Search NASA⌕ Search

SEARCH · Search NASA

Results for “compute intensive problems”

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 451 records · Page 25
All sourcesNASA NTRSDOE OSTI
DOE OSTI2018

Coupled Electron-Photon Monte Carlo Radiation Transport for Next-Generation Computing Systems

Traditional Monte Carlo particle transport codes are expected to run inefficiently on next-generation architectures as they are memory-intensive and highly divergent. Since electrons and photons also behave differently, the future for coupled electron-photon radiation transport looks even worse. This project describes preliminary efforts to improve the performance of Monte Carlo particle transport codes when using accelerators like the graphics processing unit (GPU). Two key issues are addressed: how to handle memory-intensive tallies, and how to reduce divergence. Tallying on the GPU can be done efficiently by post-processing particle data, or by using a feature called warp shuffle for summing scores in parallel during the simulation. Reducing divergence is possible by using an event-based algorithm for particle tracking instead of the traditional history-based one. Although performance tests presented in this work show that the history-based algorithm generally outperformed the event-based one for simple problems, this outcome will likely change as the complexity of the code increases.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2014

Synthetic Graph Generation for Data-Intensive HPC Benchmarking: Scalability, Analysis and Real-World Application

The benchmarking effort within the Extreme Scale Systems Center at Oak Ridge National Laboratory seeks to provide High Performance Computing benchmarks and test suites of interest to the DoD sponsor. The work described in this report is a part of the effort focusing on graph generation. A previously developed benchmark, SystemBurn, allows the emulation of a broad spectrum of application behavior profiles within a single framework. To complement this effort, similar capabilities are desired for graph-centric problems. This report described the in-depth analysis of the generated synthetic graphs' properties at a variety of scales using different generator implementations and examines their applicability to replicating real world datasets.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Desmearing Bonse–Hart USANS data using Bayesian Gaussian process regression

Ultra-small-angle neutron scattering (USANS) enables access to micrometer-scale structures but is intrinsically affected by strong, anisotropic resolution smearing arising from slit-geometry optics. As a result, recovery of the intrinsic scattering intensity constitutes an ill-posed inverse problem, and commonly used iterative desmearing methods lack rigorous uncertainty quantification. We present a Bayesian desmearing framework for slit-geometry USANS based on Gaussian process regression. In this approach, the scattering intensity is modeled as a smooth random function, and the instrumental point spread function is incorporated explicitly as a forward operator. The resulting formulation yields a closed-form maximum a posteriori solution with well-defined credibility intervals. Computational benchmarks and experimental validation using combined USANS and small-angle neutron scattering (SANS) measurements demonstrate that the framework enables stable desmearing, suppresses experimental noise, and preserves physically meaningful structural features under realistic conditions.

Tung, Chi-Huan [Oak Ridge National Laboratory (ORN↗
NASA NTRS2012

Assessment of Polarization Effect on Efficiency of Levenberg-Marquardt Algorithm in Case of Thin Atmosphere Over Black Surface

The Levenberg-Marquardt algorithm [1, 2] provides a numerical iterative solution to the problem of minimization of a function over a space of its parameters. In our work, the Levenberg-Marquardt algorithm retrieves optical parameters of a thin (single scattering) plane parallel atmosphere irradiated by collimated infinitely wide monochromatic beam of light. Black ground surface is assumed. Computational accuracy, sensitivity to the initial guess and the presence of noise in the signal, and other properties of the algorithm are investigated in scalar (using intensity only) and vector (including polarization) modes. We consider an atmosphere that contains a mixture of coarse and fine fractions. Following [3], the fractions are simulated using Henyey-Greenstein model. Though not realistic, this assumption is very convenient for tests [4, p.354]. In our case it yields analytical evaluation of Jacobian matrix. Assuming the MISR geometry of observation [5] as an example, the average scattering cosines and the ratio of coarse and fine fractions, the atmosphere optical depth, and the single scattering albedo, are the five parameters to be determined numerically. In our implementation of the algorithm, the system of five linear equations is solved using the fast Cramer s rule [6]. A simple subroutine developed by the authors, makes the algorithm independent from external libraries. All Fortran 90/95 codes discussed in the presentation will be available immediately after the meeting from sergey.v.korkin@nasa.gov by request.

