Search NASA⌕ Search

SEARCH · Search NASA

Results for “sequential experiments”

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 775 records · Page 43

Curriculum-based Reinforcement Learning for Distribution System Critical Load Restoration

This paper focuses on the critical load restoration problem in distribution systems following major outages. To provide fast online response and optimal sequential decision-making support, a reinforcement learning (RL) based approach is proposed to optimize the restoration. Due to the complexities stemming from the large policy search space, renewable uncertainty, and nonlinearity in a complex grid control problem, directly applying RL algorithms to train a satisfactory policy requires extensive tuning to be successful. To address this challenge, this paper leverages the curriculum learning (CL) technique to design a training curriculum involving a simpler steppingstone problem that guides the RL agent to learn to solve the original hard problem in a progressive and more effective manner. We demonstrate that compared with direct learning, CL facilitates controller training to achieve better performance. To study realistic scenarios where renewable forecasts used for decision-making are in general imperfect, the experiments compare the trained RL controllers against two model predictive controllers (MPCs) using renewable forecasts with different error levels and observe how these controllers can hedge against the uncertainty. Results show that RL controllers are less susceptible to forecast errors than the baseline MPCs and can provide a more reliable restoration process.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Highly scalable distributed-memory sparse triangular solution algorithms.

This paper presents a highly efficient distributed-memory parallel sparse triangular solver. The triangular solution phase is often performed following factorization phase in the sparse linear solvers and has become increasingly computa­tionally expensive for direct solvers with many right hand sides (RHSs) or preconditioned iterative solvers. However, the low arithmetic intensity and sequential nature of the triangular solve algorithm pose performance challenges for its large-scale distributed-memory parallelization. In this work, we propose several strategies to enhance scalability of an algorithm with 2D block cyclic process layout. First, an asynchronous binary-tree-based communication scheme implemented via non-blocking MPI functions is leveraged to broadcast partial solutions and reduce partial updates among a subset of processes for each block column and row of the triangular matrix, respectively. This scheme reduces message latency, improves communication load balance and significantly accelerates asynchronous execution of the triangular solve. In addition, efficient BLAS operations and threading implementations are exploited to accelerate local computations and further reduce process idle time. The proposed strategies are implemented in SuperLU_DIST and numerical experiments show up to 4.4x improvement with one right-hand side and up to 6.1x improvement with 50 right- hand sides on 4096 processes, compared to the current release. This is the first time that sparse triangular solution is demonstrated strong scaling on more than 4000 cores.

Liu, Yang↗

Validation and Verification of MCNP6 Against High-Energy Experimental Data and Calculations by Other Codes. III. The MPI Testing Primer

This primer presents a set of Validation and Verification (V&V) MCNP6 results calculated in parallel, with MPI, obtained mostly using its event generators at intermediate and high-energies, but also several examples of results using data libraries at energies below 150 MeV, against various recent experimental data. It is the third part of a set of MCNP6 Testing Primers, after its first, LA-UR-11-05129, and second, LA-UR-11-05627, publications. The MCNP6 test-problems discussed here are presented in the /VALIDATION CEM/, /VALIDATION LAQGSM/, and /DELAYED PARTICLES/ subdirectories in the MCNP6 /Testing/ directory. README files that contain short descriptions of every input file, the experiment, the quantity of interest that the experiment measures and its description in the MCNP6 output files, and the publication reference of that experiment are presented for every test-problem. Templates for plotting the corresponding results with xmgrace as well as pdf files with figures representing the final results of our V&V efforts are presented. Several technical “bugs” in MCNP6 while running it in parallel with MPI using its event generators were discovered during our current V&V of MCNP6 and were fixed mostly by Dr. Jeffrey S. Bull. Our results show that MCNP6 using its CEM03.03, LAQGSM03.03, Bertini, INCL+ABLA, and ISABEL event generators describes, as a rule, reasonably well different intermediate- and high-energy measured data and agrees very well with calculations by other codes. Just as expected, the MCNP6 results obtained with MPI are practically the same as the ones calculated with MCNP6 in a sequential mode. This primer isn’t meant to be read from cover to cover. Readers may skip some sections and go directly to a test-problem they are interested in.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Motion detection in astronomical and ice floe images

