September 2015 - Allocations-admins - mailman3.cels.anl.gov

[LCRC Accounts] Yearly Allocation Request for Cat_Biomass
by accounts＠lcrc.anl.gov 10 Sep '15

10 Sep '15

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Rajeev Surendran Assary Project Name: Cat_Biomass Division: MSD Project title: Computational Studies of Biomass Catalysis Associated funding: BES, EERE Other Systems: CNM Argonne, Blue Gene/P Argonne, NERSC Science: This computational project will be performed to gain fundamental understanding of the chemistry of biomass reactions and help to design catalysts that improve the efficiency of biomass conversion to transportation fuels or industrial chemicals. The project is an integral part of the computational modeling as part of the Computational Biomass Pyrolysis consortium funded by Energy Efficiency and Renewable Energy (EERE). Efficient chemical transformation of biomass is essential to produce sustainable energy and industrial chemicals. Conversion of biomass to useful chemicals include sequence of chemical transformation including, C-O bond cleavage and C-C bond formation. Project description: Main challenges facing the biomass conversion reactions are a plethora of undesired reactions and lack of efficient catalysts for various chemical reactions associated with biomass conversion process. Various reactions include dehydration, rehydration, isomerization, aldol/retro aldol reaction, ring opening/closing, carbon-carbon coupling, hydrogenation, and hydrodeoxygenation reactions. A detailed understanding of the chemistry of both desired and undesired reactions, together with the understanding of catalysis is essential for catalyst design for efficient biomass conversion. Our principle objective here to identify guidelines to search for efficient C-O bond cleaving and C-C bond forming catalysts from first principles to assist experimentalists to discover efficient catalysts. C-C bond formation The computation research involves a detailed investigation of the reaction mechanisms and factors controlling the catalytic gas/liquid phase transformation of primary and secondary products obtained from catalytic fast pyrolysis to distillate chemicals. The primary objective is to discover multifunctional catalysts that enable the conversion of furans and lower molecular weight carbohydrates to diesel fuels. This subtask as part of the consortium would enable us to discover new and improved catalysts essential for the bio-oil upgrading and hence the BETO target of $3/gallon for biofuel by 2022. Computational studies for catalytic materials discovery will be in constant feedback loop with the consortium and Techno Economic Analysis (TEA), and the data produced will be available to the scientific community via Computational Pyrolysis Consortium database, conference presentations, and peer reviewed publications. C-C coupling reactions that we would like to address is Aldol reaction combined with hydro cyclo addition reactions. The reaction sequence here is aldol reaction, hydrogenation in tandem with cyclo-addition reactions. Use first principle-based calculations (density functional theory, ab initio methods) to understand possible reaction pathways for converting furans and low molecular weight carbohydrate molecules (C1 to C4) obtained from biomass pyrolysis to C8-C21 distillate class compounds. Explore catalytic reactions (reaction energetics of individual reactions steps associated with the reaction pathway will be computed) that can increase the carbon chain length by coupling reactions using first principles calculations. This task will be carried out in close collaboration with National Renewable Energy Laboratory due to their extensive experience in vapor phase upgrading and catalyst synthesis. Industry partnership: Project URL: http://cpcbiomass.org/ Current FY Hours Used: undetermined amount New FY Requested allocation: 160000 Q1: 40000 Q2: 40000 Q3: 40000 Q4: 40000 Justification: Gaussian 09/VASP software is scalable up to 64 processes. Here main bottle neck is large number of calculations and long execution times than the parelliization. CPMD and CP2K are high performance software packages, scalable to 64 nodes. Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for disceng
by accounts＠lcrc.anl.gov 10 Sep '15

