Search NASA⌕ Search

SEARCH · Search NASA

Results for “PARALLEL PROCESSING”

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 307 records · Page 17

Scanning Electrochemical Microscopy: An Evolving Toolbox for Revealing the Chemistry within Electrochemical Processes

The parallel development of ultramicroelectrodes (UMEs) and groundbreaking scanning probe microscopy techniques in the late 1980s led to the development of the scanning electrochemical microscope. Scanning electrochemical microscopy (SECM) was born from the idea of using a tiny electrode to measure the local electrochemical behavior at operating electrodes. From its foundations, the technique displayed an inherent versatility in measuring sample properties beyond topography. It allowed experimenters to measure and map chemical reactions occurring at diverse interfaces, from inspecting the reversibility of redox mediators at metal electrodes, to detecting the hallmarks of cellular respiration on living plant leaves. Related but distinct electrochemical scanning probe techniques, such as electrochemical atomic force microscopy (EC-AFM), scanning ion conductance microscopy (SICM), and scanning electrochemical cell microscopy (SECCM) have developed in parallel. These techniques have demonstrated exquisite spatial resolution down to the nanoscale regime. However, it is the proposition of this review that SECM remains unmatched at revealing the chemical aspects of electrochemistry. Furthermore, it is our intention to review and demonstrate that the versatile architecture of SECM continues to evolve and address fundamental and emerging challenges in the fields of energy storage and conversion, chemical biology, materials science, and environmental chemistry, among others.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Parallel Battery: The Framework and Process for an Intelligent and Ecological Battery System and Related Services

The concept, framework, process methodology and applications of parallel battery were proposed from both virtual and real aspects.The parallel battery was an application of ACP-based parallel intelligence in battery and related energy system areas.The real battery system was running with its equivalent, and the artificial battery system was in a virtual space, in a parallel and interactive manner.The artificial battery system contained the descriptive, predictive, and prescriptive functions on the real battery and related energy systems.There was a closed-loop workflow between the real battery system and the artificial battery system, which iteratively optimizes the battery and related energy systems, leading to a new paradigm of intelligent and ecological parallel battery system management.

ACP approach↗

Large-Scale Welding Process Simulation by GPU Parallelized Computing

The computational design of industrially relevant welded structures is extremely time consuming due to coupled physics and high nonlinearity. Previously, most welding distortion and residual stress simulations have been limited to small coupons and reduced order (from three-dimensional [3D] to two-dimensional [2D]), or inherent strain approximations were used for large structures. In this current study, an explicit finite element code based on a graphics processing unit was utilized to perform 3D transient thermomechanical simulation of structural components during welding. Laser brazing of aluminum alloy panels as representative of automotive manufacturing scenarios was simulated to predict out-of-plane distortion under different clamping conditions. The predicted deformation pattern and magnitude were validated by laser scanning data of physical assemblies. In addition, the code was used to investigate residual stresses developed during multipass arc welding of a nuclear industry pressurizer surge nozzle and subsequent welding repair where a 3D simulation was necessary. Taking the experimental data as reference, the 3D model predicted better residual stress distribution than a typical 2D asymmetrical model. Stress evolution in welding repair was also presented and discussed in this study. Furthermore, the efficient numerical model made it feasible to use integrated computational welding engineering to simulate welding processes for large-scale structures.

97 MATHEMATICS AND COMPUTING↗

Singular value decomposition utilizing parallel algorithms on graphical processors

