August 2012 - Allocations-admins - mailman3.cels.anl.gov

[LCRC Accounts] Yearly Allocation Request from CEES-Si
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Maria Chan Project Name: CEES-Si Division: CNM Project title: Investigations of silicon surfaces and interfaces for lithium ion batteries Associated funding: DOE Energy Frontier Research Center -- Center for Electrical Energy Storage (CEES) Other Systems: CNM Carbon Science: This project uses density functional theory (DFT) to investigate silicon as an anode material for lithium ion batteries. Silicon is under intense study as an alternative to carbon-based anode materials due to enhanced safety and superior capacity. The Center for Electrical Energy Storage (CEES), an Argonne-based DOE Energy Frontier Research Center, is heavily invested in the study of Si anodes. The lithiation of crystalline Si results in vast structural changes, in contrast to graphitic electrode materials in which Li is inserted into fixed sites, with limited structural distortions. At room temperature, lithiation transforms crystalline Si into amorphous lithiated silicides LixSi. This renders traditional techniques to model battery materials, based on total energy calculations of crystalline phases, inadequate. In the previous FY, we have developed a new approach to sample configurations of LixSi which is connected to crystalline Si by Li insertion or subsequent removal, which allows us to understand the atomistic nature of the non-equilibrium reaction. Using this newly-developed configurational-sampling technique and Fusion allocation from the previous FY, we have developed an understanding of the drastic anisotropy [1] observed in the lithiation of crystalline Si micro- and nano-structures. In addition, we have also explained the effects of dopants on the lithiation behaviors [2] and investigated the large differences between Si and Ge in terms of anisotropy and dopant effects [3]. Following these successes, we are planning to continue the investigation of Si nanostructures, as well as surface and interfacial chemistries, both of which have been demonstrated experimentally to strongly affect electrochemical performance. The use of nanostructures, such as nanowires, is known experimentally to improve the cycleability of Si, presumably due to effective strain relaxations. The atomistic mechanism of such relaxations, however, is not well known. From our previous studies, we have preliminary findings on the expansion of Si nanowires upon lithiation, but found a profound effect of the lack of surface passivation on the lithiation process. Since then we have established proper surface passivation procedures and are poised for a more systematic study of various nanowire orientations and sizes. Nanoparticles are also heavily used in experimental studies and the effects of sizes on lithiation behavior is also of interest. In addition, surface and interfacial chemistries play a crucial role in the cycling performance of Si. We have identified several organic surface species that are of relevance, and would also like to investigate the effect of products of surface-electrolyte interphases, such as fluorides and carbonates, on the lithium insertion and removal mechanism. [1] M. K. Y. Chan, C. Wolverton, and J. Greeley, “First Principles Simulations of the Electrochemical Lithiation and Delithiation of Faceted Crystalline Silicon," Journal of the American Chemical Society, accepted 2012. [2] B. Long, M. K. Y. Chan, J. Greeley and A. Gewirth, “Dopant Modulated Li Insertion in Si for Battery Anodes,” Journal of Physical Chemistry C, doi:10.1021/jp2060602, (2011). [3] M. K. Y. Chan, B. Long, A. Gewirth and J. Greeley, “The first-cycle electrochemical lithiation of crystalline Ge and Si – a comparison,” in preparation. Project description: Lithiation simulations by configurational sampling – total-energy DFT calculations For (a) Si nanowires of [100], [110] and [111] orientations and diameters 1-2 nm, (b) Si nanoparticles of diameters 1-2 nm, (c) Si (111) surfaces with organic passivating species CH3, OCH3 and SCH3, and (d) interfaces of Si with solid-electrolyte interphase products LiF and Li2CO3, we will carry out lithiation simulation calculations using our newly-developed configurational sampling algorithm. This involves total-energy DFT calculations and relaxations of a series of configurations with Li inserted into interstitial sites, with the lowest-energy relaxed configuration serving as a starting point for next Li insertion. The computational costs per sampling calculation are: (a) 20-80 core-hours, (b) 40-100 core-hours, (c) 10-20 core-hours, and (d) 20-40 core-hours. The number of calculations involved is 200-2000 for each starting configuration. The total core-hours requested for the above categories are: (a) 200,000, (b) 100,000, (c) 20,000 and (d) 80,000 core-hours. Using the projector-augmented wave method, only valence electrons will be included which reduces computational costs. We will use well-established DFT codes VASP, which use a plane wave basis set. For large systems such as nanostructures, VASP has been tested on Fusion for computation involving systems with up to 7000 electrons on up to 512 cores. For routine surface calculations, hard scaling efficiency of 70%-80% has been demonstrated for 64 to 128 cores compared to 8 cores, thus enabling efficient calculations of large numbers of configurations. Please note that this renewal proposal consists of some work in the previous proposal that has been delayed due to a change in position and responsibilities as well as medical leave. Project URL: http://www.anl.gov/energy-storage-science/ Current FY Hours Used: undetermined amount New FY Requested allocation: 400000 Q1: 100000 Q2: 100000 Q3: 100000 Q4: 100000 Justification: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request from Fathom