Two approaches are presented for establishing correspondence between small areas in pairs of successive images for motion detection. The first one, based on local correlation, is used on a pair of successive Voyager images of the Jupiter which differ mainly in locally variable translations. This algorithm is implemented on a sequential machine (VAX 780) as well as the Massively Parallel Processor (MPP). In the case of the sequential algorithm, the pixel correspondence or match is computed on a sparse grid of points using nonoverlapping windows (typically 11 x 11) by local correlations over a predetermined search area. The displacement of the corresponding pixels in the two images is called the disparities to cubic surfaces. The disparities at points where the error between the computed values and the surface values exceeds a particular threshold are replaced by the surface values. A bilinear interpolation is then used to estimate disparities at all other pixels between the grid points. When this algorithm was applied at the red spot in the Jupiter image, the rotating velocity field of the storm was determined. The second method of motion detection is applicable to pairs of images in which corresponding areas can experience considerable translation as well as rotation.

Manohar, M.↗

Thermal Hydroquinone Oxidation on Co/N-doped Carbon Proceeds by a Band-Mediated Electrochemical Mechanism

Molecules catalyze aerobic oxidation reactions via redox cycling at the metal center to effect sequential activation of O 2 and the substrate. Metal surfaces can catalyze the same transformations by coupling independent half-reactions for oxygen reduction and substrate oxidation mediated via the exchange of band-electrons. Metal/nitrogen-doped carbons (MNCs) are promising catalysts for aerobic oxidation that consist of molecule-like active sites embedded in a conductive carbon host. Owing to their combining molecular and metallic features, it remains unclear whether they proceed via the sequential redox cycling pathways of molecules or the band-mediated pathways of metals. Herein, we simultaneously track the potential of the catalyst and the rate of the turnover of aerobic hydroquinone oxidation on a cobalt-based MNC catalyst contacted to a catalytically inactive carbon electrode. By comparing operando measurements of rate and potential with the current-voltage behavior of each constituent half-reaction under identical conditions, we show that these molecular materials can display the band-mediated reaction mechanisms of extended metallic solids. Further, we show that the action of these band-mediated mechanisms explains fractional reaction orders in both oxygen and hydroquinone, the time evolution of catalyst potential and rate, and the dependence of rate on overall reaction free energy. Selective poisoning experiments suggest that oxygen reduction proceeds at cobalt sites, whereas quinone oxidation proceeds at native oxidic defects on the carbon. These findings highlight that molecule-like active sites can take advantage of band-mediated mechanisms when coupled to conductive hosts.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

PETSc/TAO Users Manual (Rev. 3.19)

This manual describes the use of the Portable, Extensible Toolkit for Scientific Computation (PETSc) and the Toolkit for Advanced Optimization (TAO) for the numerical solution of partial differential equations and related problems on high-performance computers. PETSc/TAO is a suite of data structures and routines that provide the building blocks for the implementation of large-scale application codes on parallel (and serial) computers. PETSc uses the MPI standard for all distributed memory communication. PETSc/TAO includes a large suite of parallel linear solvers, nonlinear solvers, time integrators, and opti mization that may be used in application codes written in Fortran, C, C++, and Python (via petsc4py; see Getting Started). PETSc provides many of the mechanisms needed within parallel application codes, such as parallel matrix and vector assembly routines. The library is organized hierarchically, enabling users to employ the level of abstraction that is most appropriate for a particular problem. By using techniques of object-oriented programming, PETSc provides enormous flexibility for users. PETSc is a sophisticated set of software tools; as such, for some users it initially has a much steeper learning curve than packages such as MATLAB or a simple subroutine library. In particular, for individuals without some computer science background, experience programming in C, C++, python, or Fortran and experience using a debugger such as gdb or lldb, it may require a significant amount of time to take full advantage of the features that enable efficient software use. However, the power of the PETSc design and the algorithms it incorporates may make the efficient implementation of many application codes simpler than “rolling them” yourself. For many tasks a package such as MATLAB is often the best tool; PETSc is not intended for the classes of problems for which effective MATLAB code can be written. There are several packages, built on PETSc, that may satisfy your needs without requiring directly using PETSc. We recommend reviewing these packages functionality before starting to code directly with PETSc. PETSc can be used to provide a “MPI parallel linear solver” in an otherwise sequential, or OpenMP parallel code. This approach cannot provide extremely large improvements in the application time by utilizing large numbers of MPI processes but can still improve the performance. Certainly all parts of a previously sequential code need not be parallelized but the matrix generation portion must be parallelized to expect true scalability to large numbers of MPI processes. See PCMPI for details on how to utilize the PETSc MPI linear solver server. Since PETSc is under continued development, small changes in usage and calling sequences of routines will occur. PETSc has been supported for twenty-five years; see mailing list information on our website for information on contacting support.