One of the current challenges in underwater acoustic array signal processing is the detection of quiet targets in the presence of noise. In order to enable robust detection, one of the key processing steps requires data and replica whitening. This, in turn, involves the eigen-decomposition of the sample spectral matrix, Cx = 1/K xKX(k)XH(k) where X(k) denotes a single frequency snapshot with an element for each element of the array. By employing the singular value decomposition (SVD) method, the eigenvectors and eigenvalues can be determined directly from the data without computing the sample covariance matrix, reducing the computational requirements for a given level of accuracy (van Trees, Optimum Array Processing). (Recall that the SVD of a complex matrix A involves determining V, , and U such that A = U VH where U and V are orthonormal and is a positive, real, diagonal matrix containing the singular values of A. U and V are the eigenvectors of AAH and AHA, respectively, while the singular values are the square roots of the eigenvalues of AAH.) Because it is desirable to be able to compute these quantities in real time, an efficient technique for computing the SVD is vital. In addition, emerging multicore processors like graphical processing units (GPUs) are bringing parallel processing capabilities to an ever increasing number of users. Since the computational tasks involved in array signal processing are well suited for parallelization, it is expected that these computations will be implemented using GPUs as soon as users have the necessary computational tools available to them. Thus, it is important to have an SVD algorithm that is suitable for these processors. This work explores the effectiveness of two different parallel SVD implementations on an NVIDIA Tesla C2050 GPU (14 multiprocessors, 32 cores per multiprocessor, 1.15 GHz clock - peed). The first algorithm is based on a two-step algorithm which bidiagonalizes the matrix using Householder transformations, and then diagonalizes the intermediate bidiagonal matrix through implicit QR shifts. This is similar to that implemented for real matrices by Lahabar and Narayanan ("Singular Value Decomposition on GPU using CUDA", IEEE International Parallel Distributed Processing Symposium 2009). The implementation is done in a hybrid manner, with the bidiagonalization stage done using the GPU while the diagonalization stage is done using the CPU, with the GPU used to update the U and V matrices. The second algorithm is based on a one-sided Jacobi scheme utilizing a sequence of pair-wise column orthogonalizations such that A is replaced by AV until the resulting matrix is sufficiently orthogonal (that is, equal to U ). V is obtained from the sequence of orthogonalizations, while can be found from the square root of the diagonal elements of AH A and, once is known, U can be found from column scaling the resulting matrix. These implementations utilize CUDA Fortran and NVIDIA's CUB LAS library. The primary goal of this study is to quantify the comparative performance of these two techniques against themselves and other standard implementations (for example, MATLAB). Considering that there is significant overhead associated with transferring data to the GPU and with synchronization between the GPU and the host CPU, it is also important to understand when it is worthwhile to use the GPU in terms of the matrix size and number of concurrent SVDs to be calculated.

99 GENERAL AND MISCELLANEOUS↗

Parallel Implementation of Nonadditive Gaussian Process Potentials for Monte Carlo Simulations

A strategy is presented to implement Gaussian process potentials in molecular simulations through parallel programming. Attention is focused on the three-body nonadditive energy, though all algorithms extend straightforwardly to the additive energy. The method to distribute pairs and triplets between processes is general to all potentials. Results are presented for a simulation box of argon, including full box and atom displacement calculations, which are relevant to Monte Carlo simulation. Data on speed-up are presented for up to 120 processes across four nodes. A 4-fold speed-up is observed over five processes, extending to 20-fold over 40 processes and 30-fold over 120 processes.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Accelerating population balance-Monte Carlo simulation for coagulation dynamics from the Markov jump model, stochastic algorithm and GPU parallel computing

This paper proposes a comprehensive framework for accelerating population balance-Monte Carlo (PBMC) simulation of particle coagulation dynamics. By combining Markov jump model, weighted majorant kernel and GPU (graphics processing unit) parallel computing, a significant gain in computational efficiency is achieved. The Markov jump model constructs a coagulation-rule matrix of differentially-weighted simulation particles, so as to capture the time evolution of particle size distribution with low statistical noise over the full size range and as far as possible to reduce the number of time loopings. Here three coagulation rules are highlighted and it is found that constructing appropriate coagulation rule provides a route to attain the compromise between accuracy and cost of PBMC methods. Further, in order to avoid double looping over all simulation particles when considering the two-particle events (typically, particle coagulation), the weighted majorant kernel is introduced to estimate the maximum coagulation rates being used for acceptance–rejection processes by single-looping over all particles, and meanwhile the mean time-step of coagulation event is estimated by summing the coagulation kernels of rejected and accepted particle pairs. The computational load of these fast differentially-weighted PBMC simulations (based on the Markov jump model) is reduced greatly to be proportional to the number of simulation particles in a zero-dimensional system (single cell). Finally, for a spatially inhomogeneous multi-dimensional (multi-cell) simulation, the proposed fast PBMC is performed in each cell, and multiple cells are parallel processed by multi-cores on a GPU that can implement the massively threaded data-parallel tasks to obtain remarkable speedup ratio (comparing with CPU computation, the speedup ratio of GPU parallel computing is as high as 200 in a case of 100 cells with 10 000 simulation particles per cell). These accelerating approaches of PBMC are demonstrated in a physically realistic Brownian coagulation case. The computational accuracy is validated with benchmark solution of discrete-sectional method. The simulation results show that the comprehensive approach can attain very favorable improvement in cost without sacrificing computational accuracy.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A seamless acquisition digital storage oscilloscope with three-dimensional waveform display