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Timothy J. Tautges Project Name: Fathom Division: MCS Project title: Fathom performance engineering Associated funding: DOE/NE, DOE/ASCR, DOE/BER Other Systems: (making proposals to Intrepid, Surveyor) Science: There are two primary areas of science goals for this project. For nuclear reactor simulation, we seek to construct very large mesh models, in the range of 50M-200M elements. These meshes will represent whole-core reactor models, and will support both thermal-fluid (Nek) and neutron transport (UNIC) simulations. The specific science goals for these codes include: resolving mixing behavior in both thermal and fast reactor cores evaluating the accuracy versus computational time for coupled thermal fluid / neutronics computations evaluating the accuracy of heterogeneous material neutron transport modeling, compared to lumped parameter models used in other codes support for various OECD benchmark participation, including the Vettenfahl T-junction and the Mitas benchmarks For climate modeling, we will continue developing the ParCAL library, based on MOAB; this effort seeks to parallelize the NCL climate data analysis tool. Our efforts this year will be to continue filling out the capabilities in the original NCL tool which are performed in ParCAL/MOAB, and demonstrating these capabilities on real data analysis efforts. Project description: MOAB is a library for representing, querying, and modifying mesh and associated data. MOAB supports the finite element zoo, plus polygons and polyhedra, and, more recently, includes an interface for structured mesh. MOAB uses an efficient array-based storage model, allowing applications direct access to memory storage for field data and, soon, to connectivity and coordinate data. MOAB uses an MPI model for parallel mesh, where mesh models appear locally as a serial mesh, with additional data available about parallel constructs such as inter-processor mesh interfaces and ghost layers. The primary means for accessing MOAB's parallel representation are to read/initialize and write mesh from/to a file. MOAB uses parallel HDF5 for parallel I/O of unstructured meshes, and parallel netcdf for structured mesh parallel I/O. Substantial work has been done to optimize this file I/O, as well as the resolution of shared inter-processor interface mesh and ghost layer exchange. Parallel I/O has been demonstrated on unstructured hex and tet meshes up to 64M elements, on up to 64k processors of Intrepid. Various performance issues are still being addressed, including enhancements to MPI-IO for single-file parallel I/O (in cooperation with Rob Latham at ANL). MOAB targets a variety of applications. For nuclear energy applications, the Nek CFD code and the UNIC neutronics code have been modified to read mesh from MOAB. Nek has been modified to export its solution variables through MOAB as well. The Nuclear Energy co-design project CESAR plans to make extensive use of MOAB, for mesh-based interactions as well as solution transfer between meshes. MOAB supports several climate applications, including ice sheet modeling and climate data visualization and analysis. MOAB will also be involved in at least one SciDAC application project, involving fusion edge modeling. For reactor simulation, we have also developed the Reactor Geometry (and mesh) Generator (RGG). This tool begins in the "assygen" stage by generating geometric and mesh models for a number of assembly types, then, in the "coregen" stage, copy/moves/merges these assemblies into an overall core lattice, with some assemblies copied many times. We have demonstrated good parallel speedups of the coregen stage of RGG, in some cases obtaining super-linear speedups due to decreased swapping and thrashing of virtual memory. This year's technical work will focus on: - testing and analyzing parallel I/O performance reading reactor simulation and climate data - testing and analyzing performance of parallel mesh-to-mesh transfer functionality in MOAB - performing parallel coupled multi-physics simulations for the NEAMS program Project URL: http://trac.mcs.anl.gov/projects/ITAPS/wiki/MOAB Current FY Hours Used: undetermined amount New FY Requested allocation: 100000 Q1: 25000 Q2: 25000 Q3: 25000 Q4: 25000 Justification: MOAB has demonstrated parallel I/O out to 1024 cores on fusion, and 64k cores on Intrepid. For example, MOAB can read a structured mesh and a single solution variable, for a mesh of approximately 160M cells, in about ten seconds on 1024 processors, and less than 3 seconds on 64 processors. For unstructured meshes, MOAB reads meshes up to 64M elements in less than a hundred seconds on Intrepid. Execution times in both these cases seems to level off above about 128 processors, and we are working with Rob Latham of ALCF to understand why this happens. RGG demonstrates good scaling out to the total number of assemblies in a given core model, usually 100-300 for whole-core models. In the case of RGG, we have demonstrated super-linear speedups in some cases, due to fitting the entire model in memory that eliminates the need for virtual memory swapping (and thrashing, for large models). Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Request: Quantum_catalysis