Korkin, S.↗
NASA NTRS1996

Computational Methods for Frictional Contact With Applications to the Space Shuttle Orbiter Nose-Gear Tire

A computational procedure is presented for the solution of frictional contact problems for aircraft tires. A Space Shuttle nose-gear tire is modeled using a two-dimensional laminated anisotropic shell theory which includes the effects of variations in material and geometric parameters, transverse-shear deformation, and geometric nonlinearities. Contact conditions are incorporated into the formulation by using a perturbed Lagrangian approach with the fundamental unknowns consisting of the stress resultants, the generalized displacements, and the Lagrange multipliers associated with both contact and friction conditions. The contact-friction algorithm is based on a modified Coulomb friction law. A modified two-field, mixed-variational principle is used to obtain elemental arrays. This modification consists of augmenting the functional of that principle by two terms: the Lagrange multiplier vector associated with normal and tangential node contact-load intensities and a regularization term that is quadratic in the Lagrange multiplier vector. These capabilities and computational features are incorporated into an in-house computer code. Experimental measurements were taken to define the response of the Space Shuttle nose-gear tire to inflation-pressure loads and to inflation-pressure loads combined with normal static loads against a rigid flat plate. These experimental results describe the meridional growth of the tire cross section caused by inflation loading, the static load-deflection characteristics of the tire, the geometry of the tire footprint under static loading conditions, and the normal and tangential load-intensity distributions in the tire footprint for the various static vertical loading conditions. Numerical results were obtained for the Space Shuttle nose-gear tire subjected to inflation pressure loads and combined inflation pressure and contact loads against a rigid flat plate. The experimental measurements and the numerical results are compared.

Tanner, John A.↗
NASA NTRS1996

Computational methods for frictional contact with applications to the Space Shuttle orbiter nose-gear tire: Comparisons of experimental measurements and analytical predictions

A computational procedure is presented for the solution of frictional contact problems for aircraft tires. A Space Shuttle nose-gear tire is modeled using a two-dimensional laminated anisotropic shell theory which includes the effects of variations in material and geometric parameters, transverse-shear deformation, and geometric nonlinearities. Contact conditions are incorporated into the formulation by using a perturbed Lagrangian approach with the fundamental unknowns consisting of the stress resultants, the generalized displacements, and the Lagrange multipliers associated with both contact and friction conditions. The contact-friction algorithm is based on a modified Coulomb friction law. A modified two-field, mixed-variational principle is used to obtain elemental arrays. This modification consists of augmenting the functional of that principle by two terms: the Lagrange multiplier vector associated with normal and tangential node contact-load intensities and a regularization term that is quadratic in the Lagrange multiplier vector. These capabilities and computational features are incorporated into an in-house computer code. Experimental measurements were taken to define the response of the Space Shuttle nose-gear tire to inflation-pressure loads and to inflation-pressure loads combined with normal static loads against a rigid flat plate. These experimental results describe the meridional growth of the tire cross section caused by inflation loading, the static load-deflection characteristics of the tire, the geometry of the tire footprint under static loading conditions, and the normal and tangential load-intensity distributions in the tire footprint for the various static vertical-loading conditions. Numerical results were obtained for the Space Shuttle nose-gear tire subjected to inflation pressure loads and combined inflation pressure and contact loads against a rigid flat plate. The experimental measurements and the numerical results are compared.

