Search NASA⌕ Search

SEARCH · Search NASA

Results for “Global Arrays”

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 73 records · Page 4

Distributed, Intelligent Edge-Sensing for a Smarter Grid

The electric grid is undergoing major transformations and developments resulting in unprecedented levels of volatility, uncertainty, and stress on grid infrastructure. Smart sensors and methods aiding in advanced visibility and situational awareness are key for tackling these issues. In this work, a decentralized architecture is proposed, where sensing, local computation and control capability are embedded in the edge devices, communicating with a set of trusted 'data mules' in a 'delay-tolerant' manner, while functioning autonomously. This system has been designed and implemented as an overall platform – called Global Asset Monitoring, Management and Analytics (GAMMA) Platform intended to provide the backbone for a global array of sensors and actuators. Further, as a building block for advanced current sensing solutions, a smart, low-cost ‘clip-on’ current sensor based on PCB-embedded Rogowski coil has been developed. The sensor hosts a novel signal conditioning stage allowing an 'auto-tuning' feature, resulting in a universal current sensor design for measuring a wide range of currents, including faults for smart grid applications. Finally, the research proposes a method to instrument and monitor key parameters for the most common electric utility asset – the pole-top distribution transformer. The work done in this research enables scalable, edge-intelligent sensing solutions for monitoring grid infrastructure, allowing utility operators to gain advanced visibility in an economical way.

Kulkarni, Shreyas Bhalchandra↗

Computational Chemistry at the Petascale: Are We There Yet?

The field of electronic structure is struggling to get efficient parallel implementation on Petascaleclass hardware. One notable exception has been the achievement of Qbox, a planewavepseudopotential electronic structure code that obtained a performance of 207 TFlops on aBlueGene/L computer.Qbox makes use of the message-passing MPI library for parallelization. Instead, NWChemmakes use of the Global Arrays library; this allows the software developer to reach a high levelof abstraction and, at the same time, to use one-sided communication to efficiently exploitthe network hardware. In the remainder of the paper, we will discuss recent benchmarksand scientic results obtained with NWChem on a parallel computer whose theoretical peakperformance is in excess of 1 PFlops.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Enabling a Highly-Scalable Global Address Space Model for Petascale Computing

Over the past decade, the trajectory to the petascale has been built on increased complexity and scale of the underlying parallel architectures. Meanwhile, software de- velopers have struggled to provide tools that maintain the productivity of computational science teams using these new systems. In this regard, Global Address Space (GAS) programming models provide a straightforward and easy to use addressing model, which can lead to improved produc- tivity. However, the scalability of GAS depends directly on the design and implementation of the runtime system on the target petascale distributed-memory architecture. In this paper, we describe the design, implementation, and optimization of the Aggregate Remote Memory Copy Interface (ARMCI) runtime library on the Cray XT5 2.3 PetaFLOPs computer at Oak Ridge National Laboratory. We optimized our implementation with the flow intimation technique that we have introduced in this paper. Our optimized ARMCI implementation improves scalability of both the Global Arrays (GA) programming model and a real-world chemistry application NWChem from small jobs up through 180,000 cores.

97 MATHEMATICS AND COMPUTING↗

A local-view array library for partitioned global address space C++programs

Multidimensional arrays are an important data structure in many scientific applications. Unfortunately, built-in support for such arrays is inadequate in C++, particularly in the distributed setting where bulk communication operations are required for good performance. In this paper, we present a multidimensional library for partitioned global address space (PGAS) programs, supporting the one-sided remote access and bulk operations of the PGAS model. The library is based on Titanium arrays, which have proven to provide good productivity and performance. These arrays provide a local view of data, where each rank constructs its own portion of a global data structure, matching the local view of execution common to PGAS programs and providing maximum flexibility in structuring global data. Unlike Titanium, which has its own compiler with array-specific analyses, optimizations, and code generation, we implement multidimensional arrays solely through a C++library. The main goal of this effort is to provide a library-based implementation that can match the productivity and performance of a compiler-based approach. We implement the array library as an extension to UPC++, a C++library for PGAS programs, and we extend Titanium arrays with specializations to improve performance. We evaluate the array library by porting four Titanium benchmarks to UPC++, demonstrating that it can achieve up to 25% better performance than Titanium without a significant increase in programmer effort..