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A new project on the LCRC cluster has been requested. Please forward the information on to the LCRC Allocation sub-committee. Applicant's name: Maksym Kryvohuz Applicant's institution: ANL Applicant's division: CSE Project Name: Quantum_catalysis Project title: Quantum effects in enzyme catalysis Associated funding: Director’s Fellowship Other Systems: None Science: The objective of this project is to perform modeling of hydrogen transfer step in several enzyme-catalyzed reactions, in which significant nuclear quantum effects were experimentally observed. The origin of high efficiency of enzymes as catalysts remains unclear, yet fundamental understanding of the mechanisms underlying the function of enzymes would assist in the design of bio-inspired synthetic catalysts. It has been suggested that nuclear quantum effects such as tunneling of hydrogen through the reaction barrier can contribute to the efficiency of enzymatic catalysis, yet its role in catalysis is unclear. In this project, a multi-dimensional quantum modeling of the rate-limiting proton transfer step catalyzed by methylamine dehydrogenase (MADH) will be performed and its rate as well as primary kinetic isotope effect will be calculated. It has been estimated that 99% of this reaction occurs by tunneling at physiological temperatures. The role of dy namical effects of enzyme vibrations on the proton-transfer step as well as nuclear quantum effects such as multidimensional tunneling will be analyzed at the ab initio level. The results of this project will provide novel understanding of the role of quantum effects in enzyme catalysis and suggest ways to improve catalytic efficiency of bio-inspired synthetic catalysts. The proposed project supports Argonne’s M2D2 initiative of synthesis by design as well as on development of biomimetic catalysts for efficient conversion of biomass into fuels. Project description: Quantum effects of H-transfer step in enzyme catalyzed reactions will be studied via the instanton reaction rate theory [J.Chem.Phys. v.134, p.114103 (2011)] which allows one to rigorously incorporate nuclear quantum effects in multiple dimensions with on-the-fly computations of accurate electronic reactive potential energy surface (PES). An efficient code for parallel computations of reaction rates and PES has been developed and successfully applied to describe quantum kinetics of H-transfer reactions in 13 and 23-atomic systems. In the present project H-transfer reactions in enzymes will be studied using the same technique. For this purpose, from 40 to 80 atoms of the active site of MADH enzyme will be treated quantum mechanically with electronic structure calculations of reactive PES performed in Gaussian 09 package. The remaining 7250-7200 atoms of MADH as well as 3000-4000 atoms of solvent will be treated classically with molecular mechanics for ce field. The numerical algorithm is based on determination of tunneling trajectories under the multidimensional reaction barrier by optimization. The tunneling trajectory is divided in 16 points and its optimal shape is determined by variational principle together with on-the-fly calculations of PES at these 16 points. One round of optimization requires about 10 iterations, which results in the total number of 16*10=160 single-point electronic structure calculations (with frequency analysis, i.e., computation of matrix of second derivatives which is essential for the process of optimization) to find an optimized tunneling path at a single temperature. Tunneling trajectories at several temperatures (from 5 to 6 different values) need to be determined to obtain Arrhenius plot, which requires (160 configurations)*(5 temperatures)= 800-1000 single point electronic structure calculations for a system of 40-80 atoms. Averaging over about 40 configurations of enzyme/solvent enviro nment is necessary to get an average value of a canonical reaction rate constant. Thus (1000 single point calculations)*(40 environment configurations) = 40000 single-point computations (with frequency analyses) constitute the minimum number of computations required for one enzyme system. DFT mPW1K or B3LYP methods will be used for on-the-fly electronic structure calculations of PES. Numerical tests shown below indicate that an organic system of 45 atoms requires about 24 core-hours for a single job with frequency analysis. The estimated number of computer time for the project is therefore (24 core-hours)*(40000 computations) = 960,000 core-hours. Number of people to run the above simulations: 1. Project URL: Requested allocation: 960000 Q1: 240000 Q2: 240000 Q3: 240000 Q4: 240000 Justification: A test of performance of B3LYP and mPW1K methods with 6-31+G(d,p) basis for a frequency analysis of a fixed geometry 45-atom organic molecule (PAF-1 monomer), which contains 20 atoms of H and 25 atoms of C was performed with Gaussian 09 code using 8, 16, 24, and 32 cores. The performance results are summarized in the table below. Substantial speed-up is achieved with the increase in the number of cores. Nodes x Cores Wall Time (min) B3LYP mPW1K 1 x 8 159 172 2 x 8 82 91 3 x 8 58 62 4 x 8 45 49 The requester has used undetermined amount hours of their initial startup project. In addition to approving an initial amount, please specify a Category and Subcategory for this project. For a list of the current selection of approved categories, please see: https://wiki.lcrc.anl.gov/wiki/Processes/Categories Once the Allocation committee has approved the project, please go to the Project Management page to create it: https://accounts.lcrc.anl.gov/projects.php Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request from dft_biomass