Tanner, John A.↗
DOE OSTI2026

Joint Optimization of Multimodal Transit Frequency and Shared Autonomous Vehicle Fleet Size with Hybrid Metaheuristic and Nonlinear Programming

Shared autonomous vehicles (SAVs) bring competition to traditional transit services but redesigning multimodal transit network can utilize SAVs as feeders to enhance service efficiency and coverage. This paper presents an optimization framework for the joint multimodal transit frequency and SAV fleet size problem, a variant of the transit network frequency setting problem. The objective is to maximize total transit ridership (including SAV-fed trips and subtracting boarding rejections) across multiple time periods under budget constraints, considering endogenous mode choice (transit, point-to-point SAVs, driving) and route selection, while allowing for strategic route removal by setting frequencies to zero. Due to the problem’s non-linear, non-convex nature and the computational challenges of large-scale networks, we develop a hybrid solution approach that combines a metaheuristic approach (particle swarm optimization) with nonlinear programming for local solution refinement. To ensure computational tractability, the framework integrates analytical approximation models for SAV waiting times based on fleet utilization, multimodal network assignment for route choice, and multinomial logit mode choice behavior, bypassing the need for computationally intensive simulations within the main optimization loop. Applied to the Chicago metropolitan area’s multimodal network, our method illustrates a 33.3% increase in transit ridership through optimized transit route frequencies and SAV integration, particularly enhancing off-peak service accessibility and strategically reallocating resources.

Ng, Max↗
DOE OSTI2025

A Performance Model of In-Situ Techniques

The computational capacity of High-Performance Computing (HPC) systems increases continuously with the rapid development of central processing units (CPUs) and graphic processing units (GPUs), while the in-/output (IO) subsystem develops relatively slowly and storage capacity is also limited. Data-intensive applications, which are designed to leverage the high computational capacity of HPC resources, typically generate a considerable amount of data for post-processing visualizations and data analytics. The limited IO speed and storage space could lead to constraints in the actual performance of these applications and, therefore, scientific discovery. In-situ techniques, where data is visualized/analysed while still in memory rather than through disk, can contribute to alleviating these problems as they can reduce or even fully avoid data writing/reading through the IO subsystem to/from storage. However, the overall efficiency of insitu techniques crucially depends on the characteristics of both the in-situ tasks and the applications, and the resource distribution among them. Therefore, choosing the right in-situ approach (synchronous, asynchronous, or hybrid) and resource allocation is essential to minimize overhead and maximize the benefits of concurrent execution. In this paper, we present a performance model of in-situ techniques to find the most beneficial in-situ approach and the preferred resource configuration. We verify the high accuracy of our approach with over 6800 measurements and provide use cases with different applications.