In traditional digital storage oscilloscope (DSO), sampled data need to be processed after each acquisition. During data processing, the acquisition is stopped and oscilloscope is blind to the input signal. Thus, this duration is called dead time. With the rapid development of modern electronic systems, the effect of infrequent events becomes significant. To capture these occasional events in shorter time, dead time in traditional DSO that causes the loss of measured signal needs to be reduced or even eliminated. In this paper, a seamless acquisition oscilloscope without dead time is proposed. In this oscilloscope, three-dimensional waveform mapping (TWM) technique, which converts sampled data to displayed waveform, is proposed. With this technique, not only the process speed is improved, but also the probability information of waveform is displayed with different brightness. Thus, a three-dimensional waveform is shown to the user. To reduce processing time further, parallel TWM which processes several sampled points simultaneously, and dual-port random access memory based pipelining technique which can process one sampling point in one clock period are proposed. Furthermore, two DDR3 (Double-Data-Rate Three Synchronous Dynamic Random Access Memory) are used for storing sampled data alternately, thus the acquisition can continue during data processing. Therefore, the dead time of DSO is eliminated. In addition, a double-pulse test method is adopted to test the waveform capturing rate (WCR) of the oscilloscope and a combined pulse test method is employed to evaluate the oscilloscope's capture ability comprehensively. The experiment results show that the WCR of the designed oscilloscope is 6 250 000 wfms/s (waveforms per second), the highest value in all existing oscilloscopes. The testing results also prove that there is no dead time in our oscilloscope, thus realizing the seamless acquisition.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Development Status of the PEBBLES Code for Pebble Mechanics: Improved Physical Models and Speed-up

PEBBLES is a code for simulating the motion of all the pebbles in a pebble bed reactor. Since pebble bed reactors are packed randomly and not precisely placed, the location of the fuel elements in the reactor is not deterministically known. Instead, when determining operating parameters the motion of the pebbles can be simulated and stochastic locations can be found. The PEBBLES code can output information relevant for other simulations of the pebble bed reactors such as the positions of the pebbles in the reactor, packing fraction change in an earthquake, and velocity profiles created by recirculation. The goal for this level three milestone was to speedup the PEBBLES code through implementation on massively parallel computer. Work on this goal has resulted in speeding up both the single processor version and creation of a new parallel version of PEBBLES. Both the single processor version and the parallel running capability of the PEBBLES code have improved since the fiscal year start. The hybrid MPI/OpenMP PEBBLES version was created this year to run on the increasingly common cluster hardware profile that combines nodes with multiple processors that share memory and a cluster of nodes that are networked together. The OpenMP portions use the Open Multi-Processing shared memory parallel processing model to split the task across processors in a single node that shares memory. The Message Passing Interface (MPI) portion uses messages to communicate between different nodes over a network. The following are wall clock speed up for simulating an NGNP-600 sized reactor. The single processor version runs 1.5 times faster compared to the single processor version at the beginning of the fiscal year. This speedup is primarily due to the improved static friction model described in the report. When running on 64 processors, the new MPI/OpenMP hybrid version has a wall clock speed up of 22 times compared to the current single processor version. When using 88 processors, a speed up of 23 times is achieved. This speedup and other improvements of PEBBLES combine to make PEBBLES more capable and more useful for simulation of a pebble bed reactor. This report details the implementation and effects of the speedup work done over the course of the fiscal year.