97 MATHEMATICS AND COMPUTING↗

Workflow for Developing and Operating Subsurface Hydrogen Storage Facilities in Porous Reservoirs

Long-duration (seasonal) storage of natural gas (NG), which primarily consists of methane (CH 4 ), has been practiced for more than a hundred years at underground gas storage (UGS) facilities that use depleted hydrocarbon reservoirs, saline aquifers, and salt caverns. To enable hydrogen (H 2 ) to be used as a long-duration, energy-storage medium, similar facilities are envisioned for underground H 2 storage (UHS) of either H 2 or H 2 /NG mixtures. Experience with UGS can be used to guide recommended practices for developing and operating UHS facilities in porous reservoirs. The most important factors (formation/fluid properties and engineering choices) that influence the performance of UHS reservoirs have been identified and quantified in previous studies. These factors and choices influence phenomena that determine the sweep efficiency of the stored working gas. These phenomena include viscous fingering, hysteretic capillary trapping, and gravity override of the working gas, as well as the upconing of nonproductive fluid that determine the sweep efficiency of the stored working gas. This report describes initial recommended-practices and a project-development workflow for UHS facilities that utilize porous reservoirs, based on the current state-of-knowledge about H 2 behavior in the subsurface. The workflow sequentially addresses all aspects of UHS project development, including the identification of H 2 sources and users, site ranking and down-selection, geologic and reservoir-engineering characterization, reservoir design, testing, risk management, commissioning, operations, and monitoring for a UHS facility. The goal is to enable UHS facilities to be developed in an efficient and timely manner, while carefully managing project risks. This workflow is similar to that which has been developed for UGS facilities (see Figure 1 of API, 2022), with the addition of tasks and subtasks specific to H 2 and UHS. The project-development workflow is broken down into three major stages: (1) define the H 2 use case; (2) rank, down-select, and characterize potential, candidate UHS sites; and (3) reservoir design, integrity testing, risk assessment, commissioning, operations, and monitoring for selected UHS sites. Each major stage is further broken down into tasks and subtasks, which are described at a high level. This report also provides more detailed descriptions of all tasks and subtasks that involve reservoir analysis and testing.

08 HYDROGEN↗

Self-Driving Telescopes: Autonomous Scheduling of Astronomical Observation Campaigns with Offline Reinforcement Learning

