Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallel processors”

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 1,603 records · Page 89

Parallel multi-time step integration on a transputer system

An algorithm which allows different nodes of the finite element mesh to be integrated with different time steps is presented for second order finite element systems. The implementation of this algorithm on a system of transputer processors is discussed and a numerical example is used to evaluate the efficiency of the algorithm.

Smolinski, P.↗

Message Passing and Shared Address Space Parallelism on an SMP Cluster

Currently, message passing (MP) and shared address space (SAS) are the two leading parallel programming paradigms. MP has been standardized with MPI, and is the more common and mature approach; however, code development can be extremely difficult, especially for irregularly structured computations. SAS offers substantial ease of programming, but may suffer from performance limitations due to poor spatial locality and high protocol overhead. In this paper, we compare the performance of and the programming effort required for six applications under both programming models on a 32-processor PC-SMP cluster, a platform that is becoming increasingly attractive for high-end scientific computing. Our application suite consists of codes that typically do not exhibit scalable performance under shared-memory programming due to their high communication-to-computation ratios and/or complex communication patterns. Results indicate that SAS can achieve about half the parallel efficiency of MPI for most of our applications, while being competitive for the others. A hybrid MPI+SAS strategy shows only a small performance advantage over pure MPI in some cases. Finally, improved implementations of two MPI collective operations on PC-SMP clusters are presented.

Shan, Hongzhang↗

A 2nd Generation Parallel Advancing Front Grid Generator

A scalable, parallel advancing grid generation technique has been developed for complex geometries and meshes with large size variations. The key innovation compared to previous techniques is the use of a domain-defining grid that has the same fine surface triangulation as the final mesh desired, but a much coarser interior mesh. In this way, the domain to be gridded is uniquely defined, overcoming a shortcoming of previous approaches. This domain-defining grid is then partitioned according to the estimated number of elements to be generated, allowing for a balanced distribution of work among the processors. The domain defining grid is also used to redistribute the elements and points after grid generation, and during the subsequent mesh improvement. Timings show that the proposed approach is scalable and able to produce large grids of high quality in a modest amount of clocktime. With the proposed parallel grid generator, a major impediment to a completely scalable simulation pipeline (grid generation, solvers, post-processing) has been removed, opening the way for truly large-scale computations using unstructured, body-fitted grids.

Grid Generation↗

High-Performance, Multi-Node File Copies and Checksums for Clustered File Systems

Modern parallel file systems achieve high performance using a variety of techniques, such as striping files across multiple disks to increase aggregate I/O bandwidth and spreading disks across multiple servers to increase aggregate interconnect bandwidth. To achieve peak performance from such systems, it is typically necessary to utilize multiple concurrent readers/writers from multiple systems to overcome various singlesystem limitations, such as number of processors and network bandwidth. The standard cp and md5sum tools of GNU coreutils found on every modern Unix/Linux system, however, utilize a single execution thread on a single CPU core of a single system, and hence cannot take full advantage of the increased performance of clustered file systems. Mcp and msum are drop-in replacements for the standard cp and md5sum programs that utilize multiple types of parallelism and other optimizations to achieve maximum copy and checksum performance on clustered file systems. Multi-threading is used to ensure that nodes are kept as busy as possible. Read/write parallelism allows individual operations of a single copy to be overlapped using asynchronous I/O. Multinode cooperation allows different nodes to take part in the same copy/checksum. Split-file processing allows multiple threads to operate concurrently on the same file. Finally, hash trees allow inherently serial checksums to be performed in parallel. Mcp and msum provide significant performance improvements over standard cp and md5sum using multiple types of parallelism and other optimizations. The total speed-ups from all improvements are significant. Mcp improves cp performance over 27x, msum improves md5sum performance almost 19x, and the combination of mcp and msum improves verified copies via cp and md5sum by almost 22x. These improvements come in the form of drop-in replacements for cp and md5sum, so are easily used and are available for download as open source software at http://mutil.sourceforge.net.

