SEARCH · Engineering Papers
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.
These points are covered in this presentation: Distributed GPU stencil, non-contiguous data; Equivalence of strided datatypes and minimal representation; GPU communication methods; Deploying on managed systems; Large messages and MPI datatypes; Translation and canonicalization; Automatic model-driven transfer method selection; and Interposed library implementation.
Multigrid methods are well suited to large massively parallel computer architectures because they are mathematically optimal and display good parallelization properties. Since current architecture trends are favoring regular compute patterns to achieve high performance, the ability to express structure has become much more important. The hypre software library provides high-performance multigrid preconditioners and solvers through conceptual interfaces, including a semistructured interface that describes matrices primarily in terms of stencils and logically structured grids. This paper presents a new semistructured algebraic multigrid (SSAMG) method built on this interface. The numerical convergence and performance of a CPU implementation of this method are evaluated for a set of semistructured problems. In conclusion, SSAMG achieves significantly better setup times than hypre’s unstructured AMG solvers and comparable convergence. In addition, the new method is capable of solving more complex problems than hypre’s structured solvers.
Systematic software construction offers the potential of elevating software engineering from an art-form to an engineering discipline. The desired result is more predictable software development leading to better quality and more maintainable software. However, the overhead costs associated with the formalisms, mathematics, and methods of systematic software construction have largely precluded their adoption in real-world software development. In fact, many mainstream software development organizations, such as Microsoft, still maintain a predominantly oral culture for software development projects; which is far removed from a formalism-based culture for software development. An exception is the limited domain of safety-critical software, where the high-assuiance inherent in systematic software construction justifies the additional cost. We believe that systematic software construction will only be adopted by mainstream software development organization when the overhead costs have been greatly reduced. Two approaches to cost mitigation are reuse (amortizing costs over many applications) and automation. For the last four years, NASA Ames has funded the Amphion project, whose objective is to automate software reuse through techniques from systematic software construction. In particular, deductive program synthesis (i.e., program extraction from proofs) is used to derive a composition of software components (e.g., subroutines) that correctly implements a specification. The construction of reuse libraries of software components is the standard software engineering solution for improving software development productivity and quality.
Nuclear criticality modeling and simulations rely on the quality of the existing evaluated nuclear data libraries such as Evaluated Nuclear Data File (ENDF)/B, the Joint Evaluated Fission and Fusion (JEFF) nuclear data library, or the Japanese Evaluated Nuclear Data Library (JENDL). In some cases, the cross-section evaluations of those libraries were found to be deficient in describing criticality benchmarks accurately. More than two decades ago, the US Nuclear Criticality Safety Program (NCSP) established a Nuclear Data (ND) task which encompassed experiments and evaluations. In response to this, the Oak Ridge National Laboratory (ORNL) formed a Nuclear Criticality and Data group which performed ND experiments, data analysis, and evaluations to produce ENDF files for the ND libraries as identified in the NCSP Five-Year Plan. Before being submitted to the ENDF library, files were processed and tested for performance by running benchmark calculations. This procedure was centralized in the ORNL group and is now often referred to as the ND pipeline. NCSP collaborates with the Joint Research Center (JRC) of the European Commission in Geel, Belgium, to perform high-resolution neutron-induced cross section measurements at the Geel Linear Accelerator (GELINA). The objective is to address emerging ND problems in criticality calculations. Difficulties with ND include insufficient neutron energy range, missing covariances, and previously unrecognized inaccuracies with experiments. New neutron total and capture cross sections of 90 Zr in the neutron energy range from 100 eV to several hundred keV were recently performed. These measured data will be used, together with existing high-resolution transmission data from a metallic 90 Zr sample, to improve representation of the cross sections.
Quantum circuit execution is a central task in quantum computation. Due to inherent quantum-mechanical constraints, quantum computing workflows often involve a considerable number of independent measurements over a large set of slightly different quantum circuits. Here we discuss a simple model for parallelizing such quantum circuit executions that is based on introducing a large array of virtual quantum processing units (mapped to HPC nodes in our case) as a parallel quantum computing platform. Implemented within the XACC framework, the model can readily take advantage of its backend-agnostic features, enabling parallel quantum computing/simulation over any target backend supported by XACC. We illustrate the performance of this approach by demonstrating strong scaling in two pertinent domain science problems, namely in computing the gradients for the multi-contracted variational quantum eigensolver and in data-driven quantum circuit learning, where we vary the number of qubits and the number of circuit layers. Here, the latter simulation leverages the cuQuantum library to run efficiently on GPU-accelerated HPC platforms.