Modern astronomical experiments are designed to achieve multiple scientific goals, from studies of galaxy evolution to cosmic acceleration. These goals require data of many different classes of night-sky objects, each of which has a particular set of observational needs. These observational needs are typically in strong competition with one another. This poses a challenging multi-objective optimization problem that remains unsolved. The effectiveness of Reinforcement Learning (RL) as a valuable paradigm for training autonomous systems has been well-demonstrated, and it may provide the basis for self-driving telescopes capable of optimizing the scheduling for astronomy campaigns. Simulated datasets containing examples of interactions between a telescope and a discrete set of sky locations on the celestial sphere can be used to train an RL model to sequentially gather data from these several locations to maximize a cumulative reward as a measure of the quality of the data gathered. We use simulated data to test and compare multiple implementations of a Deep Q-Network (DQN) for the task of optimizing the schedule of observations from the Stone Edge Observatory (SEO). We combine multiple improvements on the DQN and adjustments to the dataset, showing that DQNs can achieve an average reward of 87%+-6% of the maximum achievable reward in each state on the test set. This is the first comparison of offline RL algorithms for a particular astronomical challenge and the first open-source framework for performing such a comparison and assessment task.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Exploring the Photophysics of N-Type Polymer N2200

The polymer interphase is a complex environment comprised of polymer chains, solvent molecules, and electrolyte ions. However, the microenvironments that these ingredients produce is both poorly-understood and difficult to investigate given the system's inherent structural heterogeneity. Nonetheless, a fundamental understanding of the structure, electrochemical behavior, and excited-state dynamics is a key prerequisite to optimizing devices based on such a system - for instance, photoelectrochemical cells for the direct, light-driven production of solar hydrogen, which may minimize efficiency losses compared to sequential light harvesting and electrochemical hydrogen production in separate but linked devices. This work combines transient absorption (TA) spectroscopy, which is a powerful tool for the observation of a variety of excited states on a femto- to microsecond timescale, with time-resolved microwave conductivity (TRMC) spectroscopy, which selectively probes the dynamics of free charge carriers (FC), of a blend of a donor polymer (PTB7-Th) and acceptor polymer (N2200) in a bulk heterojunction. Initial results indicate an extension of FC lifetime and homogenization of the microenvironment on a whole-film level, not just at the surface of the polymer; however, many questions are still under active investigation, including the changes in energy of the involved polaron states, effective conjugation length, FC mobility and migration distance, etc. Additional experiments regarding the photoinitiated behavior of the blend in different environments and on different timescales are ongoing.

charge recombination↗

A Performance Model for Allocating the Parallelism in a Multigrid-in-Time Solver

The traditional way to numerically solve time-dependent problems is to sequentially march through time, solving for one time step and then the next. The parallelism in this approach is limited to the spatial dimension, which is quickly exhausted, causing the gain from using more processors to solve a problem to diminish. One approach to overcome this barrier is to use methods that are parallel in time. These methods have the potential to achieve dramatically better performance compared to time-stepping approaches, but achieving this performance requires carefully choosing the amount of parallelism devoted to space versus the amount devoted to time. Here, we present a performance model that, for a multigrid-in-time solver, makes the decision on when to switch to parallel-in-time and on how much parallelism to devote to space vs. time. In our experiments, the model selects the best parallel configuration in most of our test cases and a configuration close to the best one in all other cases.

97 MATHEMATICS AND COMPUTING↗

Identification of Low-level Point Radioactive Sources using a sensor network

Identification of a low-level point radioactive source amidst background radiation is achieved by a network of radiation sensors using a two-step approach. Based on measurements from three or more sensors, a geometric difference triangulation method or an N-sensor localization method is used to estimate the location and strength of the source. Then a sequential probability ratio test based on current measurements and estimated parameters is employed to finally decide: (1) the presence of a source with the estimated parameters, or (2) the absence of the source, or (3) the insufficiency of measurements to make a decision. This method achieves specified levels of false alarm and missed detection probabilities, while ensuring a close-to-minimal number of measurements for reaching a decision. This method minimizes the ghost-source problem of current estimation methods, and achieves a lower false alarm rate compared with current detection methods. This method is tested and demonstrated using: (1) simulations, and (2) a test-bed that utilizes the scaling properties of point radioactive sources to emulate high intensity ones that cannot be easily and safely handled in laboratory experiments.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Oxygen Production from Lunar Regolith using Ionic Liquids