Kolano, Paul Z.↗

Algorithms and Libraries

This exploratory study initiated our inquiry into algorithms and applications that would benefit by latency tolerant approach to algorithm building, including the construction of new algorithms where appropriate. In a multithreaded execution, when a processor reaches a point where remote memory access is necessary, the request is sent out on the network and a context--switch occurs to a new thread of computation. This effectively masks a long and unpredictable latency due to remote loads, thereby providing tolerance to remote access latency. We began to develop standards to profile various algorithm and application parameters, such as the degree of parallelism, granularity, precision, instruction set mix, interprocessor communication, latency etc. These tools will continue to develop and evolve as the Information Power Grid environment matures. To provide a richer context for this research, the project also focused on issues of fault-tolerance and computation migration of numerical algorithms and software. During the initial phase we tried to increase our understanding of the bottlenecks in single processor performance. Our work began by developing an approach for the automatic generation and optimization of numerical software for processors with deep memory hierarchies and pipelined functional units. Based on the results we achieved in this study we are planning to study other architectures of interest, including development of cost models, and developing code generators appropriate to these architectures.

Dongarra, Jack↗

Scaling the SIESTA Magnetohydrodynamics Equilibrium Code

We report the results of a scaling effort that increases both the speed and resolution of the SIESTA magnetohydrodynamic equilibrium code. SIESTA is capable of computing three-dimensional plasma equilibria with magnetic islands at high spatial resolutions for toroidally confined plasmas. Starting with a small-scale parallel im- plementation, we identified scale-dependent bottlenecks of the code and developed scalable alternatives for each performance-significant functionality, cumulatively increasing the runtime speed and scalability with the number of processors. The net outcome is an improvement in speed by over ten-fold, utilizing a few thousand processors, en- abling SIESTA to simulate high spatial-resolution scenarios in under an hour for the first time.

Seal, Sudip K↗

Extension of 4-8 Texture Hierarchies to Large Video Processing and Visualization

The purpose of this Techbase was to reduce to practice the tiled 4-8 texture hierarchy for the display of video imagery (i.e. sequences of frames). The immediate intent was to demonstrate its use in the analysis and display of sensor imagery. As sensors are increasing in resolution the physical amount of imagery that needs to be displayed can quickly overwhelm most display systems. For example, a sensor with a horizontal resolution of over 8000 pixels would generate an image over 10 feet wide on a standard 72 DPI display. Breaking an image into tiles, and then decomposing each tile into a multiresolution hierarchy, allows a user (or software) to efficiently select and display only those parts of the image that are of interest to the user. The originator of the idea of 4-8 Texture Hierarchies was Dr. Mark Duchaineau, and we consulted with him in much of our work. We also consulted with Dan Knight, from SequoiaTek Corp., who is a contractor responsible for implementing the viewers for our applications. Most of the code for actual 4-8 Texture Hierarchy generation already existed; a large focus of the Techbase was to determine how to best use what was available for video imagery. The majority of progress was made in specifying and implementing the software framework, which turned out to be rather involved. This framework is to support the creation, storage, and display of images, both tiled and untiled. A first albeit incomplete version was successfully tested in the field in August 2007. The framework structures the process of collecting and processing images conceptually as a pipeline, where work is passed along and a different operation is performed at each stage. In practice, the pipeline is implemented by a group of processes (not threads), or 'workers', each responsible for a specific type of operation. Associated with these workers is a pool of memory (cache). As each process finishes its work, it places the results into the cache and sends a message to the process responsible for the next operation, including in the message an identifier allowing the recipient to locate the data to be worked on. Conceptually this is akin to transferring mail from person to person by placing it into and removing it from a group of mail slots, rather than handing it to the next person directly. So for example, to capture, tile, and store imagery to disk, one worker would be responsible for obtaining an image from an electro-optical sensor. Another would be responsible for breaking the image into tiles, and another for performing a decorrelating transform for entropy reduction and multiresolution representation. Still another would be responsible for compression, and another for storage. To maximize robustness in an experimental environment, we created a manager that could monitor, kill, restart, and configure the entire processing pipeline, including the shared cache and worker processes. In the event of hardware or software failure, individual elements of the pipeline were restored automatically and with minimum data loss. Working with Mark Duchaineau we attempted to determine a set of ideal coefficients for the wavelet used to create the multiresolution hierarchy. While a workable set exists, it produces undesirable artifacts at high levels of loss, so more work needs to be done to identify a more ideal set. The tiler and data structures for tiling and display have been added to the code base we developed. As currently implemented it takes approximately one second to tile and build a multiresolution hierarchy for an 11-megapixel image on a 3.8 GHz Pentium 4. This is too slow for some uses, so some work will probably need to be done to optimize the code. We may be able to leverage commodity graphics processors to gain extra speed. With the advent of quad-core CPUs it may also be possible to get required throughput by tiling the image and then having several processes run in parallel to build the multiresolution hierarchies for the resulting tiles. Additionally, more work needs to be done to define a data format for storing image tiles on disk, as well as retrieving and displaying them. As this functionality is desirable (and required, if we are to make large data easily viewable) we will be spending a lot of time developing this. As a final note, it was recently brought to our attention that Oracle provides an extension to its database technology that provides most, if not all, of the functionality that we are working on as part of this project. We will be investigating this further; if a suitable solution already exists it may be more valuable time-wise to integrate it into projects that require storing and displaying large images.