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Bin Liu Project Name: dft_biomass Division: CNM Project title: DFT studies of catalytic biomass conversion Associated funding: IACT (EFRC) Other Systems: Carbon cluster (Center for Nanoscale Materials) Science: Periodic density functional theory (DFT) calculations are performed to understand the fundamental mechanisms involving the catalytic transformation reactions of biomass-derived compounds, such as, glycerol decomposition for hydrogen fuel productions, or furfural hydrogenations to furfuryl alcohol, on metal surfaces. These are some of the major routes for fuel and chemical productions using renewable biomass sources. However, the complications due to side competing pathways associated present grand activity and selectivity challenges for general technological development. Our DFT studies have delved into detailed analyses of the mechanisms governing the respective reaction networks on different single metals. Research efforts are also engaged in developing and applying robust correlation relationships to facilitate an efficient evaluation of the complex reaction networks. Bimetallic alloys that may possess much improved reforming capabilities than any of its individu al components. Some of the alloy catalysts have been tested experimentally, the other alloys have only been proposed in theory. Overall, the knowledge regarding the roles of alloying and associated properties is scarce. Using the DFT-based computational strategies established in this study, we hope to be able to reveal the mechanism leading to the enhanced properties and help future catalyst designs. Project description: As part of the IACT (short for the Institute for Atom-efficient Chemical Transformations) supported by Energy Frontier Research Center at Argonne, the DFT calculations are performed on a number of computational projects related to catalytic biomass processing using the Vienna ab initio Simulation Package (VASP). In the previous two years, FY2011-12, we have extensively examined the thermodynamics and kinetics regarding glycerol decomposition pathways via dehydrogenation, C-C and C-O bond scissions on a number of transition metals, including Pt, Pd, Rh, Ni and Cu [1-3]. In FY2013, we plan to extend our analysis to bimetallic alloys has shown reforming activities and selectivities several factors higher than the corresponding single component. Similar to our approach to glycerol decomposition analysis, we evaluated a number of transition metals to elucidate the fundamental trend for the key elementary steps using the allocations from FY2011-2012. An analogous scaling correlation has been developed to study the hydrogenation intermediates of ring compounds. Experimental studies indicate that the hydrogenation of furfural is contended strongly by other reaction types, such as, C-C bond scission, C-O bond scission, or ring opening. In addition, the reaction routes are strongly influenced by the metal catalysts. Using the thermodynamics obtained from a combination of explicit DFT calculations and scaling correlations, and coupled with the BEP relationships, a quick evaluation of the overall furfural transformation can be achieved. In FY2013, one of the research focuses will continue to develop the correlation relationships by performing DFT calculations. Another major goal is to identify the common ground of various tr ansformation processes of the biomass model compounds, so that we would have a unified perspective on these complex reaction networks. To complete the proposed tasks, the necessary computation will approximately involve 60% of optimizations, and 40% of transition state search (primarily using NEB, combined with the Dimer algorithm). A typical optimization job uses 8 nodes (64 processor), and the job convergence approximately takes 24 hours. For NEB calculations, usually 10 or 14 nodes are used, depending on the number of interpolated images. Each NEB calculation typically takes 40~50 hours, plus 100 CPU hours for additional analysis, such as, vibrational frequency calculations. For FY2013, Dr. Bin Liu of CNM will still be the primary investigator of this project with Dr. Jeff Greeley as the co-investigator. To complete the above work, we are requesting 400,000 core hours for FY2013, with 100,000 for each quarter. This is an estimate based on our preliminary work running on Fusion in the previous FYs. We anticipate that our need for Fusion resources will be relatively steady in the coming FY. Project URL: http://www.anl.gov/catalysis-science/ Current FY Hours Used: undetermined amount New FY Requested allocation: 400000 Q1: 100000 Q2: 100000 Q3: 100000 Q4: 100000 Justification: Recent improvements in the parallelization of VASP combined with the improved hardware capabilities (e.g. the expanded memory) allow for reasonably efficient near-linear scaling for jobs running on up to 80 cores (10 nodes). Typically, a DFT job uses 4~8 nodes at a timescale of 48 hours. Future usage will require larger number of nodes (8~10 nodes) due to a shift of our focus on transition state calculations. As the computation time for a single job of such calculation can be a week or even more. Large allocation is preferred to guarantee the completion of this investigation. In the mean time, we will efficiently use the requested allocation by carefully selecting the key reactions to produce outstanding science. We hope to continue to use Fusion cluster to satisfy our scientific interest. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request from WRF-CHEM-GVAX
