Search NASA⌕ Search

SEARCH · Search NASA

Results for “Software Development”

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 397 records · Page 22

Development of flexible, scalable, low cost readout for beam tests of the high granularity calorimeter for the CMS endcap

As part of the development of the High Granularity Calorimeter for the CMS Endcap at HL-LHC, The CMS collaboration is conducting a comprehensive series of beam tests. The first beam test, with a single HGC module was carried out in March of 2016 at the Fermilab Test Beam Facility, continuing to a 16 module test in July of 2016. We describe here the development of a low cost readout system that is simple to implement and is able to grow with the system under test. The system is based on the low cost Zynq SoC that allows simple DAQ development in a Linux environment. For this application we used the Digilent ZedBoard, which allows high speed LVDS links and Linux software development on a single commercial board. A small custom FPGA board designed to comply with the VITA 57 Field Programmable Mezzanine Card standard implements the interface to the readout ASIC mounted on the HGC sensor modules and provides the LVDS links to the ZedBoard, either directly over the FMC connector or via a custom carrier card. This architecture provides scalability, ease of development and low cost.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

The MCDB System for Management and Analysis of Petabyte-Scale Uncertain Data

Analysts working with very large data sets often use statistical models to “guess” at unknown, inaccurate, or missing information associated with the data. For example, a distant object viewed through an optical lens will have its position slightly shifted by imperfections in the lens. Thus, rather than considering the object’s observed position to be absolutely correct, it makes sense to take into account the lens’s imperfections to obtain a probabilistic guess as to the object’s true position. For another example, it might be important to associate some sort of error distribution with each of the individual sensors in an array of magnetometers. This error distribution may be complex and include spatially-driven covariances, because errors in nearby sensors are likely to be correlated (caused, for example, by the presence of some nearby, fixed metal object). This project is concerned with the design and implementation of a prototype data management system called the Monte Carlo Database System, or MCDB for short. MCDB allows an expert-level analyst or statistician to attach arbitrary stochastic models to very large data sets in order to “guess” the values for unknown or inaccurate data, such as the actual position of the observed object in the lens example above. When the resulting data set is analyzed, the underlying stochastic models are used to generate hundreds or thousands of possible data set instances, and each of those possible instances is analyzed separately by MCDB. Thus, MCDB does not just give a single answer to the analysis, but it actually gives an empirical distribution of query results that embody the underlying uncertainty, and can in turn be analyzed using standard statistical techniques. The stochastic models in MCDB are implemented as user-defined, external C++ libraries called Variable Generation functions (VG functions for short). Because the VG function interface is exceedingly general, it allows MCDB to be used in a very wide variety of application domains, in conjunction with virtually any stochastic model. Ultimately, we ended up calling the system that we developed over the course of the project SimSQL, to denote the fact that it allowed for stochastic simulation of data. The software developed during the lifetime of the project is available for download from the SimSQL project website: http://cmj4.web.rice.edu/SimSQL/SimSQL.html.

42 ENGINEERING↗

Effects of overlapping sources on cosmic shear estimation: Statistical sensitivity and pixel-noise bias

The next generation of dark-energy imaging surveys — so called “Stage-IV” surveys, such as that of the Rubin Observatory Legacy Survey of Space and Time (LSST) — will cross a threshold in the number density of detected sources on the sky that requires qualitatively different image analysis and measurement techniques compared to the current generation of Stage-III surveys. In Stage-IV surveys, a significant amount of the cosmologically useful information is due to sources whose images overlap with those of other sources on the sky. Here, we focus on the weak gravitational lensing probe, for which we expect the largest impact since the cosmic shear signal is primarily encoded in the estimated shapes of observed galaxies and thus directly impacted by overlaps. We introduce a framework based on the Fisher formalism to analyze the effect of the overlapping sources (“blending”) on the estimation of cosmic shear. This method gives concrete predictions for the minimum loss of information due to noise and blending for any choice of “deblending” scheme and shape-measurement algorithm. Our studies account for undetected sources but do not address their full effects and biases they may introduce. We use simulated images and predict this impact of blending for three surveys: the Dark Energy Survey (DES), the Hyper-Suprime Cam Subaru Strategic Program (HSC-SSP), and the Rubin LSST. Our methodology successfully estimates the statistical sensitivity to weak lensing for DES and HSC early results. For LSST, we present the expected loss in statistical sensitivity for the ten-year survey due to blending. We find that for approximately 62% of galaxies that are likely to be detected in full-depth LSST images, at least 1% of the flux in their pixels is from overlapping sources. We also find that the statistical correlations between measures of overlapping galaxies and, to a much lesser extent (0.2%) the higher shot noise level due to their presence, decrease the effective number density of galaxies, N eff , by ~ 18%. We calculate an upper limit on N eff of 39.4 galaxies per arcmin 2 in r band. We study the impact of stars on as a function of stellar density and illustrate the diminishing returns of extending the survey into lower Galactic latitudes. We extend the simulation-based Fisher formalism to predict the expected increase in pixel-noise bias due to blending for maximum-likelihood (ML) shape estimators. We find that noise bias depends sensitively on the particular shape estimator and measure of ensemble-average shape that is used, and properties of the galaxy that include redshift-dependent quantities such as size and luminosity. The source code for these studies is available online.[The documented software developed for the catalog-level studies are available in the open-source LSST DESC github repository https://github.com/LSSTDESC/WeakLensingDeblending. The software for analyzing one or two galaxies with user-defined parameters is in the open-source github repository https://github.com/ismael-mendoza/ShapeMeasurementFisherFormalism.]