Kamil, A↗

MONITORING THE BIDIRECTIONAL RELATIVISTIC JETS OF THE RADIO GALAXY 1946+708

We report on a multifrequency, multi-epoch campaign of Very Long Baseline Interferometry (VLBI) observations of the radio galaxy 1946+708 using the Very Long Baseline Array and a Global VLBI array. From these high-resolution observations, we deduce the kinematic age of the radio source to be {approx}4000 years, comparable with the ages of other Compact Symmetric Objects. Ejections of pairs of jet components appears to take place on time scales of ten years and these components in the jet travel outward at intrinsic velocities between 0.6c and 0.9c. From the constraint that jet components cannot have intrinsic velocities faster than light, we derive H {sub 0} > 57 km s{sup -1} Mpc{sup -1} from the fastest pair of components launched from the core. We provide strong evidence for the ejection of a new pair of components in {approx}1997. From the trajectories of the jet components, we deduce that the jet is most likely to be helically confined, rather than being purely ballistic in nature.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Finding maximum on an array processor with a global bus

The problem of finding the maximum of a set of values stored one/processor on an n x n array of processors is analyzed. The array has a time-shared global bus in addition to conventional processor-processor links. A two-phase algorithm for finding the maximum is presented that uses conventional links during the first phase and the global bus during the second. This algorithm is faster than algorithms that use either only the global bus or only the conventional links. Two types of interconnection patterns (the eighth nearest neighbor and the fourth nearest neighbor) are analyzed. In both cases it is shown that the time required to find the maximum using the two-phase algorithm is 0(n to the 2/3-power) assuming the propagation speed of the global bus to be a constant independent of the size of the array. In the case where the propagation speed is logarithmic in the number of processors, the time to find the maximum is 0(/n-squared log n/1/3), for both types of arrays. Extensions to q-dimensional arrays show that the two-phase algorithm is superior for any fixed value of q.

Bokhari, S. H.↗

Designing Scalable PGAS Communication Subsystems on Cray Gemini Interconnect

The Cray Gemini Interconnect has been recently introduced as a next generation network architecture for building multi-petaflop supercomputers. Cray XE6 systems including LANL Cielo, NERSC Hopper, ORNL Titan and proposed NCSA BlueWaters leverage the Gemini Interconnect as their primary Interconnection network. At the same time, programming models such as the Message Passing Interface (MPI) and Partitioned Global Address Space (PGAS) models such as Unified Parallel C (UPC) and Co-Array Fortran (CAF) have become available on these systems. Global Arrays is a popular PGAS model used in a variety of application domains including hydrodynamics, chemistry and visualization. Global Arrays uses Aggregate Re- mote Memory Copy Interface (ARMCI) as the communication runtime system for Remote Memory Access communication. This paper presents a design, implementation and performance evaluation of scalable and high performance communication subsystems on Cray Gemini Interconnect using ARMCI. The design space is explored and time-space complexities of commu- nication protocols for one-sided communication primitives such as contiguous and uniformly non-contiguous datatypes, atomic memory operations (AMOs) and memory synchronization is presented. An implementation of the proposed design (referred as ARMCI-Gemini) demonstrates the efficacy on communication primitives, application kernels such as LU decomposition and full applications such as Smooth Particle Hydrodynamics (SPH) application.

ARMCI↗

Spectral stability of Taylor's vortex array

In a global sense it is shown that the two-dimensional Taylor vortex array, an exact solution of the Navier-Stokes equation, is absolutely and monotonically stable with respect to infinitesimal disturbances of all discrete frequencies as long as the viscosity is positive. It is suggested that the Taylor vortex array may also be stable with respect to finite amplitude disturbances.

Lin, S. P.↗

Evaluation of global horizontal irradiance to plane-of-array irradiance models at locations across the United States

We report an evaluation of the accuracy of combinations of models that estimate plane-of-array (POA) irradiance from measured global horizontal irradiance (GHI). This estimation involves two steps: 1) decomposition of GHI into direct and diffuse horizontal components and 2) transposition of direct and diffuse horizontal irradiance (DHI) to POA irradiance. Measured GHI and coincident measured POA irradiance from a variety of climates within the United States were used to evaluate combinations of decomposition and transposition models. A few locations also had DHI measurements, allowing for decoupled analysis of either the decomposition or the transposition models alone. Results suggest that decomposition models had mean bias differences (modeled versus measured) that vary with climate. Transposition model mean bias differences depended more on the model than the location. Lastly, when only GHI measurements were available and combinations of decomposition and transposition models were considered, the smallest mean bias differences were typically found for combinations which included the Hay/Davies transposition model.