10 Sep '15

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Justin M Wozniak Project Name: disceng Division: MCS Project title: Discovery Engines Associated funding: ANL LDRD Other Systems: We are using the remnants of PADS for interactive data analysis and visualization. We are also porting APS codes to ALCF resources. Science: The science objectives of this project are to improve data management and analysis techniques for X-ray crystallography and improve experiment analysis turnaround time. This will include data movement improvements as well as novel use of simulation-based analysis. Project description: The initial computational goals are limited: 1) Store/archive data as it comes in from the APS 2) Restore data if data loss occurs on PADS 3) Perform initial studies of scalability of novel DIFFEV/Swift application The programming model used is Swift/T, which executes as an MPI program at run time. The data management will be performed with Globus Online. No known scalability issues exist- Swift/T scales to hundreds of thousands of cores. We plan to experiment with realtime data streaming from experiments at the APS to PADS for visualization using LCRC as a persistent store. We will also experiment with various software integration techniques for DIFFEV/Swift. We have no software requirements above Fusion defaults. No previous runs have been performed- this is very preliminary. The data management operations will include little or no calculation. The DIFFEV/Swift work has the potential to utilize up to 100,000 cores. The existing DIFFEV is known to be very efficient. We expect about 10 project members. No single core runs are planned. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 8000 Q1: 2000 Q2: 2000 Q3: 2000 Q4: 2000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for Nek5000
by accounts＠lcrc.anl.gov 10 Sep '15

10 Sep '15

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Aleksandr Obabko Project Name: Nek5000 Division: MCS Project title: Scalable CFD Algorithms Associated funding: Applied Math Research Program, Nuclear Energy Advanced Modeling and Simulation Other Systems: ALCF Intrepid (BG/P), 128-core Linux cluster, Cosmea Science: Our mission is to develop and deploy scalable algorithms for simulation-based science applications that rely on numerical solution of partial differential equations. Our primary vehicle for algorithm development is the fluid/thermal simulation code, Nek5000, and its off-shoots, NekCEM (for computational electromagnetics) and NekLBM (for lattice Boltzmann methods). There are approximately 30 research groups worldwide currently using Nek5000. We use Nek5000---and the science applications brought to us through our user community---as vehicles for the development and test of new numerical algorithms that include stabilized high-order methods, scalable linear solvers, advanced timestepping methods, and so forth. We emphasize that these algorithms cannot be developed in isolation from applications, as many of the relevant issues do not arise at small scale or in simple model problems devoid of complex physics and/or geometry. Thus, to address algorithmic issues at the forefront of science-based simulation we require a code that is capable of solving problems _at_scale_ and that is sufficiently flexible to allow adaptation to new physics. Project description: Nek5000 is a petascale fluid/thermal simulation code used for a host of applications including reactor thermal hydraulics, combustion, MHD, oceanography, spatiotemporal chaos, and vascular flow modeling. Nek5000 is a grid-based PDE (partial differential equation) solver in three space dimensions and time. The code has long been at the forefront of high-order methods and scalable algorithms. It was one of the first comprehensive software tools developed for distributed memory parallel platforms. It pioneered the development of high-order discretizations that, because of their ability to accurately capture interactions over a broad range of scales, are now deemed an essential ingredient of petascale simulation science and that are being adapted across virtually all disciplines (e.g., fusion, aeronautics, biofluids and geophysics). We have access to resources on BG/Q through INCITE and discretionary allocations. This request is to address mid-range capabilities. Our goal is to use Blues to test the performance of mid-size problems (with 10,000 to 100000) elements with various compilers and compiler options. We also intend to set up a suite of large problems to ensure that various builds are working correctly. We also intend to run production cases for problems in nuclear engineering such as boiling heat transfer (multiphase flows), thermal stripping and conjugate heat-transfer. This year we plan to continue to test various large-element cases which have been run with LES models using 3 different uRANS models that have been implemented in Nek5000. These runs typically need about 100-150K core-hours per case. We aim to study 4-6 such cases with the allocation this year. Industry partnership: None Project URL: http://www.mcs.anl.gov/~fischer/sem1b Current FY Hours Used: undetermined amount New FY Requested allocation: 475000 Q1: 100000 Q2: 125000 Q3: 125000 Q4: 125000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

