, our results indicate the optimal configurations on the GPU occur at a crossover point between those that maintain good cache utilization and those that saturate computational throughput. This result is likely to be extremely difficult to predict with an empirical performance model for this particular algorithm because it has an unstructured memory access pattern that varies locally for individual rays and globally for the selected viewpoint. Our results also show that optimal parameters on modern architectures are markedly different from those in previous studies run on older architectures. And, given the dramatic performance variation across platforms for both optimal algorithm settings and performance results, there is a clear benefit for production visualization and analysis codes to adopt a strategy for performance optimization through auto-tuning. These benefits will likely become more pronounced in the future as the number of cores per chip and the cost of moving data through the memory hierarchy both increase.

97 MATHEMATICS AND COMPUTING↗

Archive Management of NASA Earth Observation Data to Support Cloud Analysis

NASA collects, processes and distributes petabytes of Earth Observation (EO) data from satellites, aircraft, in situ instruments and model output, with an order of magnitude increase expected by 2024. Cloud-based web object storage (WOS) of these data can simplify the execution of such an increase. More importantly, it can also facilitate user analysis of those volumes by making the data available to the massively parallel computing power in the cloud. However, storing EO data in cloud WOS has a ripple effect throughout the NASA archive system with unexpected challenges and opportunities. One challenge is modifying data servicing software (such as Web Coverage Service servers) to access and subset data that are no longer on a directly accessible file system, but rather in cloud WOS. Opportunities include refactoring of the archive software to a cloud-native architecture; virtualizing data products by computing on demand; and reorganizing data to be more analysis-friendly.

Lynnes, Christopher↗

Archive Management of NASA Earth Observation Data to Support Cloud Analysis

NASA collects, processes and distributes petabytes of Earth Observation (EO) data from satellites, aircraft, in situ instruments and model output, with an order of magnitude increase expected by 2024. Cloud-based web object storage (WOS) of these data can simplify the execution of such an increase. More importantly, it can also facilitate user analysis of those volumes by making the data available to the massively parallel computing power in the cloud. However, storing EO data in cloud WOS has a ripple effect throughout the NASA archive system with unexpected challenges and opportunities. One challenge is modifying data servicing software (such as Web Coverage Service servers) to access and subset data that are no longer on a directly accessible file system, but rather in cloud WOS. Opportunities include refactoring of the archive software to a cloud-native architecture; virtualizing data products by computing on demand; and reorganizing data to be more analysis-friendly. Reviewed by Mark McInerney ESDIS Deputy Project Manager.

Lynnes, Christopher↗