This DOE grant supported the development of electronic structure theory, algorithms, and computer programs for simulating singlet fission and multiple excision generation processes from Sept. 2013 to Aug. 2018 (which was terminated due to the PI’s departure from Northwestern University to industry). One of the main achievements is the development of the so-called active space decomposition (ASD) method hat describes exciton dynamics in organic semiconductors and spin dynamics in strongly correlated materials. The resulting software was interfaced with the Brilliantly Advanced General Electronic-structure Library (BAGEL) package and widely disseminated to the research community.
A computer program framework, YAHFC, to model low-energy nuclear reactions is presented. The framework allows for reactions with incident particles ranging from protons/neutrons to alphas and is designed to address reactions that ultimately lead to the formation of compound nuclear systems that then decay statistically as outlined in concepts of Hauser and Feshbach. Additionally, instead of a reaction, it is also possible to model the decay of a nuclear system with an initial excitation and population. The code models nuclear decays with a Monte Carlo process that tracks the decay of each state. This allows for an exact representation of the spectra for all emitted particles in each of the final exit channels and the possibility of generating reaction data for simulation purposes. The program is interfaced with the optical model code system FRESCOX to calculate transmission coefficients as well as the effects of coupled channels and other direct excitations via the distorted wave Born approximation (DWBA). Modules are included to account for nuclear processes such as width corrections, pre-equilibrium emission, and fission. The program is controlled by a series of input commands and while a set of input parameters exists for each projectile and target, the input commands allow for complete control over each input parameter. Extensive data files are produced and a program is provided that converts YAHFC data files into nuclear data library entries in the generalized nuclear data structure (GNDS).
This report is an institutional record of experiments conducted to explore performance of a vendor installation of CephFS on the SNL stria cluster. Comparisons between CephFS, the Lustre parallel file system, and NFS were done using the IOR and MDTEST benchmarking tools, a test program which uses the SEACAS/Trilinos IOSS library, and the checkpointing activity performed by the LAMMPS molecular dynamics simulation.
This report provides the documentation for the revised procedure to generate the MC 2 -3 gamma library with the latest versions of PreGAMMA and GenGAMMA which have recently been updated to fix the program errors and inconsistent data processing for some isotopes identified in a recent verification and validation study of the MC 2 -3 gamma library. A few associated errors associated with gamma processing were also identified and fixed in the MC 2 -3 code. Addressing the changes in MC 2 -3, the procedure to generate the PMATRX and GAMISO datasets is discussed as well. The input data of PreGAMMA and GenGAMMA are discussed in detail using sample input data. In addition, the structure and formats of the MC 2 -3 gamma library are provided in detail.
The computational bias of criticality safety computer codes must be established through the validation of the codes to critical experiments. A large collection of suitable experiments has been vetted by the International Criticality Safety Benchmark Evaluation Project (ICSBEP) and made available in the International Handbook of Evaluated Criticality Safety Benchmark Experiments (ICSBEP Handbook). More than 600 cases from this handbook have been prepared and reviewed within the Verified, Archived Library of Inputs and Data (VALID), which is maintained by the Reactor and Nuclear Systems Division at Oak Ridge National Laboratory. The performance of the KENO V.a and KENO-VI Monte Carlo codes within the SCALE 6.2.4 code system is assessed using the VALID models of benchmark experiments. A range of nuclear cross section libraries based on Evaluated Nuclear Data File (ENDF)/B-VII.1 in both multigroup (MG) and continuous energy (CE) formats is considered. The critical experiments available to validate the KENO V.a code cover 15 broad categories of systems. These systems use a range of fissile materials, including a range of uranium enrichments, various plutonium isotopic vectors, and some mixed uranium/plutonium oxides. The physical forms of the fissile material also vary and are represented as metal, solutions, or arrays of rods or plates in a water moderator. The neutron energy spectra of the systems also vary and cover fast, intermediate, mixed, and thermal spectra. Over 550 of the total cases use the KENO V.a code for the four nuclear data libraries considered in this report.