10 SYNTHETIC FUELS↗

Rapid evaluation and quality control of next generation sequencing data with FaQCs

Background: Next generation sequencing (NGS) technologies that parallelize the sequencing process and produce thousands to millions, or even hundreds of millions of sequences in a single sequencing run, have revolutionized genomic and genetic research. Because of the vagaries of any platform's sequencing chemistry, the experimental processing, machine failure, and so on, the quality of sequencing reads is never perfect, and often declines as the read is extended. These errors invariably affect downstream analysis/application and should therefore be identified early on to mitigate any unforeseen effects. Results: Here we present a novel FastQ Quality Control Software (FaQCs) that can rapidly process large volumes of data, and which improves upon previous solutions to monitor the quality and remove poor quality data from sequencing runs. Both the speed of processing and the memory footprint of storing all required information have been optimized via algorithmic and parallel processing solutions. The trimmed output compared side-by-side with the original data is part of the automated PDF output. We show how this tool can help data analysis by providing a few examples, including an increased percentage of reads recruited to references, improved single nucleotide polymorphism identification as well as de novo sequence assembly metrics. Conclusion: FaQCs combines several features of currently available applications into a single, user-friendly process, and includes additional unique capabilities such as filtering the PhiX control sequences, conversion of FASTQ formats, and multi-threading. The original data and trimmed summaries are reported within a variety of graphics and reports, providing a simple way to do data quality control and assurance.

59 BASIC BIOLOGICAL SCIENCES↗

Automatic identification and classification of Palomar Transient Factory astrophysical objects in GLADE

Palomar Transient Factory (PTF) is a comprehensive detection system for the identification and classification of transient astrophysical objects. In this paper, we make two significant contributions to the PTF pipeline. First, we present an experimental study that evaluates a novel implementation of the real-time classifier in GLADE - a parallel data processing system that combines the efficiency of a database with the extensibility of map-reduce. We show how each stage in the classifier maps optimally into GLADE tasks by taking advantage of the unique features of the system - range-based data partitioning, columnar storage, multi-query execution, and in-database support for complex aggregate computation. Second, we introduce a novel parallel similarity join algorithm for advanced transient classification. In conclusion, we implement this algorithm in GLADE and execute it on a massive supercomputer with more than 3,000 threads, achieving more than three orders of magnitude improvement over the PostgreSQL solution.

79 ASTRONOMY AND ASTROPHYSICS↗

Status of the Development of In-Tank/At-Tank Separations Technologies for High-Level Waste Processing for the U.S. Department of Energy

Within the U.S. Department of Energy’s (DOE) Office of Technology Innovation and Development, the Office of Waste Processing manages a research and development program related to the treatment and disposition of radioactive waste. At the Savannah River (South Carolina) and Hanford (Washington) Sites, approximately 90 million gallons of waste are distributed among 226 storage tanks (grouped or collocated in “tank farms”). This waste may be considered to contain mixed and stratified high activity and low activity constituent waste liquids, salts and sludges that are collectively managed as high level waste (HLW). A large majority of these wastes and associated facilities are unique to the DOE, meaning many of the programs to treat these materials are “first-of-a-kind” and unprecedented in scope and complexity. As a result, the technologies required to disposition these wastes must be developed from basic principles, or require significant re-engineering to adapt to DOE’s specific applications. Of particular interest recently, the development of In-tank or At-Tank separation processes have the potential to treat waste with high returns on financial investment. The primary objective associated with In-Tank or At-Tank separation processes is to accelerate waste processing. Insertion of the technologies will (1) maximize available tank space to efficiently support permanent waste disposition including vitrification; (2) treat problematic waste prior to transfer to the primary processing facilities at either site (i.e., Hanford’s Waste Treatment and Immobilization Plant (WTP) or Savannah River’s Salt Waste Processing Facility (SWPF)); and (3) create a parallel treatment process to shorten the overall treatment duration.