Ju, Yi [Max Planck Computing and Data Facility, Ga↗
DOE OSTI2018

Epi Archive: Automated Synthesis of Global Notifiable Disease Data

ObjectiveLANL has built software that automatically collects global notifiable disease data, synthesizes the data, and makes it available to humans and computers within the Biosurveillance Ecosystem (BSVE) as a novel data stream. These data have many applications including improving the prediction and early warning of disease events.IntroductionMost countries do not report national notifiable disease data in a machine-readable format. Data are often in the form of a file that contains text, tables and graphs summarizing weekly or monthly disease counts. This presents a problem when information is needed for more data intensive approaches to epidemiology, biosurveillance and public health.While most nations likely store incident data in a machine-readable format, governments are often hesitant to share data openly for a variety of reasons that include technical, political, economic, and motivational issues1.A survey conducted by LANL of notifiable disease data reporting in over fifty countries identified only a few websites that report data in a machine-readable format. The majority (>70%) produce reports as PDF files on a regular basis. The bulk of the PDF reports present data in a structured tabular format, while some report in natural language.The structure and format of PDF reports change often; this adds to the complexity of identifying and parsing the desired data. Not all websites publish in English, and it is common to find typos and clerical errors.LANL has developed a tool, Epi Archive, to collect global notifiable disease data automatically and continuously and make it uniform and readily accessible.MethodsWe conducted a survey of the national notifiable disease reporting systems notating how the data are reported and in what formats. We determined the minimal metadata that is required to contextualize incident counts properly, as well as optional metadata that is commonly found.The development of software to regularly ingest notifiable disease data and make it available involves three or four main steps: scraping, detecting, parsing and persisting.Scraping: we examine website design and determine reporting mechanisms for each country/website, as well as what varies across the reporting mechanisms. We then designed and wrote code to automate the downloading of the data for each country. We store all artifacts presented as files (PDF, XLSX, etc.) in their original form, along with appropriate metadata for parsing and data provenance.Detecting: This step is required when parsing structured non-machine-readable data such as tabular data in PDF files. We combined the Nurminen methodology of PDF table detection with in-house heuristics to find the desired data within PDF reports2.Parsing: We determined what to extract from each dataset and parsed these data into uniform data structures, correctly accommodating the variations in metadata (e.g., time interval definitions) and the various human languages.Persisting: We store the data in the Epi Archive database and make it available on the internet and through the BSVE. The data is persisted into a structured and normalized SQL database.ResultsThe Epi Archive tool currently contains national and/or subnational notifiable disease data from twenty nations. When a user accesses the Epi Archive site, they are prompted with four fields: country, subregion, disease of interest, and date duration. Upon form submission, a time series is generated from the users’ specifications. The generated graph can then be downloaded into a CSV file if a user is interested in performing personal analysis. Additionally, the data from Epi Archive can be reached through a REST API (Representational State Transfer Application Programming Interface).ConclusionsLANL, as part of a currently funded DTRA effort, is automatically and continually collecting global notifiable disease data. While 20 nations are in production, more are being brought online in the near future. These data are already being utilized and will have many applications including improving the prediction and early warning of disease events.References[1] van Panhuis WG, Paul P, Emerson C, et al. A systematic review of barriers to data sharing in public health. BMC Public Health. 2014. 14:1144. doi:10.1186/1471-2458-14-1144[2] Nurminen, Anssi. "Algorithmic extraction of data in tables in PDF documents." (2013).

Khalsa, Hari S.↗
NASA NTRS1994

Automated mainframe data collection in a network environment

The progress and direction of the computer industry have resulted in widespread use of dissimilar and incompatible mainframe data systems. Data collection from these multiple systems is a labor intensive task. In the past, data collection had been restricted to the efforts of personnel specially trained on each system. Information is one of the most important resources an organizations has. Any improvement in an organization's ability to access and manage that information provides a competitive advantage. This problem of data collection is compounded at NASA sites by multi-center and contractor operations. The Centralized Automated Data Retrieval System (CADRS) is designed to provide a common interface that would permit data access, query, and retrieval from multiple contractor and NASA systems. The methods developed for CADRS have a strong commercial potential in that they would be applicable for any industry that needs inter-department, inter-company, or inter-agency data communications. The widespread use of multi-system data networks, that combine older legacy systems with newer decentralized networks, has made data retrieval a critical problem for information dependent industries. Implementing the technology discussed in this paper would reduce operational expense and improve data collection on these composite data systems.

Gross, David L.↗
NASA NTRS2004

Winds, Water Budgets and Stable Isotopes in Tropical Cyclones using TRMM and QUICKSCAT