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Rao Kotamarthi Project Name: WRF-CHEM-GVAX Division: EVS Project title: Weather and Aerosol forecast products for the Ganges Valley Aerosol Project Associated funding: DOE-BER Other Systems: NERSC Science: The Ganges valley region is one of the largest and most rapidly developing sections of the Indian subcontinent. Impacts of changes in precipitation patterns, temperature, and the flow of the snow-fed rivers could be immense. Recent satellite-based measurements have indicated that the upper Ganges valley has some of the highest persistently observed aerosol optical depth values. The aerosol layer covers a vast region, extending across the Indo-Gangetic Plain to the Bay of Bengal during the winter and early spring of each year. This is one of the few regions showing a trend toward increasing surface dimming and enhanced mid-tropospheric warming. The consequences of aerosols and associated pollution for surface insolation over the Ganges valley and monsoons in particular are not well understood. Our proposed field study will provide critical data to address these hypotheses and will contribute to developing better parameterizations for tropical clouds, convection, an d aerosol-cloud interactions. The primary science questions for the mission are as follows: (a) What are the sources and sinks of the aerosols comprising the aerosol cloud over this region and what effects does it have on clear-sky radiation fields, as well as on cloud micro- and macrophysical properties? What effect does the diurnal PBL cycle in the Ganges valley have on the atmospheric distribution of aerosols? (b) What effect do increasing aerosols in the Ganges Valley have on the ISM? What effect do aerosols have on shortwave radiative forcing, mid-tropospheric heating and convective activity over continental India? To achieve these science objectives an implementation has been devised in collaboration with scientists in India. The primary anchor facility for the project is the ARM mobile facility. A deployment starting in approximately June 2011 and lasting to April 2012 at the ARIES observatory located in Nainital is anticipated for AMF ground operations. During this one-year deployment of the AMF there will be an intensive periods, extending from February 1st, 2012 to April 1st 2012 (winter intensive). Project description: The Ganges Valley Aerosol eXperiment (GVAX), led by Dr. Rao Kotamarthi, will begin its one-year ground-based observation with the Atmospheric Radiation Monitoring (ARM) mobile facility at Nainital, India,starting June, 2011. This major field study involving multiple Department of Energy laboratories will focus on investigating the potential impact of aerosols on regional scale radiative forcing changes over the Ganges Valley. Measurements of radiative flux, cloud, convection, and aerosol physical, chemical, and optical properties will be made to develop core data sets for scientific analysis. The measurements will range from those made at surface locations that yield point and profile data sets to a few selected aircraft platform-based measurements. However, these measurements and the subsequent data sets must be placed in a larger regional context in order to relate the data to a particular dynamic state of the atmosphere and begin developing hypothe ses that explain observed phenomena. The regional context will be developed within the conditions generated by a 3-D regional scale climate model. Baseline simulations of 3-D regional scale model will provide the necessary analytical support for many of the scientific studies done under GVAX. In addition, the 3-D model will be a test bed for high performance computing data storage and visualization. We have implemented and run the latest 3.3 beta version of the Weather Research and Forecasting (WRF) model with an online capability of simulating chemistry and aerosols (WRF-Chem) on the Argonne Fusion clusters. The WRF-Chem model has been optimized and uses domain decomposition to divide total work load on the distributed-memory systems. Without chemistry and aerosol modules, the standard WRF CONUS benchmark (48 hr and 12 km forecast over Continental US) completes in 15 minutes on a teraflop system (~250 processors). For the GVAX campaign, the principal investigator will simulate a model domain from 55E to 95E and 0 to 36N, divided into 367x331 12-km grid cells with 28 vertical layers from the surface to the top of the atmosphere. The model boundary conditions will be updated on a daily basis from the gridded NOAA/NCEP global forecasts, which are available online within about 36 hrs to 48 hrs after real time. Therefore, there is a need to perform at least 72 hr model simulations in order to provide 36 hr forecasts required for the field experiment. The test runs on the WRF-Chem model have been completed on a number of nodes. The regional model scales up to 250 cores and needs ~10 hours to complete a 72-hr forecast. The speedup achieved with 72 cores is about 13 relative to 4 cores. Memory used on each node decreases from 4.5 GB, 1.5GB, 1.1 GB, to 0.9 GB while the number of nodes increases from 4, 16, 36, to 72 cores. The ideal computing time to finish daily forecasts is around 3~5 hours. Thus, the daily forecasts is expected to require about 200-250 cores. Because of gas-phase and aerosol chemistry for pollutant forecast output, the WRF-Chem is much more expensive in computation time compared to running the WRF model only. However, forecasts of the geographic distribution and transport pattern of atmospheric gases and aerosols are required for the GVAX campaign. The above estimated time does not include pre-processing of the GFS meteorology fields, post-processing and visualization of the forecast results. In addition t o the model runs, these processes will be done in serial and cost an additional several hours on Fusion. We have setup a Live Access Server (LAS) node at Argonne to serve the model results to the user community. The site (atmos.anl.gov:8080/las) is now operational