The objective of this work and future follow-on work is to develop a safe, efficient, and recyclable method for oxygen and/or metals extraction from lunar regolith, in support of establishing a manned lunar outpost. The approach is to solubilize the oxides that comprise lunar regolith in media consisting of ionic liquids (ILs) and/or their mixtures at temperatures at or below 300 C. Once in solution, electrolysis can either be performed in-situ to generate oxygen at the anode and hydrogen and/or metals (silicon, iron, aluminum, titanium, etc.) at the cathode. Alternatively, the water that is generated during the solubilization process can be distilled out and condensed into a separate IL and then electrolysized to produce hydrogen and oxygen. In the case of lunar regolith, this method could theoretically produce 44g oxygen per 100g of regolith. The oxygen can be used for human life support and/or as an oxidizer for rocket fuels, and the metals can be used as raw materials for construction and/or device fabrication. Moreover, the hydrogen produced can be used to re-generate the acidic medium, which can then be used to process additional regolith, thereby making the materials recyclable and limiting upmass requirements. An important advantage of IL acid systems is that they are much "greener" and safer than conventional materials used for regolith processing such as sulfuric or hydrochloric acids. They have very low vapor pressures, which means that they contain virtually no toxic and/or flammable volatile content, they are relatively non-corrosive, and they can exhibit good stability in harsh environments (extreme temperatures, hard vacuum, etc.). Furthermore, regolith processing can be achieved at lower temperatures than other processes such as molten oxide electrolysis or hydrogen reduction, thereby reducing initial power requirements. Six ILs have been synthesized and tested for their capability to dissolve lunar simulant, and for electrochemical and thermal stability. The results showed that ILs can be very efficient electrolytes; in particular IL/phosphoric-acid mixtures appear extremely promising for solubilizing lunar simulant. Results from preliminary experiments for distillation of water produced from the oxygen within the metal oxides of the simulant and the hydrogen from the acid indicates that over 75% of the oxygen from the simulant can be harvested as water at a temperature of 150 C. A method for collection of oxygen from electrolysis of the water derived from solubilizing simulant was developed by using a liquid nitrogen trap to liquefy and collect the oxygen. Although precise quantification of the liquid oxygen trapped is difficult to obtain, the amount of hydrogen and oxygen collected from electrolysis of water in this system was greater than 98%. This set-up also included a portable mass spectrometer for the identification of gases released from electrolysis cells. Regeneration of ILs through re-protonation was also demonstrated. Four sequential re-generations of an IL following solubilization of simulant showed no significant differences in amounts of simulant dissolved. Follow-on work for this project should include more studies of IL/phosphoric acid systems. Also, much more work is necessary for defining methods for electrolysis and purification of metals from regolith solubilized in ILs, and for developing a system to use the produced hydrogen to regenerate the spent IL. Finally, design and development of flight breadboard and prototype hardware is required.

Paley, Mark Steven↗

Sequential Separation of Linear, Mono–, and Di–Branched Hexane Isomers on a Robust Coordination Polymer with Nonbonding Flexibility

Efficient separation of hexane isomers is a crucial process for upgrading gasoline. Herein, the sequential separation of linear, mono-, and di-branched hexane isomers by a robust stacked 1D coordination polymer termed as Mn-dhbq ([Mn(dhbq)(H 2 O) 2 ], H 2 dhbq = 2,5-dihydroxy-1,4-benzoquinone) is reported. The interchain space of the activated polymer is of optimal aperture size (5.58 Å) that could exclude 2,3-dimethylbutane, while the chain structure can discriminate n-hexane with high capacity (1.53 mmol g –1 at 393 K, 6.67 kPa) by high-density open metal sites (5.18 mmol g –1 ). With the temperature- and adsorbate-dependent swelling of interchain spaces, the affinity between 3-methylpentane and Mn-dhbq can be deliberately controlled from sorption to exclusion, and thus a complete separation of ternary mixture can be achieved. Column breakthrough experiments confirm the excellent separation performance of Mn-dhbq. Furthermore, the ultrahigh stability and easy scalability further highlight the application prospect of Mn-dhbq for separation of hexane isomers.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Fragmentation pathways of tungsten hexacarbonyl clusters upon electron ionization