Water vapor is the most abundant greenhouse gas in the atmosphere. Changes in its concentration and distribution are controlled by the hydrologic cycle. Because of its capacity to absorb and emit long wave radiation, release latent heat during condensation in storms and reflect short wave radiation when clouds form it has a major impact on Global climate change. The stable isotope ratios of water are H20 H2l6O and H0 H2l6O. These ratios change whenever water undergoes a phase change. They also change in both rain and water vapor whenever an air parcel is exposed to rain. In addition the relative changes in the two ratios differ as a &nction of the relative humidity. In short, the stable isotope ratios in water vapor in the atmosphere contain an integrated history of the processes affecting the concentration and distribution of water vapor in the atmosphere. Therefore the measurement and interpretation of changes in these stable isotope ratios are a powerful tool matched by no other method in tracing the transport history of water in the atmosphere. Our initial studies under this grant focused on the changes of the stable isotope ratios of precipitation and water vapor in tropical cyclones. The changes in time and space were found to be very large and to trace the transport of water in the storms reflecting changes in basic structural features. Because the stable isotope ratios of rains from tropical cyclones are so low flooding associated with land falling tropical cyclones introduces a negative isotopic spike into the coastal surface waters. In addition the stable isotope ratios of water vapor in the vicinity of tropical cyclones is anomalously low. This suggests that carbonate shelled organisms such as ostracoda living in coastal waters have the potential to record the isotopic spike and thereby provide a long term record of tropical storm activity in sediment cores containing fossil shells. Likewise, tree rings in coastal environments offer a similar potential. We have analyzed the oxygen isotopic composition of ostrcoda shells formed in the floodwaters of Tropical Storm Allison (2001) and discovered the negative isotopic 1 16 spike. Because we had learned that storm activity has a major impact on the stable isotope ratios of water vapor in the tropics and sub-tropics we decided to analyze the isotopic compositions of water vapor in different locations in the tropics. We did this in Puerto Escondido, Mexico in July 1998, near Kwajalein Island in the Pacific in 1999 as part of a TRMM summer field program and in 2001 in Key West, Florida as part of the CAMEX 4 summer field program. Our isotopic studies along with our earlier tropical cyclone studies showed that the low isotopic ratios in water vapor induced by exposure to rains the storms persisted for 48 hours often far away from the original storm site. We also noted that positive isotopic spikes were introduced into atmospheric water vapor if winds were high and extensive sea spray was present. These findings have a significant impact on the interpretation of the stable isotope studies of tropical ice cores found in the high mountain regions of the tropics. The assumption made in interpreting the ice core record is that the source water vapor evaporated from the sea surface is in near isotopic equilibrium with the seawater and undergoes a decrease during its transport that reflects the change in temperature from the sea surface to the site of the ice core. Because an additional isotopic depletion occurs at the sea surface source area that depends on the intensity, duration and size of the tropical rain system the isotopic variations found in the ice cores must take into account changes in past storm activity in the tropics. These systems must be an important source of water vapor to the ice cores because they charge the troposphere with water vapor to a far greater vertical height than evaporation in quiescent regions. Finally, an interest in increased heat transfer in thnterior of tropical cyclones resulting from greater amounts of sea spray is a topic of considerable interest to the research community. Increases in sea spray may be related to rapid changes in the intensity of hurricanes, a feature of hurricanes that currently is very poorly forecast. Project CBLAST of the Hurricane Research Division of NOAA is an active program that uses P3 research aircraft to evaluate this problem. An instrument has been designed and built at the University of Houston that will be placed on the P3 research aircraft during the 2004 hurricane season. It continuously measures the salt content of rain in hurricanes. Changes in the salt content of the rains should reflect the abundance of sea spray at the sea surface. In this way maps of sea spray intensity in hurricanes may be forthcoming. This should help computer modelers who simulate hurricanes to better understand the potential of changes in sea spray to change the intensity of hurricanes. The cost of designing, building and installing this instrument was borne largely by funding from this NASA grant. A list of presentations at national meetings and publications that were as the result of funding from this NASA grant are found in the report.