and we plan to regularly update the website with model results on a daily basis 3-D regional scale model simulations will provide the necessary analytical support for many of the scientific studies done under GVAX. These modeling results of air pollution and meteorological fields will become one of the central ‘data’ products for the experiments archived at the ARM websites. The expectation is that there will be a significant number of scientists from around the world who will be interested in GVAX data. In addition, the 3-D model will be a test bed for high performance computing data storage and visualization. The large volumes of model outputs generated would be used to test or assess the feasibility of using the cloud computing systems (such as Magellan) to perform/extract a pre-determined set of forecast products as and when needed by scientists distributed at disparate locations world-wide. The project this year will use the resources to perform diagnostic simulations for the year 2011-2012 and explore the data and model comparison issues arising from the GVAX project. Project URL: http://www.arm.gov/sites/amf/pgh/ Current FY Hours Used: undetermined amount New FY Requested allocation: 400000 Q1: 100000 Q2: 100000 Q3: 100000 Q4: 100000 Justification: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request from Atmos_Chem
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Rao Kotamarthi Project Name: Atmos_Chem Division: EVS Project title: Process Scale Modeling of Atmospheric Chemistry and Transport Associated funding: DOE,DoD Other Systems: NERSC Science: The scientific aim of the proposed modeling activity is to further the understanding of the chemistry and distribution of aerosols over certain regional domains and calculate the heating rates as a result of increased aerosol loading primarily over India. A second objective is to use Ensemble Adjusted Kalman Filter technique for data assimilation of satellite data sets into global chemical transport models to estimate model biases in calculating trace atmospheric gas constituents and aerosols A third objective is to perform dynamic downscaling of climate model simulation results for regional scale applications. In particular, we will focus on the uncertainties in downscaled projections resulting from boundary conditions, the utility of spectral nudging and grid scale nudging for improving the energy balance of regional scale climate models over long integration times. Project description: The models employed for these studies include (a) WRF-CHEM (b) NRCM and (3) GEOS-CHEM. Each of these models has been ported to the fusion luster and uses to perform calculations by our group. All the codes use a mix of MPI and OpenMP programming models and can scale for as few as 4 nodes to several hundred. The GEOS-CHEM model doesn't scale beyond 32 nodes due to gridding constraints. The NRCM (Nested Regional Scale Climate Model) and the WRF-CHEM models are highly scalable and limited by job turnover times and queue limitations. The project will have approximately 4 members. Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 475000 Q1: 100000 Q2: 125000 Q3: 150000 Q4: 100000 Justification: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request from lcrc-app-engr
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: John Valdes Project Name: lcrc-app-engr Division: LCRC Project title: LCRC Application Engineer support work Associated funding: Other Systems: Science: Project description: This project will be used by the LCRC application engineers to do support work for users of the LCRC cluster. This support might be in the form of installation and debugging of both LCRC core and user applications, benchmarking and performance analysis, user startup help, etc. Project URL: http://www.lcrc.anl.gov Current FY Hours Used: undetermined amount New FY Requested allocation: 24000 Q1: 6000 Q2: 6000 Q3: 6000 Q4: 6000 Justification: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request from support
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: John Valdes Project Name: support Division: LCRC Project title: Systems administration tasks on the LCRC Cluster(s). Associated funding: Other Systems: Science: Project description: This project will be used by the support team for systems tasks. This will include debugging problems, testing upgrades and modifications, and running scaling and performance tests for diverse areas such as scheduling algorithms, networks, filesystem upgrades, application installations, system management, etc. Project URL: http://www.lcrc.anl.gov Current FY Hours Used: undetermined amount New FY Requested allocation: 8000 Q1: 2000 Q2: 2000 Q3: 2000 Q4: 2000 Justification: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request from SHARP_TH
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Elia Merzari Project Name: SHARP_TH Division: NE Project title: SHARP Thermal-Hydraulics Development Associated funding: DOE-Nuclear Energy Other Systems: Intrepid. NE clusters. Science: To validate, develop and extend the thermal-hydraulics capabilities of the multi-physics nuclear reactor analysis platform package SHARP. This project comprises several activities in the SHARP thermal-hydraulics development and validation efforts in support of the Nuclear Energy Advanced Modeling and Simulation (NEAMS) program under DOE-NE: Item 1) Simulation of the SIBERIA experiment performed in Russia and part of a collaborative benchmark to test all shear stress predictions in case of partial blockage of the channels. The aim is to validate RANS models using STAR-CCCM+ using both LES data generated with NEK and available experimental data. Item 2) Using the