Re: [allocations-admins] [LCRC Accounts] Project Request: Materials_Genome_SCs
by John Blaas 10 Sep '15

10 Sep '15

Antia is quite anxious to start computations and I have already explained to her that she will need to submit a request for time next year as well. I see no issue in granting the time. - JB On Thu, Sep 10, 2015 at 8:35 AM, accounts(a)lcrc.anl.gov <accounts(a)lcrc.anl.gov> wrote: > 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: Antia Sanchez Botana > Applicant's institution: ANL > Applicant's division: MSD > Project Name: Materials_Genome_SCs > Project title: Materials Genome Approach for Discovery of HTS > Associated funding: Center for Emergent Superconductivity, an Energy > Frontier Research Center funded by the US DOE, Office of Science > Other Systems: > Science: This project uses density functional theory-based methods to predict both the crystal and electronic structures of candidate correlated electron materials and high temperature superconductors. > Our goal will be to provide a survey and database of potential correlated electron materials, with the capability to identify trends, project promising directions and explore competing orders giving rise/enhancing superconductivity. > Project description: We intend to create a framework for designing novel superconductors combining first principles electronic-structure tools, materials databases, and evolutionary algorithms, capable of exploring large configurational spaces. > > In a first stage, this framework will be applied to transition metal nitrochlorides (TNCl (T= Ti, Zr, Hf)) and cuprates. We will use different doping mechanisms and atomic substitutions to design novel compositions. We will screen for local, configurational and thermodynamic stability to isolate new compounds with high probabibility of synthesis. > > We will intensively use DFT codes such as VASP, WIEN2k and Quantum Espresso as well as crystal structure prediction methods that merge ab initio total energy calculations and a specific evolutionary algorithm like USPEX, CALYPSO and MAISE. These have been proven to be efficient for systems with up to 100-200 atoms/cell. Jobs requesting up to 10 nodes will be launched for a typical DFT-based calculation. > > We plan to run intensive calculations of a 1000 different compounds containing up to 150 atoms/cell each. These calculations have proven to be efficiently run in related clusters in NERSC. > > > Industry partnership: > Project URL: > Requested allocation: 100000 > Q1: 0 > Q2: 0 > Q3: 0 > Q4: 100000 > Justification: > Storage requirements: > > > 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 > _______________________________________________ > allocations-admins mailing list > allocations-admins(a)lcrc.anl.gov > https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins

1 0

[LCRC Accounts] Yearly Allocation Request for 1atomCat
by accounts＠lcrc.anl.gov 10 Sep '15