Lawrence, James R.↗
NASA NTRS2011

Experiment in Onboard Synthetic Aperture Radar Data Processing

Single event upsets (SEUs) are a threat to any computing system running on hardware that has not been physically radiation hardened. In addition to mandating the use of performance-limited, hardened heritage equipment, prior techniques for dealing with the SEU problem often involved hardware-based error detection and correction (EDAC). With limited computing resources, software- based EDAC, or any more elaborate recovery methods, were often not feasible. Synthetic aperture radars (SARs), when operated in the space environment, are interesting due to their relevance to NASAs objectives, but problematic in the sense of producing prodigious amounts of raw data. Prior implementations of the SAR data processing algorithm have been too slow, too computationally intensive, and require too much application memory for onboard execution to be a realistic option when using the type of heritage processing technology described above. This standard C-language implementation of SAR data processing is distributed over many cores of a Tilera Multicore Processor, and employs novel Radiation Hardening by Software (RHBS) techniques designed to protect the component processes (one per core) and their shared application memory from the sort of SEUs expected in the space environment. The source code includes calls to Tilera APIs, and a specialized Tilera compiler is required to produce a Tilera executable. The compiled application reads input data describing the position and orientation of a radar platform, as well as its radar-burst data, over time and writes out processed data in a form that is useful for analysis of the radar observations.

Holland, Matthew↗
NASA NTRS2003

Development of Numerical Codes for Modeling Electromagnetic Behavior at High Frequencies Near Large Objects

A study into the problem of determining electromagnetic solutions at high frequencies for problems involving complex geometries, large sizes and multiple sources (e.g. antennas) has been initiated. Typical applications include the behavior of antennas (and radiators) installed on complex conducting structures (e.g. ships, aircrafts, etc..) with strong interactions between antennas, the radiation patterns, and electromagnetic signals is of great interest for electromagnetic compatibility control. This includes the overall performance evaluation and control of all on-board radiating systems, electromagnetic interference, and personnel radiation hazards. Electromagnetic computational capability exists at NASA LaRC, and many of the codes developed are based on the Moment Method (MM). However, the MM is computationally intensive, and this places a limit on the size of objects and structures that can be modeled. Here, two approaches are proposed: (i) a current-based hybrid scheme that combines the MM with Physical optics, and (ii) an Alternating Direction Implicit-Finite Difference Time Domain (ADI-FDTD) method. The essence of a hybrid technique is to split the overall scattering surface(s) into two regions: (a) a MM zone (MMZ) which can be used over any part of the given geometry, but is most essential over irregular and "non-smooth" geometries, and (b) a PO sub-region (POSR). Currents induced on the scattering and reflecting surfaces can then be computed in two ways depending on whether the region belonged to the MMZ or was part of the POSR. For the MMZ, the current calculations proceed in terms of basis functions with undetermined coefficients (as in the usual MM method), and the answer obtained by solving a system of linear equations. Over the POSR, conduction is obtained as a superposition of two contributions: (i) currents due to the incident magnetic field, and (ii) currents produced by the mutual induction from conduction within the MMZ. This effectively leads to a reduction in the size of linear equations from N to N - Npo with N being the total number of segments for the entire surface and Npo the number of segments over the POSR. The scheme would be appropriate for relatively large, flat surfaces, and at high frequencies. The ADI-FDTD scheme provides for both transient and steady state analyses. The restrictive Courant-Friedrich-Levy (CFL) condition on the time-step is removed, and so large time steps can be chosen even though the spatial grids are small. This report includes the problem definition, a detailed discussion of both the numerical techniques, and numerical implementations for simple surface geometries. Numerical solutions have been derived for a few simple situations.

Joshi, R. P.↗
DOE OSTI2016

SU-G-JeP2-02: A Unifying Multi-Atlas Approach to Electron Density Mapping Using Multi-Parametric MRI for Radiation Treatment Planning