SHARP toolset support activities related to SHARP in current NEUPs. In particular, the modeling of experiments related to nuclear fuel bundles responses to earthquake, currently being performed at George Washington University. This exercise is particularly important because it will help validated the SHARP toolset in a complex experiment. NEK (high fidelity CFD code) will be coupled through MOAB with DIABLO (a FE stress analysis code). Item 3) This project will support a potentially funded NEUP related to design of completely passive spent fuel pools. Both issues are key in addressing the issues raised by a Fukushima type scenario. Item 4) Using the commercial CFD code STAR-CCM+ we plan to study in further detail the Argonne MAX experiment. In previous years, preliminary calculations for the MAX experiment were performed. A data set has now been produced and validation studies on our tools will be performed. Project description: Item 1, (STAR-CCM+ simulation of “SIBERIA”) STAR-CCM+ employs the finite volume formulation of the Reynolds-Averaged Navier-Stokes (RANS) method to determine the local three-dimensional velocity distribution in the channel. The results are compared with available experimental data for Validation. Item 2, (Simulation of the George Washington University experiment) This experiment present complex fluid structure interaction (FSI) between fluid and water induced by a shaker table. A wide variety of options will be tested for FSI (i.e., NEK coupling with DIABLO). This calculations will be typically large and involve transient analysis. To simulate highly unstable turbulent flows the solver utilizes an LES approach, where the filtered equations of motion are solved. All scales smaller that the filter size, are modeled using a dynamic subgrid scale (SGS) model. The stress analysis and the corresponding moving boundaries and deformations will be handled by the coupling with DIABLO through MOAB. Item 3 (Support of Purdue NEUP) As part of this activity, several design iterations will be performed to design a reliable and. completely passive (i.e., operating in natural convection) spent fuel pool. Each iterations will be studied using the CFD code STAR-CCM+ in RANS mode. A sensitivity analysis on modeling options will be performed. Item 4, (MAX experiment support) Using insight from previous work steady state and full transient URANS calculations will be performed using the inlet data now available. Results will be compared with available experimental data and used to determine and guide the future experimental matrix and possibly an international benchmark. Access to the Fusion cluster would be crucial in performing all the proposed work. The STAR-CCM+ code and the Nek5000 code have already been installed on Fusion. Compilers, such as the Intel Fortran Compiler necessary to compile users subroutines are already installed. All jobs run on Fusion will be run in parallel. All calculations proposed require large computational resources due to the size of the problem. Two to Four users are expected for this project. Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 814 Q1: 114 Q2: 200 Q3: 400 Q4: 100 Justification: At Argonne and elsewhere, the CFD codes STAR-CCM+ is routinely applied to simulations that use up to thousands of cores, and good parallel performance has been observed. Optimal performance is usually observed in steady state analysis is observed usually with 50,000 - 100,000 meshes per core. Nek5000, has proved scalability even below the ranges proposed. The typical size of the problems proposed are the following: Item 1) SIBERIA experiment : 10,000,000 meshes in steady state. This would require typically 128 cores to be performed. Considering that a problem would likely converge within 24h of continuous calculation and that we would like to perform at least 20 production runs with 3,000 h of development this would translate in approximately 64,000 core-hours. Item 2) GW experiment : Fusion will be used here for test runs rather than production runs. These will involve several (around 30) runs with 64 cores for 48 hours. For a total of approximately 100,000 hours. Item 3) Support of the Purdue NEUP: 100,000,000 for the largest cases in steady state. A large number of design testing through CFD will be performed. These will involve approximately 15 runs (only final production runs will be performed here) on 32 nodes for 24 hours. Costing approximately 400,000 including development time. Item 4) MAX simulation: 5,000,000 meshes. This would require typically 64 cores to be performed. Long transient are likely to be necessary with 200h of consecutive calculations. Considering that we would like to perform at least 20 production transient runs, with 3,000 h of development this would translate in approximately 250,000 core-hours. The uneven distribution of the request is due to the likelihodd that item n.3 work will be concentrated in the 3rd quarter due to testing of several design options. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request from Metastable_Materials
by accounts＠lcrc.anl.gov 28 Aug '12

28 Aug '12