Electron ionization of neat tungsten hexacarbonyl (W(CO){sub 6}) clusters has been investigated in a crossed electron-molecular beam experiment coupled with a mass spectrometer system. The molecule is used for nanofabrication processes through electron beam induced deposition and ion beam induced deposition techniques. Positive ion mass spectra of W(CO){sub 6} clusters formed by electron ionization at 70 eV contain the ion series of the type W(CO){sub n}{sup +} (0 ≤ n ≤ 6) and W{sub 2}(CO){sub n}{sup +} (0 ≤ n ≤ 12). In addition, a series of peaks are observed and have been assigned to WC(CO){sub n}{sup +} (0 ≤ n ≤ 3) and W{sub 2}C(CO){sub n}{sup +} (0 ≤ n ≤ 10). A distinct change of relative fragment ion intensity can be observed for clusters compared to the single molecule. The characteristic fragmentation pattern obtained in the mass spectra can be explained by a sequential decay of the ionized organometallic, which is also supported by the study of the clusters when embedded in helium nanodroplets. In addition, appearance energies for the dissociative ionization channels for singly charged ions have been estimated from experimental ion efficiency curves.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Synchrotron X-radiolysis of l -cysteine at the sulfur K-edge: Sulfurous products, experimental surprises, and dioxygen as an oxidoreductant

In-situ inventory of sulfurous products from the sulfur K-edge synchrotron X-radiolysis of L-cysteine, in solid-phase and anaerobic (pH 5) and air-saturated (pH 5, pH 7, and pH 9) solutions without and with 40% glycerol is reported. Sequential K-edge X-ray Absorption Spectroscopic (XAS) spectra were acquired. L-cysteine degraded systematically in the X-ray beam. Radiolytic products were inventoried by fits using the XAS spectra of sulfur model compounds. Solid L-cysteine declined to 92% fraction after a single K-edge XAS scan. After six scans 60 % remained, accompanied by 14 % cystine, 16 % thioether, 5.4 % of elemental sulfur, and smaller fractions of more highly oxidized products. In air-saturated pH 5 solution, 73 % of L-cysteine remained after ten scans, with 2 % cystine and 19 % elemental sulfur. Oxidation increased with 40% glycerol, yielding 67 %, 5 % and 23 % fractions, respectively, after ten-scans. Higher pH solutions exhibited less radiolytic chemistry. All the reactivity followed first-order kinetics. The anaerobic experiment displayed two reaction phases, with sharp changes in kinetics and radiolytic chemistry. Unexpectedly, the radiolytic oxidation of L-cysteine was increased in anaerobic solution. After ten-scans only 60% of the L-cysteine remained, along with 17 % cystine, 22 % elemental sulfur, and traces of more highly oxidized products. A new aerobic reaction cycle is hypothesized, wherein dissolved dioxygen captures radiolytic H• or e$-\atop{aq}$, enters HO 2 •/O 2 •-, reductively quenches cysteine thiyl radical, and cycles back to O 2 . This cycle is suggested to suppress the radiolytic production of cystine in aerobic solution.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Adaptive scapula bone remodeling computational simulation: Relevance to regenerative medicine