Purpose: MRI has a number of advantages over CT as a primary modality for radiation treatment planning (RTP). However, one key bottleneck problem still remains, which is the lack of electron density information in MRI. In the work, a reliable method to map electron density is developed by leveraging the differential contrast of multi-parametric MRI. Methods: We propose a probabilistic Bayesian approach for electron density mapping based on T1 and T2-weighted MRI, using multiple patients as atlases. For each voxel, we compute two conditional probabilities: (1) electron density given its image intensity on T1 and T2-weighted MR images, and (2) electron density given its geometric location in a reference anatomy. The two sources of information (image intensity and spatial location) are combined into a unifying posterior probability density function using the Bayesian formalism. The mean value of the posterior probability density function provides the estimated electron density. Results: We evaluated the method on 10 head and neck patients and performed leave-one-out cross validation (9 patients as atlases and remaining 1 as test). The proposed method significantly reduced the errors in electron density estimation, with a mean absolute HU error of 138, compared with 193 for the T1-weighted intensity approach and 261 without density correction. For bone detection (HU>200), the proposed method had an accuracy of 84% and a sensitivity of 73% at specificity of 90% (AUC = 87%). In comparison, the AUC for bone detection is 73% and 50% using the intensity approach and without density correction, respectively. Conclusion: The proposed unifying method provides accurate electron density estimation and bone detection based on multi-parametric MRI of the head with highly heterogeneous anatomy. This could allow for accurate dose calculation and reference image generation for patient setup in MRI-based radiation treatment planning.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

GPU-enabled extreme-scale turbulence simulations: Fourier pseudo-spectral algorithms at the exascale using OpenMP offloading

Fourier pseudo-spectral methods for nonlinear partial differential equations are of wide interest in many areas of advanced computational science, including direct numerical simulation of three-dimensional (3-D) turbulence governed by the Navier-Stokes equations in fluid dynamics. This paper presents a new capability for simulating turbulence at a new record resolution up to 35 trillion grid points, on the world's first exascale computer, Frontier, comprising AMD MI250x GPUs with HPE's Slingshot interconnect and operated by the US Department of Energy's Oak Ridge Leadership Computing Facility (OLCF). Key programming strategies designed to take maximum advantage of the machine architecture involve performing almost all computations on the GPU which has the same memory capacity as the CPU, performing all-to-all communication among sets of parallel processes directly on the GPU, and targeting GPUs efficiently using OpenMP offloading for intensive number-crunching including 1-D Fast Fourier Transforms (FFT) performed using AMD ROCm library calls. With 99% of computing power on Frontier being on the GPU, leaving the CPU idle leads to a net performance gain via avoiding the overhead of data movement between host and device except when needed for some I/O purposes. Memory footprint including the size of communication buffers for MPI_ALLTOALL is managed carefully to maximize the largest problem size possible for a given node count. Detailed performance data including separate contributions from different categories of operations to the elapsed wall time per step are reported for five grid resolutions, from 2048 3 on a single node to 32768 3 on 4096 or 8192 nodes out of 9408 on the system. Both 1D and 2D domain decompositions which divide a 3D periodic domain into slabs and pencils respectively are implemented. The present code suite (labeled by the acronym GESTS, GPUs for Extreme Scale Turbulence Simulations) achieves a figure of merit (in grid points per second) exceeding goals set in the Center for Accelerated Application Readiness (CAAR) program for Frontier. The performance attained is highly favorable in both weak scaling and strong scaling, with notable departures only for 2048 3 where communication is entirely intra-node, and for 32768 3 , where a challenge due to small message sizes does arise. Communication performance is addressed further using a lightweight test code that performs all-to-all communication in a manner matching the full turbulence simulation code. Performance at large problem sizes is affected by both small message size due to high node counts as well as dragonfly network topology features on the machine, but is consistent with official expectations of sustained performance on Frontier. Overall, although not perfect, the scalability achieved at the extreme problem size of 32768 3 (and up to 8192 nodes — which corresponds to hardware rated at just under 1 exaflop/sec of theoretical peak computational performance) is arguably better than the scalability observed using prior state-of-the-art algorithms on Frontier's predecessor machine (Summit) at OLCF. New science results for the study of intermittency in turbulence enabled by this code and its extensions are to be reported separately in the near future.