99 GENERAL AND MISCELLANEOUS↗

Formulation of consumables management models. Volume 1: Mission planning

Development of an STS (Space Transportation System) interactive computer program MPP (Mission Planning Processor) working model was conducted. A summary of the computer program development and those supporting tasks conducted is presented. Development of the MPP Computer Program is discussed. This development was supported by several parallel tasks. These tasks either directly supported the program development, or provided information for future application and/or modification to the program in relation to the flight planning and flight operations of the STS and advanced spacecraft. The supporting tasks also included development of a Space Station MPP to demonstrate the applicability of the analytical methods developed under this RTOP to more advanced spacecraft than the STS.

Torian, J. G.↗

Optimal parallel algorithms for problems modeled by a family of intervals

A family of intervals on the real line provides a natural model for a vast number of scheduling and VLSI problems. Recently, a number of parallel algorithms to solve a variety of practical problems on such a family of intervals have been proposed in the literature. Computational tools are developed, and it is shown how they can be used for the purpose of devising cost-optimal parallel algorithms for a number of interval-related problems including finding a largest subset of pairwise nonoverlapping intervals, a minimum dominating subset of intervals, along with algorithms to compute the shortest path between a pair of intervals and, based on the shortest path, a parallel algorithm to find the center of the family of intervals. More precisely, with an arbitrary family of n intervals as input, all algorithms run in O(log n) time using O(n) processors in the EREW-PRAM model of computation.

Olariu, Stephan↗

An Atmospheric General Circulation Model with Chemistry for the CRAY T3E: Design, Performance Optimization and Coupling to an Ocean Model

The design, implementation and performance optimization on the CRAY T3E of an atmospheric general circulation model (AGCM) which includes the transport of, and chemical reactions among, an arbitrary number of constituents is reviewed. The parallel implementation is based on a two-dimensional (longitude and latitude) data domain decomposition. Initial optimization efforts centered on minimizing the impact of substantial static and weakly-dynamic load imbalances among processors through load redistribution schemes. Recent optimization efforts have centered on single-node optimization. Strategies employed include loop unrolling, both manually and through the compiler, the use of an optimized assembler-code library for special function calls, and restructuring of parts of the code to improve data locality. Data exchanges and synchronizations involved in coupling different data-distributed models can account for a significant fraction of the running time. Therefore, the required scattering and gathering of data must be optimized. In systems such as the T3E, there is much more aggregate bandwidth in the total system than in any particular processor. This suggests a distributed design. The design and implementation of a such distributed 'Data Broker' as a means to efficiently couple the components of our climate system model is described.