10 Sep '15

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Ujjal Das Project Name: 1atomCat Division: MSD Project title: Supported Single-Site Catalysts for Activation of Small Molecules Associated funding: BES Chemical Sciences Other Systems: CNM Carbon cluster - 100,000 cpu hrs. Science: This proposal requests a renewal of our previous proposal titled "Supported Single-Atom Catalysts for Activation of Small Molecules" but into new directions. Over the last two years, we have primarily focused on understanding the structures of catalyst active sites, exploring various reaction pathways, building structure-function relationships, and used them to screen and predict new materials with improved activity and selectivity. The entire efforts remained within the framework of single atom catalysts consisting of elements from first-row transition (Fe, Co) and post-transition (Zn, Ga) metals. While these catalysts show excellent selectivity for propane dehydrogenation to propylene, the activity is moderate-to-low, especially in comparison with some of the state-of-the-art alkane dehydrogenation catalysts. In the current proposal, we propose to develop supported single atom catalysts consisting of late transition metals (LTM) such as Pd, Pt, Rh etc, as they are known to be highly active for activation of small molecules such as H2 and alkanes. However, these elements are difficult to stabilize as isolated atom on typical oxide support materials (SiO2, Al2O3) because of weak metal-support interaction energy. As a result, they tend to agglomerate and form larger nanoparticles even at a moderate temperature. Since many of our target chemical reactions, such as dehydrogenation of alkane, require high temperature, synthesizing stable single atom LTM catalysts in this condition is a major challenge. Recently, it has been shown that metal-support interaction energy, especially for LTM, depends strongly on the size of the support materials. For example, while a few atom platinum clusters (Pt8) are not stable on typical cerium oxide surfaces, they can be stabilized on ceria particles of a few nanometers. Since, some of our experimental colleagues have strong expertise in growing materials using atomic layer deposition (ALD) technique, we want to explore computationally how layer thickness of support metal oxide affects the metal-support interaction energy with the ultimate goal of selecting right combination of LTM and metal oxide. Project description: To begin with, we will select typical transition metal oxides such as CeO2, TiO2, MnO2 and evaluate interaction energy of single atom LTM such as Pt and Pd with these supports as a function of layer thickness of the support oxides. The diffusion energy barriers of metal atom between different surface sites will be also computed. We will perform density functional theory (DFT) calculations to determine structure, bonding, and reaction energies of the system. Periodic boundary condition (PBC) will be imposed in order to maintain periodicity of the solid support. Gradient corrected PBE density functional will be used for these calculations. The Nudged Elastic Band (NEB) method of Henkelman and coworkers will be used to calculate transition states and activation barriers for formation of different surface species. Surface phonons will be calculated for accurate estimation of reaction free energies. We will primarily use the DFT code implemented in VASP (for periodic calculations) and Gaussian package. Both VASP and Gaussian run in parallel mode. For the MPI version of VASP, increased efficiency is observed up to 32 processors. The Linda version of Gaussian scales up to 32 processors. Energy optimization in VASP takes approximately 750 CPU hours and we plan roughly 200 of such calculations. Hessian (second order energy derivative) evaluation takes roughly 1600 CPU hours and we plan to run 50 of these calculations. Optimization and Hessian evaluation using Gaussian program take approximately the same time for model clusters. We estimate that we will have to run a total of 300 Gaussian calculations, each taking 400 CPU hours. So our total request is 350,000 core hours. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 400000 Q1: 100000 Q2: 100000 Q3: 100000 Q4: 100000 Justification: Storage requirements: Default amount Thank You, The LCRC Accounts System

1 0

New LCRC report submitted for QM_Storage
by support＠lcrc.anl.gov 10 Sep '15

10 Sep '15

Full project report at https://accounts.lcrc.anl.gov/reporting/project/2015/?project=QM_Storage

1 0

[LCRC Accounts] Yearly Allocation Request for dual-fuel-sprays
by accounts＠lcrc.anl.gov 10 Sep '15