3D fast Fourier transform↗
DOE OSTI2020

CSRI Summer Proceedings 2020

The Computer Science Research Institute (CSRI) brings university faculty and students to Sandia for focused collaborative research on Department of Energy (DOE) computer and computational science problems. The institute provides an opportunity for university researchers to learn about problems in computer and computational science at DOE laboratories. Participants conduct leading-edge research, interact with scientists and engineers at the laboratories, and help transfer results of their research to programs at the labs. Some specific CSRI research interest areas are: scalable solvers, optimization, adaptivity and mesh refinement, graph-based, discrete, and combinatorial algorithms, uncertainty estimation, mesh generation, dynamic load-balancing, virus and other malicious-code defense, visualization, scalable cluster computers, data-intensive computing, environments for scalable computing, parallel input/output, advanced architectures, and theoretical computer science. The CSRI Summer Program is organized by CSRI and typically includes the organization of a weekly seminar series and the publication of a summer proceedings. In 2020, the CSRI summer program was executed completely virtually; all student interns worked from home, due to the COVID-19 pandemic.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1995

Computing Interactions Of Free-Space Radiation With Matter

High Charge and Energy Transport (HZETRN) computer program computationally efficient, user-friendly package of software adressing problem of transport of, and shielding against, radiation in free space. Designed as "black box" for design engineers not concerned with physics of underlying atomic and nuclear radiation processes in free-space environment, but rather primarily interested in obtaining fast and accurate dosimetric information for design and construction of modules and devices for use in free space. Computational efficiency achieved by unique algorithm based on deterministic approach to solution of Boltzmann equation rather than computationally intensive statistical Monte Carlo method. Written in FORTRAN.

Wilson, J. W.↗
NASA NTRS2000

HTMT-class Latency Tolerant Parallel Architecture for Petaflops Scale Computation

Computational Aero Sciences and other numeric intensive computation disciplines demand computing throughputs substantially greater than the Teraflops scale systems only now becoming available. The related fields of fluids, structures, thermal, combustion, and dynamic controls are among the interdisciplinary areas that in combination with sufficient resolution and advanced adaptive techniques may force performance requirements towards Petaflops. This will be especially true for compute intensive models such as Navier-Stokes are or when such system models are only part of a larger design optimization computation involving many design points. Yet recent experience with conventional MPP configurations comprising commodity processing and memory components has shown that larger scale frequently results in higher programming difficulty and lower system efficiency. While important advances in system software and algorithms techniques have had some impact on efficiency and programmability for certain classes of problems, in general it is unlikely that software alone will resolve the challenges to higher scalability. As in the past, future generations of high-end computers may require a combination of hardware architecture and system software advances to enable efficient operation at a Petaflops level. The NASA led HTMT project has engaged the talents of a broad interdisciplinary team to develop a new strategy in high-end system architecture to deliver petaflops scale computing in the 2004/5 timeframe. The Hybrid-Technology, MultiThreaded parallel computer architecture incorporates several advanced technologies in combination with an innovative dynamic adaptive scheduling mechanism to provide unprecedented performance and efficiency within practical constraints of cost, complexity, and power consumption. The emerging superconductor Rapid Single Flux Quantum electronics can operate at 100 GHz (the record is 770 GHz) and one percent of the power required by convention semiconductor logic. Wave Division Multiplexing optical communications can approach a peak per fiber bandwidth of 1 Tbps and the new Data Vortex network topology employing this technology can connect tens of thousands of ports providing a bi-section bandwidth on the order of a Petabyte per second with latencies well below 100 nanoseconds, even under heavy loads. Processor-in-Memory (PIM) technology combines logic and memory on the same chip exposing the internal bandwidth of the memory row buffers at low latency. And holographic storage photorefractive storage technologies provide high-density memory with access a thousand times faster than conventional disk technologies. Together these technologies enable a new class of shared memory system architecture with a peak performance in the range of a Petaflops but size and power requirements comparable to today's largest Teraflops scale systems. To achieve high-sustained performance, HTMT combines an advanced multithreading processor architecture with a memory-driven coarse-grained latency management strategy called "percolation", yielding high efficiency while reducing the much of the parallel programming burden. This paper will present the basic system architecture characteristics made possible through this series of advanced technologies and then give a detailed description of the new percolation approach to runtime latency management.

Sterling, Thomas↗