Shoulder arthroplasty success has been attributed to many factors including, bone quality, soft tissue balancing, surgeon experience, and implant design. Improved long-term success is primarily limited by glenoid implant loosening. Prosthesis design examines materials and shape and determines whether the design should withstand a lifetime of use. Finite element (FE) analyses have been extensively used to study stresses and strains produced in implants and bone. However, these static analyses only measure a moment in time and not the adaptive response to the altered environment produced by the therapeutic intervention. Computational analyses that integrate remodeling rules predict how bone will respond over time. Recent work has shown that subject-specific two- and three dimensional adaptive bone remodeling models are feasible and valid. Feasibility and validation were achieved computationally, simulating bone remodeling using an intact human scapula, initially resetting the scapular bone material properties to be uniform, numerically simulating sequential loading, and comparing the bone remodeling simulation results to the actual scapula’s material properties. Three-dimensional scapula FE bone model was created using volumetric computed tomography images. Muscle and joint load and boundary conditions were applied based on values reported in the literature. Internal bone remodeling was based on element strain-energy density. Initially, all bone elements were assigned a homogeneous density. All loads were applied for 10 iterations. After every iteration, each bone element’s remodeling stimulus was compared to its corresponding reference stimulus and its material properties modified. The simulation achieved convergence. At the end of the simulation the predicted and actual specimen bone apparent density were plotted and compared. Location of high and low predicted bone density was comparable to the actual specimen. High predicted bone density was greater than actual specimen. Low predicted bone density was lower than actual specimen. Differences were probably due to applied muscle and joint reaction loads, boundary conditions, and values of constants used. Work is underway to study this. Nonetheless, the results demonstrate three dimensional bone remodeling simulation validity and potential. Such adaptive predictions take physiological bone remodeling simulations one step closer to reality. Computational analyses are needed that integrate biological remodeling rules and predict how bone will respond over time. We expect the combination of computational static stress analyses together with adaptive bone remodeling simulations to become effective tools for regenerative medicine research.

60 APPLIED LIFE SCIENCES↗

Porting AMG2013 to Heterogeneous CPU+GPU Nodes

LLNL's future advanced technology system SIERRA will feature heterogeneous compute nodes that consist of IBM PowerV9 CPUs and NVIDIA Volta GPUs. Conceptually, the motivation for such an architecture is quite straightforward: While GPUs are optimized for throughput on massively parallel workloads, CPUs strive to minimize latency for rather sequential operations. Yet, making optimal use of heterogeneous architectures raises new challenges for the development of scalable parallel software, e.g., with respect to work distribution. Porting LLNL's parallel numerical libraries to upcoming heterogeneous CPU+GPU architectures is therefore a critical factor for ensuring LLNL's future success in ful lling its national mission. One of these libraries, called HYPRE, provides parallel solvers and precondi- tioners for large, sparse linear systems of equations. In the context of this intern- ship project, I consider AMG2013 which is a proxy application for major parts of HYPRE that implements a benchmark for setting up and solving di erent systems of linear equations. In the following, I describe in detail how I ported multiple parts of AMG2013 to the GPU (Section 2) and present results for di erent experiments that demonstrate a successful parallel implementation on the heterogeneous ma- chines surface and ray (Section 3). In Section 4, I give guidelines on how my code should be used. Finally, I conclude and give an outlook for future work (Section 5).

97 MATHEMATICS AND COMPUTING↗

First Results From Nb3Sn Coatings of 2.6 GHz Nb SRF Cavities Using DC Cylindrical Magnetron Sputtering System

A DC cylindrical magnetron sputtering system has been commissioned and operated to deposit Nb3Sn onto 2.6 GHz Nb SRF cavities. After optimizing the deposition conditions in a mock-up cavity, Nb-Sn films are deposited first on flat samples by multilayer sequential sputtering of Nb and Sn, and later annealed at 950 °C for 3 hours. X-ray diffraction of the films showed multiple peaks for the Nb3Sn phase and Nb (substrate). No peaks from any Nb- Sn compound other than Nb3Sn were detected. Later three 2.6 GHz Nb SRF cavities are coated with ~1 μm thick Nb3Sn. The first Nb3Sn coated cavity reached close to Eacc = 8 MV/m, demonstrating a quality factor Q0 of 3.2 × 108 at Tbath = 4.4 K and Eacc = 5 MV/m, about a factor of three higher than that of Nb at this temperature. Q0 was close to 1.1 × 109, dominated by the residual resistance, at 2 K and Eacc = 5 MV/m. The Nb3Sn coated cavities demonstrated Tc in the range of 17.9 – 18 K. Here we present the commissioning experience, s ystem optimization, and the first results from the Nb3Sn fabrication on flat samples and SRF cavities.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