10 Sep '15

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Sibendu Som Project Name: dual-fuel-sprays Division: ES Project title: E10 certification fuel spray model development Associated funding: DOE Office of Vehicle technologies, Fiat Chrysler Automobiles LLC. Other Systems: Fiat Chrysler Automobiles LLC. Science: This project involves development of a fuel spray model for use in CONVERGE CFD simulations for in-house use by FCA. The fuel of interest is a Haltermann gasoline certification fuel with 10% ethanol. The project is viewed to have two parts: (1) model development for the hollow-cone sprays, and (2) chemical kinetic mechanism reduction for combustion simulations. Initial emphasis will be placed on completing the injector spray model representing the fuel surrogate for the E10 certification fuel, in order to match fuel spray behavior. A parallel path will be taken for reducing the chemical kinetic mechanism. Details of the two tasks are in the following texts. Project description: A fuel spray model currently implemented in CONVERGE will be adopted to try and match injector spray data for a gasoline certification fuel provided by FCA. Since there is some experimental data for hollow cone injectors available in literature, it can help us with setting up the case. However it is expected that the spray models would need further improvement to match data from FCA, especially for the E10 certification fuel. The first step will be to isolate the uncertainties associated with fuel surrogates with the spray models. Hence, first set of simulations will involve using single or two-component mixtures for which the physical properties are well known. This approach will help identify and subsequently improve the spray sub-models associated with the hollow-cone sprays. With an improved understanding of the fuel spray model, the next step will involve incorporating the physical properties of the multi-component surrogate mixture that will be dev eloped by FCA, to mimic the E10 certification fuel. The performance of these spray and breakup models will be validated and tested for a range of injection pulse widths and backpressures. Experimental data from FCA will consist of a total of 27 conditions with parametric variations on key injection and ambient conditions for the certification fuel surrogates. It is expected that the fuel spray model will need to be developed further in order to properly simulate the physical properties of the E10 fuel. This task will start with calibration of the spray models already available in the software. If there are shortcomings or the existing models do not capture all of the physics to adequately predict the fuel spray behavior (within 5-10% of the ensemble averaged experimental data), then an improved model needs to be developed which would need to account for the full injection and needle dynamics. At this point, FCA would provide further geometrical information about the injector and the deliverables will be modified accordingly. FCA will provide the following data as determined from pressure vessel experiments for both the single/two-component mixture and E10 certification fuel surrogate: Rate of injection, Liquid fuel penetration, Fuel vapor penetration, Fuel vapor distribution, Injector and injection condition details This second task will involve the reduction of a detailed chemical kinetic mechanism for the identified E10 certification fuel surrogates. The detailed mechanism will be developed by FCA through a different contract and provided to Argonne. The detailed mechanism will then be reduced at Argonne with the aim to have a reduced mechanism with about 150 species and 500 reactions. Considering the size of the detailed mechanism, this may be a challenging task. Additional reduction in wall-clock time will be possible by integrating the recent developments made through the VERIFI program at Argonne. Recent developments with VERIFI enable load balancing based on computational time from chemical kinetics. Accuracy of the mechanisms will be judged against fundamental combustion experimental data, such as RCM, shock tubes, or similar, wherever data under relevant gasoline engine conditions is available. The performance of the spray model and chemical kinetic mechanism will also be compar ed to optical and metal engine data, as supplied by FCA. Industry partnership: The project is a work-for-others agreement between Argonne and Fiat Chrysler Group LLC. It is expected that the results of the industry project will not only advance the state-of-the-art in spray simulations but also significantly benefit Chrysler in terms of modifying their best practices for spray simulations. Argonne will be performing the simulations, post-processing results, and writing technical reports. Chrysler will provide the geometry to be modeled, give guidance on simulation results, and their internal needs.It is expected that some of the results will be publishable Project URL: http://www.transportation.anl.gov/engines/multi_dim_model_home.html Current FY Hours Used: undetermined amount New FY Requested allocation: 475000 Q1: 100000 Q2: 125000 Q3: 125000 Q4: 125000 Justification: Storage requirements: 1 TB Thank You, The LCRC Accounts System

1 0

New LCRC report submitted for Cat_Biomass
by support＠lcrc.anl.gov 10 Sep '15

10 Sep '15

Full project report at https://accounts.lcrc.anl.gov/reporting/project/2015/?project=Cat_Biomass

1 0

[LCRC Accounts] Yearly Allocation Request for SHARP-RGG
by accounts＠lcrc.anl.gov 10 Sep '15