79 ASTRONOMY AND ASTROPHYSICS↗

RAPIDS2: A SciDAC Institute for Computer Science, Data, and Artificial Intelligence (University of Delaware)

This report summarizes the activities, technical accomplishments, and outcomes of the RAPIDS2 Institute project at the University of Delaware (UD). The RAPIDS2 Institute was a large multi-institution project with the objective of assisting SciDAC and Office of Science application teams in the use of DOE supercomputing resources to achieve scientific breakthroughs. The UD team contributed to this effort through work on formal software verification. This thrust aims to reduce software developer time and effort, especially regarding debugging and testing, and to increase confidence in the correctness of the results computed by the software.

97 MATHEMATICS AND COMPUTING↗

Implementation of a Dose-based Compliance System for WCS - 19178

Waste Control Specialists LLC (WCS) operates a low-level radioactive waste (LLRW) disposal facility in Andrews, Texas. The site includes a compact waste facility (CWF) that began receiving waste in 2012, a federal waste facility (FWF) that started receiving waste in 2013, as well as a Resource Conservation and Recovery Act (RCRA) landfill and a byproduct facility. The current site license uses a complex system of concentration-based investigation levels and action levels (ILs and ALs) coupled with analytical decision levels (ADLs) that, when exceeded, require specific steps to be taken. These concentration-based ILs and ALs are media specific (i.e. air, soil, water), radionuclide specific, location specific, and disposal facility specific. This results in thousands of individual ILs, ALs, and ADLs that need to be managed and evaluated routinely in order to demonstrate compliance. Further, under the current system, each exceedance of an IL and AL can trigger comparisons to numerous additional isotopic ILs per sample, which is cumbersome, error prone, and costly. This paper describes implementation of a dose-based compliance system that has been developed to replace the existing system. The dose-based compliance system is implemented in RACER{sup C}, a cloud-based software developed by Risk Assessment Corporation (RAC). This system maintains all the analytical results related to the sampling and analysis conducted as part of the Radiological Environmental Monitoring Program (REMP) together with the geo-spatial data allowing instantaneous trending and analysis of the data. Further, the system contains the information needed to compare the analytical results to the new media-specific screening levels that have been developed, and the parameters and routines to compute doses using prescribed exposure scenarios and dose coefficients for compliance evaluation. Once measurement data are entered into RACER, the routines for continuous screening of air, soil, groundwater, and dosimeter data are triggered, and any sample result failing the screening step or exceeding the dose IL/AL is identified and flagged automatically for further review or action, as appropriate. Because RACER is cloud-based, the data, screening, and dose calculations can be accessed remotely at any time, providing a timely and efficient method for identifying any possible releases from the site to the environment as well as determining the potential dose consequences associated with a release. Furthermore, the system is accessible by the regulator providing transparency and ensuring they remain fully informed in an efficient manner. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Award ER25750: Coordinated Infrastructure for Fault Tolerance Systems Indiana University Final Report