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Peter Zapol Project Name: Metastable_Materials Division: MSD Project title: First-principles studies of growth mechanisms for metastable materials Associated funding: ANL Strategic LDRD, DOE Other Systems: NERSC, pnnl-Chinook Science: The main objective of the project is to develop first-principles models describing the growth of Group-III nitrides in order to establish a firm basis for far-from-equilibrium nitride synthesis in a wide range of compositions. In particular, the growth of InxGa1-xN alloys will be extensively studied in this work because of their widespread applications in solid-state lighting, optoelectronics, and photovoltaics. Although metal organic chemical vapor deposition (MOCVD) is the most efficient and clean deposition process for production of nitride solid-state devices, the growth process for InxGa1-xN via this technique has been observed to display a complex behavior characterized by self-sustaining spatiotemporal chemical oscillations between nanocrystalline islands of InN and elemental In droplets. While these results suggest that the nitrogen activity produced by surface-catalyzed ammonia (NH3) decomposition may be responsible for the observed unusual behavior, an at omistic model of the gas phase and surface chemistries, and catalytic reactions that govern such behavior needs to be developed. We propose calculations aimed at development of a kinetic model, which will help optimize synthesis conditions to grow InxGa1-xN materials of much higher quality for solid-state lighting. The parameters of the model describing oscillatory behavior of InN islands/In droplets during the growth will be determined from the results. Identification of rate-limiting reaction steps will lead to optimization of synthesis conditions and precursors based on this model. The results of the modeling will be validated by comparison to the experimental data obtained in ANL. This model is expected to be applicable to describe kinetically controlled processes in synthesis of other materials. Project description: We will continue our systematic study of the non-equilibrium growth of Group-III nitrides with NH3, In(CH3)3, and Ga(CH3)3 as N, In, and Ga precursors, respectively. So far, we have investigated the interaction of NH3 and its moieties with the (0001) surfaces of InN and GaN, focusing on their adsorption energies and decomposition barriers. We have also included in our modeling In and Ga from In(CH3)3 and Ga(CH3)3 onto the nitride surfaces. We will extend this work to other surface orientations, in particular to miscut surfaces. First, gas phase decomposition rates will be calculated for these compounds to establish the precursor composition interacting with the surfaces at relevant conditions. The species will be then placed on the surfaces along with appropriate NHx (x=0-3) precursors and the reaction energies will be calculated. Decomposition barriers for NH3 will be also computed on the nitride surfaces in the presence of the In and Ga species a nd compared to our earlier calculations on bare surfaces. Surface step edges and kinks will be considered in our modeling as well. These are of particular importance because of ongoing experiments to elucidate growth processes on GaN substrates with varying miscuts. Additionally, we are planning to calculate kinetic barriers for NH3 decomposition on strained surfaces since the nitrides are currently grown on substrates with a large lattice mismatch by our experimental collaborators. Also, vibrational frequency calculations will be performed for the lowest energy surfaces in the presence of NHx (x=0-3) and will be compared to the ongoing experimental measurements. The predicted precursor decomposition barriers will be used to develop a kinetic model aiming at optimizing synthesis conditions to grow nitride materials of much higher quality. All calculations will be performed with Vienna Ab-initio Simulation Package (VASP), using plane wave basis sets and PAW potentials. The kinetic barriers for NH3 on the nitride surfaces will be computed using the Nudged Elastic Band Method. The scale of the proposed calculations exceeds small computer cluster capacity. Therefore, we are requesting the LCRC resources to perform the outlined calculations. The speed and reliability of the Fusion cluster will allow us to accomplish the work in a timely manner. The calculations of the InN and GaN miscut surfaces in the presence of N and In (Ga) precursors will take approximately 96 hours each when ran on 48 processors. At least 50 different structures will be calculated. We are planning to run about 25 of them. The calculations of the reaction energy barriers (NEB), on the other hand, will require about 192 hours each when ran on 48 processors based on our previous work. At least 60 reaction barriers for NH3 decomposition (i) in the presence of In (Ga) precursors, (ii) on strained surfaces and (iii) miscut surfaces will be performed for InN and GaN surfaces. Therefore, a total allocation of 700,000 processor hours will be needed to perform the proposed calculations. Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 700000 Q1: 150000 Q2: 150000 Q3: 200000 Q4: 200000 Justification: This work pursues development of models to describe kinetically controlled growth of group-III nitrides based on first-principles approaches. The project is part of Argonne strategic Materials for Energy initiative. Three different surfaces (i.e. the (0001) and (101-0) surfaces of InN, GaN) will be investigated in detail in the presence of MOCVD precursors. Since VASP does not scale very well beyond 48 processors for our systems, it precludes the use of short-time large-scale capabilities of leadership class computer facilities. In fact, the scaling of the calculations is dramatically reduced with the number of processors beyond 48. The use of Fusion cluster will allow us to overcome the computational challenges posed by the significantly large system sizes necessary for modeling various surfaces as well as decomposition and diffusion processes, which are computationally very demanding. The detailed estimate of the required computer time is given at the e nd of the project description. Thank You, The LCRC Accounts System

1 0