10 Sep '15

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Rajeev Jain Project Name: SHARP-RGG Division: MCS Project title: CoreGen and MOAB reading/writing scalability studies Associated funding: NEAMS Other Systems: MCS Science: Multiphysics nuclear reactor core simulations are under active development in several countries. Most of these techniques require large unstructured meshes that can be hard to generate in a standalone desktop computers because of high memory requirements, limited processing power, and other complexities. We have previously reported on a hierarchical lattice-based approach for generating reactor core meshes. In this work we exploit coarse-grained parallelism develop and debug the production of large meshes for reactor assembly and reactor core mesh generation processes. Project description: We have added adaptive mesh refinement to the suite of tools. Some salient aspects of the project are: 1. We were accepted for a poster in SC 2015 based on our results on blues. Title: Scalable Mesh Generation for HPC Applications 2. The complete approach to creating reactor models is now graph-based (C++/MPI), wherein each operation can be a graph-node and this specially aids in exploiting parallelism embedded in the process. 3. We rely on HDF5 and have a tie-up with HDF5 group to study IO bottlenecks in the process. 4. We have run on 4000+ processors 5. It is observed that intel-13 optimized mode gave us best performance, all our codes are installed and working. # project members = 3 Industry partnership: Kitware Inc. (Makers of CMake) they are developing a GUI for RGG tools for designing reactor core geometries. Project URL: http://sigma.mcs.anl.gov Current FY Hours Used: undetermined amount New FY Requested allocation: 20000 Q1: 5000 Q2: 5000 Q3: 5000 Q4: 5000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for MRSA
by accounts＠lcrc.anl.gov 10 Sep '15

10 Sep '15

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Nicholson Collier Project Name: MRSA Division: DIS Project title: Agent-based Modeling of Community-Associated Methicillin-Resistant Staphylococcus Aureus Epidemiology Associated funding: NIH Other Systems: Science: This project is developing an agent-based model of the spread of community-associated methicillin-resistant Staphylococcus aureus throughout Cook County. Each agent represents an individual person in a synthetic population. The scientific objective is to better understand the epidemiology of the disease transmission. Current scenarios simulate 2.9 million people and 1.2 locations throughout Cook County for 10 years on an hourly basis. We plan to extend the simulation to other infectious diseases transmitted by human-to-human contact. Project description: This project is using Repast for High Performance Computing (HPC (http://repast.sourceforge.net) Repast HPC C++ agent-based modeling system that is designed for use on large computing clusters and supercomputers. Several papers have been produced so far: David, Michael Z., Robert S. Daum, Michael J. North, Nicholson Collier, Vanja M. Dukic, Philip Schumm, James A. Evans, Duane T. Wegener, Jocelyn R. Wilder, Charles M. Macal, Diane S. Lauderdale. 2015 (Draft). Estimating the Contribution of Urban Jails to the Spread of Community-Associated MRSA, 2001-2010, using an Agent-Based Model of Chicago. Collier, N.T., J. Ozik, and C.M. Macal. Large-scale ABM with RepastHPC: A Case-study in Parallelizing a Distributed ABM, in PADABS 2015 Proceedings. Vienna, Austria, 24-28 August 2015 (in press). Macal, C.M., M.J. North, N.T. Collier, V.M. Dukic, D.T. Wegner, M.Z. David, R.S. Daum, P. Shumm, J.A. Evans, J.R. Wilder, D.S. Lauderdale, L.G. Miller, and S.J. Eels, “Reducing Transmission of Community-Associated Methicillin-Resistant Staphylococcus aureus: A Dynamic Agent-based Simulation,” Journal of Translational Medicine (2014). Macal et al., et al., “Modeling the Spread of Community-Associated MRSA,” Proceedings of the 2012 Winter Simulation Conference, C. Laroque, J. Himmelspach, R. Pasupathy, O. Rose, and A.M. Uhrmacher, eds., IEEE/ACM, Berlin, FRG (December 2012). Thanks to Fusion, we were able to run the model for the full population of Chicago (2.9 million agents and 1.2 million geographic locations) for the first time. We used this capability to reproduce epidemic curves from a meta-analysis of community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) infection incidence in Chicago. The model results closely match the observed values. We are in the final stages for a journal article exploring "search and destroy" responses to MRSA infections. The existing MRSA model has been updated to simulate these responses. The new allocation will allow us to execute simulations runs with this new code. The results will then be the basis for a new journal article. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 225000 Q1: 75000 Q2: 50000 Q3: 50000 Q4: 50000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0