August 2017 - Allocations-admins - mailman3.cels.anl.gov

New LCRC report submitted for NuclearStructure
by support＠lcrc.anl.gov 24 Aug '17

24 Aug '17

Full project report at https://accounts.lcrc.anl.gov/reporting/project/2017/?project=NuclearStruct…

1 0

[LCRC Accounts] Yearly Allocation Request for Nek5000
by accounts＠lcrc.anl.gov 24 Aug '17

24 Aug '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Shashi Aithal 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. 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 LCRC machines to test the performance of mid-size problems (with 10,000 to 100000) elements with various compilers and compiler options. We also intend to use the time 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. Additionally, we also intend to study some oscillatory problems that require simulations for long time-transients, that typically need about 10K core-hours per case. We intend to run about 20-30 of these cases. Industry partnership: None Project URL: http://www.mcs.anl.gov/~fischer/sem1b Current FY Hours Used: undetermined amount New FY Requested allocation: 600000 Q1: 150000 Q2: 150000 Q3: 150000 Q4: 150000 Justification: Near linear strong scaling has been seen in all problems studied. Typically, 3000 grid points/core is required to ensure both strong and weak scaling. Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Request: BEM_Uncertainty
by accounts＠lcrc.anl.gov 24 Aug '17

24 Aug '17

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: Ralph T. Muehleisen Applicant's institution: ANL Applicant's division: ES Project Name: BEM_Uncertainty Project title: Building Energy Model Uncertainty and Sensitivity Associated funding: Building Technologies Office of DOE Other Systems: Science: Quantify the input sensitivity and determine the modeling uncertainty for a detailed building energy model to be used in the development of an empirical building energy model validation standard. Project description: This computing time will be used to quantify the input sensitivity and uncertainty in EnergyPlus building energy models developed to model two highly instrumented buildings. The underlying project is the development of a method of empirical validation of building energy models. Detailed and precision experiments are being made of two different experimental buildings at LBL and ORNL. Extremely high fidelity baseline EnergyPlus building energy models of the experimental facilities are studied to determine the sensitivity of the models to various inputs which are used to help improve the experimental setup. Typical uncertainties in the inputs of the baseline model are then propagated through the model using Monte Carlo methods and output predictions with uncertainty are generated. This uncertainty is then analyzed and used as a basic for determining a validation metric to be used for validation of lower fidelity models or models created with other softw are. The project will utilize python scripts, the python SAlib sensitivity library, Pandas, and the EnergyPlus building energy model. The Morris method is first used as a screening to determine which of the thousands of model inputs are to be included in the global sensitivity analysis. A global sensitivity analysis is performed to develop Sobol sensitivity indices for the selected inputs. A full uncertainty analysis of the model is performed using the Argonne / Georgia Tech building input uncertainty database and monte carlo methods to propagate the uncertainty through the models. EnergyPlus has been recently updated to better use multiple cores internally and Monte carlo analysis can naturally parallelize to utilize all available nodes so this project can make good use of available nodes and the available cores per node. In the conversion of EnergyPlus from Fortran to C++ several years ago, better internal multithreading was added. Researchers at ORNL have used, and continue to use Titan in a similar manner for their large scale building energy modeling and researchers at NREL and LBL also do similar modeling on their HPC systems. Project team members will consult with the ORNL, LBL, and NREL, teams to ensure that we are making best use of Blues or BeBop. This project will start when after new postdoc in Q1 so only minimal time is requested for Q1 Industry partnership: Project URL: Requested allocation: 155000 Q1: 5000 Q2: 50000 Q3: 50000 Q4: 50000 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

1 0

[LCRC Accounts] Yearly Allocation Request for gcmat
by accounts＠lcrc.anl.gov 23 Aug '17

23 Aug '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Jonathan Ozik Project Name: gcmat Division: GSS Project title: Global Critical Materials Modeling Associated funding: DOE Office of Energy Policy and Systems Analysis DOD Defense Logistics Agency Other Systems: Science: This project uses adaptive simulation methods to compute sensitivity and uncertainty ranges for predicting global price and production of critical materials needed for clean energy technologies. Project description: We will continue to use agent-based modeling to study the effects of producer and consumer decisions on market dynamics for rare earth elements. The project sponsors are particularly interested in understanding the uncertainty in future demand and shortfall forecasts. With that objective, the DLA has contracted with Argonne to support the development of their 2019 and 2021 reports to Congress on strategic and critical material stockpile requirements. We also intend to understand the effects of critical parameters on model outcomes. We will continue to use Repast Simphony for agent-based modeling and the Extreme-scale Model Exploration with Swift (EMEWS) framework for constructing large-scale adaptive workflows. Our goal is to continue to advance methods for adaptive simulation experiments that will be widely applicable to many kinds of models. We will carry out parameter sweeps over 100k+ to 1M+ parameter combinations, determining regions of parameter space that exhibit the best fit to empirical data, including historical prices of critical materials, and parameter sensitivity. We plan to exploit the increased concurrency of Bebop to run 2000+ process workflows with walltimes up to 20 hours, with potentially longer walltimes for the more dynamic sampling methods. The adaptive simulation workflows will allow us to strategically explore large parameter spaces, dynamically generating parameter combinations based on results of distributed simulation runs, using evolutionary algorithms and approximate Bayesian computation. 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: The project generates large amount of data in the form of simulation outputs. Outputs from large parameter sweeps involving millions of simulation runs quickly add up. This data needs to be stored for analysis and for joint access by the project members. As such, a 5TB storage allocation would be very helpful. Thank You, The LCRC Accounts System