Farrara, John D.↗

The P-Mesh: A Commodity-based Scalable Network Architecture for Clusters

We designed a new network architecture, the P-Mesh which combines the scalability and fault resilience of a torus with the performance of a switch. We compare the scalability, performance, and cost of the hub, switch, torus, tree, and P-Mesh architectures. The latter three are capable of scaling to thousands of nodes, however, the torus has severe performance limitations with that many processors. The tree and P-Mesh have similar latency, bandwidth, and bisection bandwidth, but the P-Mesh outperforms the switch architecture (a lower bound for tree performance) on 16-node NAB Parallel Benchmark tests by up to 23%, and costs 40% less. Further, the P-Mesh has better fault resilience characteristics. The P-Mesh architecture trades increased management overhead for lower cost, and is a good bridging technology while the price of tree uplinks is expensive.

Nitzberg, Bill↗

High precision computing with charge domain devices and a pseudo-spectral method therefor

The present invention enhances the bit resolution of a CCD/CID MVM processor by storing each bit of each matrix element as a separate CCD charge packet. The bits of each input vector are separately multiplied by each bit of each matrix element in massive parallelism and the resulting products are combined appropriately to synthesize the correct product. In another aspect of the invention, such arrays are employed in a pseudo-spectral method of the invention, in which partial differential equations are solved by expressing each derivative analytically as matrices, and the state function is updated at each computation cycle by multiplying it by the matrices. The matrices are treated as synaptic arrays of a neural network and the state function vector elements are treated as neurons. In a further aspect of the invention, moving target detection is performed by driving the soliton equation with a vector of detector outputs. The neural architecture consists of two synaptic arrays corresponding to the two differential terms of the soliton-equation and an adder connected to the output thereof and to the output of the detector array to drive the soliton equation.

Barhen, Jacob↗

Quantum Monte Carlo Endstation for Petascale Computing

The major achievements enabled by QMC Endstation grant include * Performance improvement on clusters of x86 multi-core systems, especially on Cray XT systems * New and improved methods for the wavefunction optimizations * New forms of trial wavefunctions * Implementation of the full application on NVIDIA GPUs using CUDA The scaling studies of QMCPACK on large-scale systems show excellent parallel efficiency up to 216K cores on Jaguarpf (Cray XT5). The GPU implementation shows speedups of 10-15x over the CPU implementation on older generation of x86. We have implemented hybrid OpenMP/MPI scheme in QMC to take advantage of multi-core shared memory processors of petascale systems. Our hybrid scheme has several advantages over the standard MPI-only scheme. * Memory optimized: large read-only data to store one-body orbitals and other shared properties to represent the trial wave function and many-body Hamiltonian can be shared among threads, which reduces the memory footprint of a large-scale problem. * Cache optimized: the data associated with an active Walker are in cache during the compute-intensive drift-diffusion process and the operations on an Walker are optimized for cache reuse. Thread-local objects are used to ensure the data affinity to a thread. * Load balanced: Walkers in an ensemble are evenly distributed among threads and MPI tasks. The two-level parallelism reduces the population imbalance among MPI tasks and reduces the number of point-to-point communications of large messages (serialized objects) for the Walker exchange. * Communication optimized: the communication overhead, especially for the collective operations necessary to determine ET and measure the properties of an ensemble, is significantly lowered by using less MPI tasks. The multiple forms of parallelism afforded by QMC algorithms make them ideal candidates for acceleration in the many-core paradigm. We presented the results of our effort to port the QMCPACK simulation code to the NVIDIA CUDA GPU platform. We restructured the CPU algorithms to express additional parallelism, minimize GPU-CPU communication, and efficiently utilize the GPU memory hierarchy. Using mixed precision on GT200 GPUs and MPI for intercommunication and load balancing, we observe typical full-application speedups of approximately 10x to 15x relative to quad-core Xeon CPUs alone, while reproducing the double-precision CPU results within statistical error. We developed an all-electron quantum Monte Carlo (QMC) method for solids that does not rely on pseudopotentials, and used it to construct a primary ultra-high-pressure calibration based on the equation of state of cubic boron nitride. We computed the static contribution to the free energy with the QMC method and obtained the phonon contribution from density functional theory, yielding a high-accuracy calibration up to 900 GPa usable directly in experiment. We computed the anharmonic Raman frequency shift with QMC simulations as a function of pressure and temperature, allowing optical pressure calibration. In contrast to present experimental approaches, small systematic errors in the theoretical EOS do not increase with pressure, and no extrapolation is needed. This all-electron method is applicable to first-row solids, providing a new reference for ab initio calculations of solids and benchmarks for pseudopotential accuracy. We compared experimental and theoretical results on the momentum distribution and the quasiparticle renormalization factor in sodium. From an x-ray Compton-profile measurement of the valence-electron momentum density, we derived its discontinuity at the Fermi wavevector finding an accurate measure of the renormalization factor that we compared with quantum-Monte-Carlo and G0W0 calculations performed both on crystalline sodium and on the homogeneous electron gas. Our calculated results are in good agreement with the experiment. We have been studying the heat of formation for various Kubas complexes of molecular hydrogen on Ti(1,2)ethylene-nH2 using Diffusion Monte Carlo. This work has been started and is ongoing. We are studying systems involving 1 and 2 Ti bonding sites with up to 10 hydrogen molecules in numerous configurations. This work will establish a benchmark that will test the accuracy of density functional calculations and establish the feasibility of our methods for similar systems.