High level waste↗

STATUS OF THE DEVELOPMENT OF IN-TANK/AT-TANK SEPARATIONS TECHNOLOGIES FOR FOR HIGH-LEVEL WASTE PROCESSING FOR THE U.S. DEPARTMENT OF ENERGY

Within the U.S. Department of Energy's (DOE) Office of Technology Innovation and Development, the Office of Waste Processing manages a research and development program related to the treatment and disposition of radioactive waste. At the Savannah River (South Carolina) and Hanford (Washington) Sites, approximately 90 million gallons of waste are distributed among 226 storage tanks (grouped or collocated in 'tank farms'). This waste may be considered to contain mixed and stratified high activity and low activity constituent waste liquids, salts and sludges that are collectively managed as high level waste (HLW). A large majority of these wastes and associated facilities are unique to the DOE, meaning many of the programs to treat these materials are 'first-of-a-kind' and unprecedented in scope and complexity. As a result, the technologies required to disposition these wastes must be developed from basic principles, or require significant re-engineering to adapt to DOE's specific applications. Of particular interest recently, the development of In-tank or At-Tank separation processes have the potential to treat waste with high returns on financial investment. The primary objective associated with In-Tank or At-Tank separation processes is to accelerate waste processing. Insertion of the technologies will (1) maximize available tank space to efficiently support permanent waste disposition including vitrification; (2) treat problematic waste prior to transfer to the primary processing facilities at either site (i.e., Hanford's Waste Treatment and Immobilization Plant (WTP) or Savannah River's Salt Waste Processing Facility (SWPF)); and (3) create a parallel treatment process to shorten the overall treatment duration. This paper will review the status of several of the R&D projects being developed by the U.S. DOE including insertion of the ion exchange (IX) technologies, such as Small Column Ion Exchange (SCIX) at Savannah River. This has the potential to align the salt and sludge processing life cycle, thereby reducing the Defense Waste Processing Facility (DWPF) mission by 7 years. Additionally at the Hanford site, problematic waste streams, such as high boehmite and phosphate wastes, could be treated prior to receipt by WTP and thus dramatically improve the capacity of the facility to process HLW. Treatment of boehmite by continuous sludge leaching (CSL) before receipt by WTP will dramatically reduce the process cycle time for the WTP pretreatment facility, while treatment of phosphate will significantly reduce the number of HLW borosilicate glass canisters produced at the WTP. These and other promising technologies will be discussed.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

A Monte Carlo neutron transport code for eigenvalue calculations on a dual-GPU system and CUDA environment

Monte Carlo (MC) method is able to accurately calculate eigenvalues in reactor analysis. Its lengthy computation time can be reduced by general-purpose computing on Graphics Processing Units (GPU), one of the latest parallel computing techniques under development. The method of porting a regular transport code to GPU is usually very straightforward due to the 'embarrassingly parallel' nature of MC code. However, the situation becomes different for eigenvalue calculation in that it will be performed on a generation-by-generation basis and the thread coordination should be explicitly taken care of. This paper presents our effort to develop such a GPU-based MC code in Compute Unified Device Architecture (CUDA) environment. The code is able to perform eigenvalue calculation under simple geometries on a multi-GPU system. The specifics of algorithm design, including thread organization and memory management were described in detail. The original CPU version of the code was tested on an Intel Xeon X5660 2.8 GHz CPU, and the adapted GPU version was tested on NVIDIA Tesla M2090 GPUs. Double-precision floating point format was used throughout the calculation. The result showed that a speedup of 7.0 and 33.3 were obtained for a bare spherical core and a binary slab system respectively. The speedup factor was further increased by a factor of {approx}2 on a dual GPU system. The upper limit of device-level parallelism was analyzed, and a possible method to enhance the thread-level parallelism was proposed. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Data-Driven Compositional Optimization in Misspecified Regimes

With a manifold growth in the scale and intricacy of systems, the challenges of parametric misspecification become pronounced. These concerns are further exacerbated in compositional settings, which emerge in problems complicated by modeling risk and robustness. In “Data-Driven Compositional Optimization in Misspecified Regimes,” the authors consider the resolution of compositional stochastic optimization problems, plagued by parametric misspecification. In considering settings where such misspecification may be resolved via a parallel learning process, the authors develop schemes that can contend with diverse forms of risk, dynamics, and nonconvexity. They provide asymptotic and rate guarantees for unaccelerated and accelerated schemes for convex, strongly convex, and nonconvex problems in a two-level regime with extensions to the multilevel setting. Surprisingly, the nonasymptotic rate guarantees show no degradation from the rate statements obtained in a correctly specified regime and the schemes achieve optimal (or near-optimal) sample complexities for general T-level strongly convex and nonconvex compositional problems.