The main purpose of the Coordinated Infrastructure for Fault Tolerance in Systems initiative has been to conduct research with a goal of providing end-to-end fault tolerance on a systemwide basis for applications and other system software. While fault tolerance has been an integral part of most high-performance computing (HPC) system software developed over the past decade, it has been treated mostly as a collection of isolated stovepipes. Visibility and response to faults has typically been limited to the particular hardware and software subsystems in which they are initially observed. Little fault information is shared across subsystems, allowing little flexibility or control on a system-wide basis, making it practically impossible to provide cohesive end-to-end fault tolerance in support of scientific applications. As an example, consider faults such as communication link failures that can be seen by a network library but are not directly visible to the job scheduler, or consider faults related to node failures that can be detected by system monitoring software but are not inherently visible to the resource manager. If information about such faults could be shared by the network libraries or monitoring software, then other system software, such as a resource manager or job scheduler, could ensure that failed nodes or failed network links were excluded from further job allocations and that further diagnosis could be performed. As a founding member and one of the lead developers of the Open MPI project, our efforts over the course of this project have been focused on making Open MPI more robust to failures by supporting various fault tolerance techniques, and using fault information exchange and coordination between MPI and the HPC system software stack from the application, numeric libraries, and programming language runtime to other common system components such as jobs schedulers, resource managers, and monitoring tools.

97 MATHEMATICS AND COMPUTING↗

Final Report: Correctness Tools for Petascale Computing

In the course of developing parallel programs for leadership computing systems, subtle programming errors often arise that are extremely difficult to diagnose without tools. To meet this challenge, University of Maryland, the University of Wisconsin—Madison, and Rice University worked to develop lightweight tools to help code developers pinpoint a variety of program correctness errors that plague parallel scientific codes. The aim of this project was to develop software tools that help diagnose program errors including memory leaks, memory access errors, round-off errors, and data races. Research at Rice University focused on developing algorithms and data structures to support efficient monitoring of multithreaded programs for memory access errors and data races. This is a final report about research and development work at Rice University as part of this project.

97 MATHEMATICS AND COMPUTING↗

A Data Exchange Interface for a Standards Based Data Integration Platform

The modern electrical grid is a complex and a data rich system, requiring several sub-systems to provide the capability and support needed to perform functions such as distributed energy resource management, outage management, and fault location and isolation, to name a few. There is a need to develop software solutions and platforms that provide an integrated infrastructure to allow interoperability between these subsystems, while leveraging the abundant data available from today’s grid. In this paper, we present the development and deployment of an application service to integrate GridAPPS-D which is an open source, standards-based software platform with SurvalentONE, a commonly used DNP3 based ADMS platform that allows operation, monitoring, analysis, restoration, and optimization of network operations. We present details of the integration architecture, including its implementation, and provide results from functional testing of the architecture on a 13 bus IEEE system. The results validate that the integration is successful and is able to provide a two-way exchange of communication and translation between different communication protocols.

Singh, Alka↗

Numerical Modeling and Optimization of the iProTech Pitching Inertial Pump (PIP) Wave Energy Converter (WEC) (Cooperative Research and Development Final Report, CRADA Number: CRD-22-22968)

This work generated a first-of-its-kind automated workflow to couple time-domain simulations of wave energy converters written in one software language with a set of design generation and evaluation scripts written in another software language. This automated workflow used an existing optimization package to analyze the sensitivity of different design parameters on the power output of a specific WEC, iProTech’s Pitching Inertial Pump (PIP). Geometric, inertial, and power take-off variables were all varied and optimized to find values that produced the highest amount of power generated over varying wave conditions. The findings on these parameter sensitivity studies are used to inform future design iterations of the PIP WEC. Including more design variables in the optimizations will only increase computational run time and further software development is needed to analyze a larger optimization.

16 TIDAL AND WAVE POWER↗

Lightweight Software Process Improvement Using Productivity and Sustainability Improvement Planning (PSIP)

Productivity and Sustainability Improvement Planning (PSIP) is a lightweight, iterative workflow that allows software development teams to identify development bottlenecks and track progress to overcome them. In this paper, we present an overview of PSIP and how it compares to other software process improvement (SPI) methodologies, and provide two case studies that describe how the use of PSIP led to successful improvements in team effectiveness and efficiency.