14 SOLAR ENERGY↗

Harmonia: A Globally Coordinated Garbage Collector for Arrays of Solid-state Drives

Solid-State Drives (SSDs) offer significant performance improvements over hard disk drives (HDD) on a number of workloads. The frequency of garbage collection (GC) activity is directly correlated with the pattern, frequency, and volume of write requests, and scheduling of GC is controlled by logic internal to the SSD. SSDs can exhibit significant performance degradations when garbage collection (GC) conflicts with an ongoing I/O request stream. When using SSDs in a RAID array, the lack of coordination of the local GC processes amplifies these performance degradations. No RAID controller or SSD available today has the technology to overcome this limitation. This paper presents Harmonia, a Global Garbage Collection (GGC) mechanism to improve response times and reduce performance variability for a RAID array of SSDs. Our proposal includes a high-level design of SSD-aware RAID controller and GGC-capable SSD devices, as well as algorithms to coordinate the global GC cycles. Our simulations show that this design improves response time and reduces performance variability for a wide variety of enterprise workloads. For bursty, write dominant workloads response time was improved by 69% while performance variability was reduced by 71%.

99 GENERAL AND MISCELLANEOUS↗

Extending ALE3D, an Arbitrarily Connected hexahedral 3D Code, to Very Large Problem Size (U)

As the number of compute units increases on the ASC computers, the prospect of running previously unimaginably large problems is becoming a reality. In an arbitrarily connected 3D finite element code, like ALE3D, one must provide a unique identification number for every node, element, face, and edge. This is required for a number of reasons, including defining the global connectivity array required for domain decomposition, identifying appropriate communication patterns after domain decomposition, and determining the appropriate load locations for implicit solvers, for example. In most codes, the unique identification number is defined as a 32-bit integer. Thus the maximum value available is 231, or roughly 2.1 billion. For a 3D geometry consisting of arbitrarily connected hexahedral elements, there are approximately 3 faces for every element, and 3 edges for every node. Since the nodes and faces need id numbers, using 32-bit integers puts a hard limit on the number of elements in a problem at roughly 700 million. The first solution to this problem would be to replace 32-bit signed integers with 32-bit unsigned integers. This would increase the maximum size of a problem by a factor of 2. This provides some head room, but almost certainly not one that will last long. Another solution would be to replace all 32-bit int declarations with 64-bit long long declarations. (long is either a 32-bit or a 64-bit integer, depending on the OS). The problem with this approach is that there are only a few arrays that actually need to extended size, and thus this would increase the size of the problem unnecessarily. In a future computing environment where CPUs are abundant but memory relatively scarce, this is probably the wrong approach. Based on these considerations, we have chosen to replace only the global identifiers with the appropriate 64-bit integer. The problem with this approach is finding all the places where data that is specified as a 32-bit integer needs to be replaced with the 64-bit integer. that need to be replaced. In the rest of this paper we describe the techniques used to facilitate this transformation, issues raised, and issues still to be addressed. This poster will describe the reasons, methods, issues associated with extending the ALE3D code to run problems larger than 700 million elements.

97 MATHEMATICS AND COMPUTING↗

Ablation dynamics in wire array Z-pinches under modifications on global magnetic field topology

The dynamics of ablation streams and precursor plasma in cylindrical wire array Z-pinches under temporal variations of the global magnetic field topology is investigated through experiments and numerical simulations. The wire arrays in these experiments are modified by replacing a pair of consecutive wires with wires of a larger diameter. This modification leads to two separate effects, both of which impact the dynamics of the precursor plasma; firstly, current is unevenly distributed between the wires and secondly, the thicker wires take longer to fully ablate. The uneven distribution of current is evidenced in the experiments by the drift of the precursor off axis due to a variation in the global magnetic field topology which modifies the direction of the ablation streams tracking the precursor position. The variation of the global magnetic field due to the presence of thick wires is studied with three-dimensional magnetohydrodynamic (MHD) simulations, showing that the global field changes from the expected toroidal field to a temporally variable topology after breakages appear in the thin wires. This leads to an observed acceleration of the precursor column towards the region closer to the thick wires and later, when thick wires also present breakages, it continues moving away from the original array position as a complicated and disperse object subject to MHD instabilities.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

On the use of term rewriting for performance optimization of legacy HPC applications

Preparing codes for next generation supercomputer systems is anticipated to require significant changes to the optimization strategies employed in established HPC applications. In this paper, we present our experience in applying term-rewriting transformations for the optimization of such applications. We have designed application-specific term-rewriting transformations to improve the scalability, enhance locality and reduce the communication overhead of the Self-Consistent Field (SCF) computational chemistry benchmark. We present the rationale for the use of term-rewriting in this manner, the design of our transformations, and the much enhanced performance of the resulting code.