The Multiphysics Object-Oriented Simulation Environment (MOOSE) framework is a foundational capability used by the Nuclear Energy Advanced Modeling and Simulation (NEAMS) program to create over 15 different simulation tools for advanced nuclear reactors. Due to this broad use, improvements to the framework in support of modeling and simulation goals are critical to the program. These improvements can take many forms, including optimization, improved user experience, streamlined application programming interfaces (APIs), parallelism, and new capabilities. The work transcribed in this report was conducted in direct support of the simulation tools and has already been deployed or will be deployed in the coming months. The capabilities were implemented in the same order as they are covered in this report: increased support of face variables, arbitrary spatial and temporal evaluation of material properties, and the addition of a triangular meshing library in libMesh.
NOODLES is a new cross-domain collaborative visualization and analysis capability being developed at NREL. The NOODLES specification is a minimal protocol that can tie different software packages and platforms together. As an example, a simulation code that speaks this protocol could stream iso-surfaces to an immersive VR space and a web browser. Another example is a team (with some members across the country) classifying and discovering features in a statistical data plot, driven from a researcher's existing Python-based workflow. In this talk, we will discuss some background on the problems that this protocol intends to solve, basic principles of the specification, the current state of supporting libraries, and some demonstrations of the protocol in action.
This presentation discusses the Verified, Archived Library of Inputs and Data (VALID) procedure which is managed per computational procedure under the overall SCALE quality assurance plan. The principle behind VALID is that two independent, qualified people prepare and review the model, while a dual, independent check acts as a primary barrier to prevent errors in the library. A final check by VALID QAC of procedural compliance is then performed. The future for VALID includes plans to add more experiments, a review for all ICSBEP cases containing Deuterium, ZEUS intermediate and fast spectrum evaluations, intermediate enrichment experiments, other cases of interest to sponsors, and mixed/intermediate spectrum experiments.
Not Available
This report is an institutional record of experiments conducted to explore performance of a vendor installation of CephFS on the SNL stria cluster. Comparisons between CephFS, the Lustre parallel file system, and NFS were done using the IOR and MDTEST benchmarking tools, a test program which uses the SEACAS/Trilinos IOSS library, and the checkpointing activity performed by the LAMMPS molecular dynamics simulation.
The auxiliary field quantum Monte Carlo algorithm in QMCPACK was ported to run on AMD GPUs using HIP. Initial performance measurements on early access hardware indicate a 4x slowdown compared to NVIDIA V100s. The slowdown was determined to be a result of suboptimal ROCM libraries for small matrix sizes coupled with some poorly performing hipified CUDA kernels. Through a preliminary analysis of HIP kernel performance we determined a strategy for pro ling and tuning kernels. We request that ROCM libraries are optimized for smaller problem sizes particularly for batched operations.
This presentation discusses the nuclear data management code FUDGE (For Updating Data and Generating Evaluations). While FUDGE is designed to support GNDS, the presentation also states that ENDF-6 and ENDL data files are also supported but must first be translated into GNDS; translators are included with FUDGE. It also states that FUDGE supports plotting, manipulating, checking for physical content, resonance reconstruction, Doppler broadening, etc. FUDGE also supports processing for Monte Carlo and deterministic transport. FUDGE is open source, and the latest public release was FUDGE-4.2.3 with support for GNDS-1.9. In summation, FUDGE capabilities include translating older formats into GNDS, translating GNDS back to ENDF-6, and visualizing, modifying, checking, and processing GNDS data. It also states that a new version of FUDGE is coming soon. The plan is to support GNDS-2.0 specification, but it may be released sooner if v2.0 is delayed.
The ForTrilinos subproject of the Exascale Computing Project (ECP) was initiated to bring the capabilities and scalability of the Trilinos numerical solver collection to Fortran scientific application codes. A novel Fortran extension to the Simplified Wrapper and Interface Generator (SWIG) tool, which automatically generates Fortran bindings from existing C/C++ library code, has been applied to key Trilinos solver libraries to generate the new ForTrilinos libraries. SWIG-Fortran has additionally been used to generate new Fortran compatibility layers for additional scientific libraries and applications. This report summarizes the products and impact of the ForTrilinos subproject.