Heroux, Michael A.↗

Neutrino interaction vertex reconstruction in DUNE with Pandora deep learning

The Pandora Software Development Kit and algorithm libraries perform reconstruction of neutrino interactions in liquid argon time projection chamber detectors. Pandora is the primary event reconstruction software used at the Deep Underground Neutrino Experiment, which will operate four large-scale liquid argon time projection chambers at the far detector site in South Dakota, producing high-resolution images of charged particles emerging from neutrino interactions. While these high-resolution images provide excellent opportunities for physics, the complex topologies require sophisticated pattern recognition capabilities to interpret signals from the detectors as physically meaningful objects that form the inputs to physics analyses. A critical component is the identification of the neutrino interaction vertex. Subsequent reconstruction algorithms use this location to identify the individual primary particles and ensure they each result in a separate reconstructed particle. A new vertex-finding procedure described in this article integrates a U-ResNet neural network performing hit-level classification into the multi-algorithm approach used by Pandora to identify the neutrino interaction vertex. The machine learning solution is seamlessly integrated into a chain of pattern-recognition algorithms. The technique substantially outperforms the previous BDT-based solution, with a more than 20% increase in the efficiency of sub-1 cm vertex reconstruction across all neutrino flavours.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A Targeted Sequencing Assay for Serotyping Escherichia coli Using AgriSeq Technology

The gold standard method for serotyping Escherichia coli has relied on antisera-based typing of the O- and H-antigens, which is labor intensive and often unreliable. In the post-genomic era, sequence-based assays are potentially faster to provide results, could combine O-serogrouping and H-typing in a single test, and could simultaneously screen for the presence of other genetic markers of interest such as virulence factors. Whole genome sequencing is one approach; however, this method has limited multiplexing capabilities, and only a small fraction of the sequence is informative for subtyping or identifying virulence potential. A targeted, sequence-based assay and accompanying software for data analysis would be a great improvement over the currently available methods for serotyping. The purpose of this study was to develop a high-throughput, molecular method for serotyping E. coli by sequencing the genes that are required for production of O- and H-antigens, as well as to develop software for data analysis and serotype identification. To expand the utility of the assay, targets for the virulence factors, Shiga toxins ( stx 1 , and stx 2 ) and intimin ( eae ) were included. To validate the assay, genomic DNA was extracted from O-serogroup and H-type standard strains and from Shiga toxin-producing E. coli , the targeted regions were amplified, and then sequencing libraries were prepared from the amplified products followed by sequencing of the libraries on the Ion S5™ sequencer. The resulting sequence files were analyzed via the SeroType Caller™ software for identification of O-serogroup, H-type, and presence of stx 1 , stx 2 , and eae . We successfully identified 169 O-serogroups and 41 H-types. The assay also routinely detected the presence of stx 1a,c,d (3 of 3 strains), stx 2c−e,g (8 of 8 strains), stx 2f (1 strain), and eae (6 of 6 strains). Taken together, the high-throughput, sequence-based method presented here is a reliable alternative to antisera-based serotyping methods for E. coli .

Elder, Jacob R.↗

Heat pump concepts for nZEB Technology developments, design tools and testing of heat pump systems for nZEB in the USA: Country report IEA HPT Annex 40 Task 2, Task 3 and Task 4 of the USA

The IEA HPT Annex 40 "Heat pump concepts for Nearly Zero Energy Buildings" deals with the application of heat pumps as a core component of the HVAC system for Nearly or Net Zero energy buildings (nZEB). This report covers Task 2 on the system comparison and optimisation and Task 3 dedicated to the development of adapted technologies for nZEB and field monitoring results of heat pump systems in nZEB. In the US team three institutions are involved and have worked on the following projects: The Oak Ridge National Laboratory (ORNL) will summarize development activities through the field demonstration stage for several integrated heat pump (IHP) systems electric ground-source (GS-IHP) and air-source (AS-IHP) versions and an engine driven AS-IHP version. The first commercial GS-IHP product was just introduced to the market in December 2012. This work is a contribution to Task 3 of the Annex. The University of Maryland will contribute a software development project to Task 2 of the Annex. The software ThermCom evaluates occupied space thermal comfort conditions accounting for all radiative and convective heat transfer effects as well as local air properties. The National Institute of Standards and Technology (NIST) is working on a field study effort on the NIST Net Zero Energy Residential Test Facility (NZERTF). This residential building was constructed on the NIST campus and officially opened in summer 2013. During the first year, between July 2013 and June 2014, baseline performance of the NZERTF was monitored under a simulated occupancy protocol. The house was equipped with an air-to-air heat pump which included a dedicated dehumidification operating mode. Outdoor conditions, internal loads and modes of heat pump operation were monitored. Field study results with respect to heat pump operation will be reported and recommendations on heat pump optimization for a net zero energy building will be provided. This work is a contribution to Task 3 of the Annex.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Combined-Accelerated Stress Testing System for Photovoltaic Modules