97 MATHEMATICS AND COMPUTING↗

Parallel processing methods for space based power systems

This report presents a method for doing load-flow analysis of a power system by using a decomposition approach. The power system for the Space Shuttle is used as a basis to build a model for the load-flow analysis. To test the decomposition method for doing load-flow analysis, simulations were performed on power systems of 16, 25, 34, 43, 52, 61, 70, and 79 nodes. Each of the power systems was divided into subsystems and simulated under steady-state conditions. The results from these tests have been found to be as accurate as tests performed using a standard serial simulator. The division of the power systems into different subsystems was done by assigning a processor to each area. There were 13 transputers available, therefore, up to 13 different subsystems could be simulated at the same time. This report has preliminary results for a load-flow analysis using a decomposition principal. The report shows that the decomposition algorithm for load-flow analysis is well suited for parallel processing and provides increases in the speed of execution.

Berry, F. C.↗

Skel: Generative Software for Producing Skeletal I/O Applications

Massively parallel computations consist of a mixture of computation, communication, and I/O. As part of the co-design for the inevitable progress towards exascale computing, we must apply lessons learned from past work to succeed in this new age of computing. Of the three components listed above, implementing an effective parallel I/O solution has often been overlooked by application scientists and was usually added to large scale simulations only when existing serial techniques had failed. As scientists teams scaled their codes to run on hundreds of processors, it was common to call on an I/O expert to implement a set of more scalable I/O routines. These routines were easily separated from the calculations and communication, and in many cases, an I/O kernel was derived from the application which could be used for testing I/O performance independent of the application. These I/O kernels developed a life of their own used as a broad measure for comparing different I/O techniques. Unfortunately, as years passed and computation and communication changes required changes to the I/O, the separate I/O kernel used for benchmarking remained static no longer providing an accurate indicator of the I/O performance of the simulation making I/O research less relevant for the application scientists. In this paper we describe a new approach to this problem where I/O kernels are replaced with skeletal I/O applications automatically generated from an abstract set of simulation I/O parameters. We realize this abstraction by leveraging the ADIOS middleware's XML I/O specification with additional runtime parameters. Skeletal applications offer all of the benefits of I/O kernels including allowing I/O optimizations to focus on useful I/O patterns. Moreover, since they are automatically generated, it is easy to produce an updated I/O skeleton whenever the simulation's I/O changes. In this paper we analyze the performance of automatically generated I/O skeletal applications for the S3D and GTS codes. We show that these skeletal applications achieve performance comparable to that of the production applications. We wrap up the paper with a discussion of future changes to make the skeletal application better approximate the actual I/O performed in the simulation.