Global Arrays↗

Scaling Linear Algebra Kernels using Remote Memory Access

This paper describes the scalability of linear algebra kernels based on remote memory access approach. The current approach differs from the other linear algebra algorithms by the explicit use of shared memory and remote memory access (RMA) communication rather than message passing. It is suitable for clusters and scalable shared memory systems. The experimental results on large scale systems (Linux-Infiniband cluster, Cray XT) demonstrate consistent performance advantages over ScaLAPACK suite, the leading implementation of parallel linear algebra algorithms used today. For example, on a Cray XT4 for a matrix size of 102400, our RMA-based matrix multiplication achieved over 55 teraflops while ScaLAPACK’s pdgemm measured close to 42 teraflops on 10000 processes.

99 GENERAL AND MISCELLANEOUS↗

First Sagittarius A* Event Horizon Telescope Results. I. The Shadow of the Supermassive Black Hole in the Center of the Milky Way

We present the first Event Horizon Telescope (EHT) observations of Sagittarius A* (Sgr A*), the Galactic center source associated with a supermassive black hole. These observations were conducted in 2017 using a global interferometric array of eight telescopes operating at a wavelength of λ = 1.3 mm. The EHT data resolve a compact emission region with intrahour variability. A variety of imaging and modeling analyses all support an image that is dominated by a bright, thick ring with a diameter of 51.8 ± 2.3 μas (68% credible interval). The ring has modest azimuthal brightness asymmetry and a comparatively dim interior. Using a large suite of numerical simulations, we demonstrate that the EHT images of Sgr A* are consistent with the expected appearance of a Kerr black hole with mass ∼4 × 10$^{6}$ M $_{⊙}$, which is inferred to exist at this location based on previous infrared observations of individual stellar orbits, as well as maser proper-motion studies. Our model comparisons disfavor scenarios where the black hole is viewed at high inclination (i > 50°), as well as nonspinning black holes and those with retrograde accretion disks. Our results provide direct evidence for the presence of a supermassive black hole at the center of the Milky Way, and for the first time we connect the predictions from dynamical measurements of stellar orbits on scales of 10$^{3}$–10$^{5}$ gravitational radii to event-horizon-scale images and variability. Furthermore, a comparison with the EHT results for the supermassive black hole M87* shows consistency with the predictions of general relativity spanning over three orders of magnitude in central mass.

79 ASTRONOMY AND ASTROPHYSICS↗

The Geoscience Laser Altimetry/Ranging System (GLARS)

The Geoscience Laser Altimetry Ranging System (GLARS) is a highly precise distance measurement system to be used for making extremely accurate geodetic observations from a space platform. It combines the attributes of a pointable laser ranging system making observations to cube corner retroreflectors placed on the ground with those of a nadir looking laser altimeter making height observations to ground, ice sheet, and oceanic surfaces. In the ranging mode, centimeter-level precise baseline and station coordinate determinations will be made on grids consisting of 100 to 200 targets separated by distances from a few tens of kilometers to about 1000 km. These measurements will be used for studies of seismic zone crustal deformations and tectonic plate motions. Ranging measurements will also be made to a coarser, but globally distributed array of retroreflectors for both precise geodetic and orbit determination applications. In the altimetric mode, relative height determinations will be obtained with approximately decimeter vertical precision and 70 to 100 meter horizontal resolution. The height data will be used to study surface topography and roughness, ice sheet and lava flow thickness, and ocean dynamics. Waveform digitization will provide a measure of the vertical extent of topography within each footprint. The planned Earth Observing System is an attractive candidate platform for GLARS since the GLAR data can be used both for direct analyses and for highly precise orbit determination needed in the reduction of data from other sensors on the multi-instrument platform. (1064, 532, and 355 nm)Nd:YAG laser meets the performance specifications for the system.

Cohen, S. C.↗