Combining the stress factors of the natural environment into a single test for photovoltaic modules requires fewer modules, fewer parallel tests, and makes it possible to discover potential weaknesses that are not known textbfa priori in new module designs. In this work, we present the necessary textbfhardware and textbfsoftware capabilities for developing, implementing, and performing Combined-Accelerated Stress Testing (C-AST) protocols for photovoltaic modules, using off-the-shelf equipment and accessible software development tools, with the aim of supporting the development of C-AST capabilities by the photovoltaic community.

14 SOLAR ENERGY↗

AAPM recommendations on dose prescription and reporting methods for permanent interstitial brachytherapy for prostate cancer: Report of Task Group 137

During the past decade, permanent radioactive source implantation of the prostate has become the standard of care for selected prostate cancer patients, and the techniques for implantation have evolved in many different forms. Although most implants use {sup 125}I or {sup 103}Pd sources, clinical use of {sup 131}Cs sources has also recently been introduced. These sources produce different dose distributions and irradiate the tumors at different dose rates. Ultrasound was used originally to guide the planning and implantation of sources in the tumor. More recently, CT and/or MR are used routinely in many clinics for dose evaluation and planning. Several investigators reported that the tumor volumes and target volumes delineated from ultrasound, CT, and MR can vary substantially because of the inherent differences in these imaging modalities. It has also been reported that these volumes depend critically on the time of imaging after the implant. Many clinics, in particular those using intraoperative implantation, perform imaging only on the day of the implant. Because the effects of edema caused by surgical trauma can vary from one patient to another and resolve at different rates, the timing of imaging for dosimetry evaluation can have a profound effect on the dose reported (to have been delivered), i.e., for the same implant (same dose delivered), CT at different timing can yield different doses reported. Also, many different loading patterns and margins around the tumor volumes have been used, and these may lead to variations in the dose delivered. In this report, the current literature on these issues is reviewed, and the impact of these issues on the radiobiological response is estimated. The radiobiological models for the biological equivalent dose (BED) are reviewed. Starting with the BED model for acute single doses, the models for fractionated doses, continuous low-dose-rate irradiation, and both homogeneous and inhomogeneous dose distributions, as well as tumor cure probability models, are reviewed. Based on these developments in literature, the AAPM recommends guidelines for dose prescription from a physics perspective for routine patient treatment, clinical trials, and for treatment planning software developers. The authors continue to follow the current recommendations on using D{sub 90} and V{sub 100} as the primary quantities, with more specific guidelines on the use of the imaging modalities and the timing of the imaging. The AAPM recommends that the postimplant evaluation should be performed at the optimum time for specific radionuclides. In addition, they encourage the use of a radiobiological model with a specific set of parameters to facilitate relative comparisons of treatment plans reported by different institutions using different loading patterns or radionuclides.

61 RADIATION PROTECTION AND DOSIMETRY↗

Development of an In-Situ Method for Cable Condition Monitoring in Nuclear Power Plants (SBIR Phase IIB Final Report)