Business & Economics↗

Hybrid AI/Cognitive Tactical Behavior Framework for LVC (LDRD Project Final Report)

This Lab-Directed Research and Development (LDRD) sought to develop technology that enhances scenario construction speed, entity behavior robustness, and scalability in Live-Virtual-Constructive (LVC) simulation. We investigated issues in both simulation architecture and behavior modeling. We developed path-planning technology that improves the ability to express intent in the planning task while still permitting an efficient search algorithm. An LVC simulation demonstrated how this enables 'one-click' layout of squad tactical paths, as well as dynamic re-planning for simulated squads and for real and simulated mobile robots. We identified human response latencies that can be exploited in parallel/distributed architectures. We did an experimental study to determine where parallelization would be productive in Umbra-based force-on-force (FOF) simulations. We developed and implemented a data-driven simulation composition approach that solves entity class hierarchy issues and supports assurance of simulation fairness. Finally, we proposed a flexible framework to enable integration of multiple behavior modeling components that model working memory phenomena with different degrees of sophistication.

99 GENERAL AND MISCELLANEOUS↗

Scalable Communication Trace Compression

Characterizing the communication behavior of parallel programs through tracing can help understand an application’s characteristics, model its performance, and predict behavior on future systems. However, lossless communication traces can get prohibitively large, causing programmers to resort to variety of other techniques. In this paper, we present a novel approach to lossless communication trace compression. We augment the sequitur compression algorithm to employ it in communication trace compression of parallel programs. We present optimizations to reduce the memory overhead, reduce size of the trace files generated, and enable compression across multiple processes in a parallel program. The evaluation shows improved compression and reduced overhead over other approaches, with up to 3 orders of magnitude improvement for the NAS MG benchmark. We also observe that, unlike existing schemes, the trace files sizes and the memory overhead incurred are less sensitive to, if not independent of, the problem size for the NAS benchmarks.

97 MATHEMATICS AND COMPUTING↗

A theoretical approach to electric breakdown behaviour and effective secondary emission coefficient in gas discharge process between two parallel-plane disk electrodes

The gas discharge characteristics are seriously influenced by the effective secondary electron-emission coefficient ({gamma}{sub eff}) originally defined by Townsend. A theoretical approach has been proposed for the determination of variation of this coefficient as a function of the geometrical parameter d/R (the ratio of inter-electrode distance to electrode radius) in the most practical range of 0.08{<=}d/R{<=}1.90. Two geometrical factors g{sub i} and g{sub p}, which are attributed to ion-induced and photon-induced terms, respectively, were introduced to the effective gamma in parallel-plane arrangement. The factors were theoretically derived and hence the effective gamma and the breakdown voltage for a specified geometrical parameter, such as d/R=0.16, were obtained. Then, the effective gamma and the Paschen curves for various geometrical parameters such as d/R=0.32, 0.63, 1.27, and 1.90 were theoretically reproduced. The obtained results showed quite a good agreement with those reported by other researchers and verified that the proposed g{sub i} and g{sub p} have properly been introduced and derived.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Scioto: A Framework for Global-ViewTask Parallelism

We introduce Scioto, Shared Collections of Task Objects, a framework for supporting task-parallelism in one-sided and global-view parallel programming models. Scioto provides lightweight, locality aware dynamic load balancing and interoperates with existing parallel models including MPI, SHMEM, CAF, and Global Arrays. Through task parallelism, the Scioto framework provides a solution for overcoming load imbalance and heterogeneity as well as dynamic mapping of computation onto emerging multicore architectures. In this paper, we present the design and implementation of the Scioto framework and demonstrate its effectiveness on the Unbalanced Tree Search (UTS) benchmark and two quantum chemistry codes: the closed shell Self-Consistent Field (SCF) method and a sparse tensor contraction kernel extracted from a coupled cluster computation. We explore the efficiency and scalability of Scioto through these sample applications and demonstrate that is offers low overhead, achieves good performance on heterogeneous and multicore clusters, and scales to hundreds of processors.

99 GENERAL AND MISCELLANEOUS↗