99 GENERAL AND MISCELLANEOUS↗

Scalability of Parallel Spatial Direct Numerical Simulations on Intel Hypercube and IBM SP1 and SP2

The implementation and performance of a parallel spatial direct numerical simulation (PSDNS) approach on the Intel iPSC/860 hypercube and IBM SP1 and SP2 parallel computers is documented. Spatially evolving disturbances associated with the laminar-to-turbulent transition in boundary-layer flows are computed with the PSDNS code. The feasibility of using the PSDNS to perform transition studies on these computers is examined. The results indicate that PSDNS approach can effectively be parallelized on a distributed-memory parallel machine by remapping the distributed data structure during the course of the calculation. Scalability information is provided to estimate computational costs to match the actual costs relative to changes in the number of grid points. By increasing the number of processors, slower than linear speedups are achieved with optimized (machine-dependent library) routines. This slower than linear speedup results because the computational cost is dominated by FFT routine, which yields less than ideal speedups. By using appropriate compile options and optimized library routines on the SP1, the serial code achieves 52-56 M ops on a single node of the SP1 (45 percent of theoretical peak performance). The actual performance of the PSDNS code on the SP1 is evaluated with a "real world" simulation that consists of 1.7 million grid points. One time step of this simulation is calculated on eight nodes of the SP1 in the same time as required by a Cray Y/MP supercomputer. For the same simulation, 32-nodes of the SP1 and SP2 are required to reach the performance of a Cray C-90. A 32 node SP1 (SP2) configuration is 2.9 (4.6) times faster than a Cray Y/MP for this simulation, while the hypercube is roughly 2 times slower than the Y/MP for this application. KEY WORDS: Spatial direct numerical simulations; incompressible viscous flows; spectral methods; finite differences; parallel computing.

Joslin, Ronald D.↗

A Robust Parallel Distributed State Estimation for Large Scale Distribution Systems

The growing need and interest in real-time monitoring of large distribution networks motivated by the rapid population of renewable sources, EVs and etc. demand a computationally efficient state estimation framework. Furthermore, this paper presents an improved computational framework for implementing a robust state estimator using a multi-core processor. The main contribution of the paper is the proposed computational framework along with two partitioning strategies which enable fast and robust state estimation for large scale radial and/or meshed distribution systems. Formulation of the proposed method and its implementation are described in detail. Performance of the estimator is tested by simulations first using a small 84-bus radial distribution system. Then the method’s scalability is demonstrated by simulations on two very large scale distribution networks one configured radially and the other meshed each containing over 12,500 buses.

42 ENGINEERING↗

TriC: Distributed-memory Triangle Counting by Exploiting the Graph Structure

Graph analytics has emerged as an important tool in the analysis of large scale data from diverse application domains such as social networks, cyber security and bioinformatics. Counting the number of triangles in a graph is a fundamental kernel with several applications such as detecting the community structure of a graph or in identifying important vertices in a graph. The ubiquity of massive datasets is driving the need to scale graph analytics on parallel systems. However, numerous challenges exist in efficiently parallelizing graph algorithms, especially on distributed-memory systems. Irregular memory accesses and communication patterns, low computation to communication ratios, and the need for frequent synchronization are some of the leading challenges. In this paper, we present TriC, our distributed-memory implementation of triangle counting in graphs using the Message Passing Interface (MPI), as a submission to the 2020 GraphChallenge competition. Using a set of synthetic and real-world inputs from the challenge, we demonstrate a speedup of up to 90x relative to previous work on 32 processor-cores of a NERSC Cori node. We also provide details from distributed runs with up to8192 processes along with strong scaling results. The observations presented in this work provide an understanding of the system-level bottlenecks at scale that specifically impact sparse-irregular workloads and will therefore benefit other efforts to parallelize graph algorithms.

Halappanavar, Mahantesh↗