This is the final report of a Small Business Innovation Research (SBIR) project that Analysis and Measurement Services Corporation (AMS) has conducted for the U.S. Department of Energy (DOE) over a seven-and-a-half year period (February 2013 through April 2020 with the awards of Phase I, Phase II and Phase IIB projects). The goal of the project was to design, develop, validate, and demonstrate a technique for in-situ testing of cable insulation polymers that will identify, locate, and provide the degree of aging for cables commonly used in nuclear power plants. During the Phase I project, AMS established that the frequency domain reflectometry (FDR) technique can successfully identify and measure cable insulation degradation that can be trended with aging. In Phase II, AMS performed extensive cable aging studies to correlate FDR measurements with other laboratory condition monitoring techniques and developed aging condition categories to quantify the severity of insulation degradation. In Phase IIB, the project expanded the research to include a wider variety of cable polymers that are used in the commercial nuclear power industry. This work also involved developing acceptance criteria to objectively assess age-related cable degradation while sharing the results of this research with industry, academia, and national laboratories to advance the state of the art in cable aging management technologies. The products developed under this research project provide the nuclear industry with an effective condition monitoring tool to support safe and long-term plant operation. Throughout the project, collaborations and support were received from a variety of industry organizations and individuals including DOE National Laboratories as well as nuclear plant utilities and several other industry experts and cable manufacturers. Contributions from these organizations included the donation of new cables, naturally aged cables, and research collaboration. The validation and commercialization of the products of this project were achieved through opportunities to test and demonstrate the technologies’ capabilities on-site at nuclear power generation and research facilities as well as in the laboratory alongside industry peers and cable testing service companies. The research resulted in a technology that can be used to identify, locate, and quantify age-related degradation in several types of cable polymers. This included developing software and hardware as well as the methodology for using the FDR technique to assess age-related degradation of installed cables. The technology developed under this project can provide nuclear plant management, engineers, and technicians with an in-situ electrical test method to determine if in-service cables need to be replaced, monitored on a periodic basis, or show no evidence of significant age-related degradation that may require action. Near the end of this Phase IIB project, the FDR product was sold to a nuclear power utility in South Korea. This sale of a dedicated aging assessment tool is the beginning of a comprehensive contract with the expectation of twelve (12) units sold to that country. Additionally, the FDR technology has been sold to several industry organizations including a nuclear research institute and the Diablo Canyon nuclear power plant. This technology is also being leased by other nuclear industry service companies for incorporation into aging management programs. The cable testing technology that was developed under this project was also integrated into a comprehensive cable aging assessment service that is being offered to the nuclear industry at the request of U.S. nuclear utilities and is currently part of onsite testing services provided by AMS.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Research software engineering: Professionalization, roles, and identity

This report summarizes the results of a study of research software professionals based on 17 interviews conducted at U.S. national laboratories and academic institutions. The study focuses on understanding the emerging professional role of research software engineer (RSE) and examining how the history of software engineering might be relevant to the emergence of research software engineering as a professional movement. The report first uses interview data to identify several models of professional identity for research software professionals. These include dominance models in which a person strongly identifies with a single professional identity (such as RSE), and multiplicity models in which an individual sees an identity like RSE as just one facet of their overall professional role. Next, it identifies three types of work situations for software professionals at national laboratories, covering those who work in dedicated RSE organizations, those who play an integral role in large multiphysics code projects, and those who work in a more isolated capacity to provide software development support for less computationally intensive research projects. It then reviews some key ideas from the sociology of professions in the context of the software engineering profession, suggesting parallels between the current RSE professional movement and aspects of the history of software engineering. It concludes by suggesting some potential practical implications of these findings for the RSE movement and professional organizations.

97 MATHEMATICS AND COMPUTING↗

Automatic Offloading C++ Expression Templates to CUDA Enabled GPUs

In the last few years, many scientific applications have been developed for powerful graphics processing units (GPUs) and have achieved remarkable speedups. This success can be partially attributed to high performance host callable GPU library routines that are offloaded to GPUs at runtime. These library routines are based on C/C++-like programming toolkits such as CUDA from NVIDIA and have the same calling signatures as their CPU counterparts. Recently, with the sufficient support of C++ templates from CUDA, the emergence of template libraries have enabled further advancement in code reusability and rapid software development for GPUs. However, Expression Templates (ET), which have been very popular for implementing data parallel scientific software for host CPUs because of their intuitive and mathematics-like syntax, have been underutilized by GPU development libraries. The lack of ET usage is caused by the difficulty of offloading expression templates from hosts to GPUs due to the inability to pass instantiated expressions to GPU kernels as well as the absence of the exact form of the expressions for the templates at the time of coding. This paper presents a general approach that enables automatic offloading of C++ expression templates to CUDA enabled GPUs by using the C++ metaprogramming technique and Just-In-Time (JIT) compilation methodology to generate and compile CUDA kernels for corresponding expression templates followed by executing the kernels with appropriate arguments. This approach allows developers to port applications to run on GPUs with virtually no code modifications. More specifically, this paper uses a large ET based data parallel physics library called QDP++ as an example to illustrate many aspects of the approach to offload expression templates automatically and to demonstrate very good speedups for typical QDP++ applications running on GPUs against running on CPUs using this method of offloading. In addition, this approach of automatic offlo- ding expression templates could be applied to other many-core accelerators that provide C++ programming toolkits with the support of C++ template.

97 MATHEMATICS AND COMPUTING↗