1 0

New LCRC report submitted for gcmat
by support＠lcrc.anl.gov 23 Aug '17

23 Aug '17

Full project report at https://accounts.lcrc.anl.gov/reporting/project/2017/?project=gcmat

1 0

[LCRC Accounts] Project Request: BebopPerformance
by accounts＠lcrc.anl.gov 23 Aug '17

23 Aug '17

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: Dan Olson Applicant's institution: ANL Applicant's division: CELS Project Name: BebopPerformance Project title: System performance analysis/improvement with Intel/Cray Associated funding: Other Systems: Science: Project description: This will be used for collaborative work with Intel and Cray to improve the performance of the Bebop system. Industry partnership: Project URL: Requested allocation: 4000000 Q1: 1000000 Q2: 1000000 Q3: 1000000 Q4: 1000000 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

1 0

New LCRC report submitted for NEXTGENOPT
by support＠lcrc.anl.gov 23 Aug '17

23 Aug '17

Full project report at https://accounts.lcrc.anl.gov/reporting/project/2017/?project=NEXTGENOPT

1 0

[LCRC Accounts] Yearly Allocation Request for DMDOHEMA_1
by accounts＠lcrc.anl.gov 23 Aug '17

23 Aug '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Baofu Qiao Project Name: DMDOHEMA_1 Division: CSE Project title: Metal Ion Transport through Nano-Structured Lipophilic Systems: Atomistic MD Simulations Associated funding: DOE-BES Other Systems: Science: The interactions and transport mechanisms of metal ions through nano-structured soft matter systems is woefully lacking in fundamental understanding. Such systems are imported in a huge range of applications, such as: extractive separations; forming meso-porous catalytic supports; nano-particle synthesis; life-mimicking systems; etc. Understanding multi-scale structures and interactions is key to developing these important applications and is at the forefront of meso-science. However, soft-matter systems containing coordinating metal ions are notoriously difficult to probe structurally, with the best techniques being synchrotron SAXS and EXAFS giving information on the atomic metal-centric scale and supramolecular scale, respectively. Interpreting such experimental data from these complex systems can give a general idea of coordination environment and the morphology of the supramolecular solvent. However, in order to gain a truly multi-scale understanding with atomi c resolution, we must couple these experimental data with MD simulations, which can be done conveniently and powerfully by directly fitting SAXS and EXAFS data with the simulated systems. This methodology will provide a breakthrough in new fundamental insights into the structural complexity of metal ion – soft matter systems. Project description: We are currently working on two projects. In the first project, we are developing the pairwise additive CHARMM force field parameters for the trivalent metal ions. The precise understanding of the speciation behavior of metal ions depends on the accurate description of the metal ions and the interacting ligands. In contrast with the ligand, the force field parameters of most of which have been well developed, the accurate potentials for the high valent metal ions are very limited. Most of them are restricted in the polarizable potential (e.g., AMOEBA potential). These polarizable potential are computationally much more expensive. Therefore, we are aiming to develop the additive potentials for the trivalent metal ions, which are expected to be of similar accuracy with the polarizable force fields and much more computationally efficient. In specific, we are developing the additive CHARMM force field for trivalent lanthanide (Ln(III)) ions. The free energy perturbation method is employed to calculate the hydration free energy of the Ln(III) ions. The recently proposed corrections to the calculated hydration free energy are applied. Moreover, the structural features are also studied. Our preliminary results show that the obtained CHARMM fore field parameters for a variety of Ln(III) ions are capable of providing comparable accuracy with the polarizable AMOEBA force field parameters. The additive CHARMM potential is around two orders of magnitude computationally faster. We are currently writing the manuscript, and will submit it to J. Phys. Chem. Lett. for publication. In the second project, we are studying how surfactants could be employed to separate trivalent metal ions. We are currently using the anionic SDS as the amphiphilc surfactant. Preliminary simulations show that the presence of trivalent metal ions drives the SDS monomers into multilayer structure, and most of the trivalent metal ions are buried inside the multilayer structure. We are currently equilibrating the structures. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 350000 Q1: 200000 Q2: 50000 Q3: 50000 Q4: 50000 Justification: Storage requirements: 1TB storage is large enough for our project. Thank You, The LCRC Accounts System

1 0

New LCRC report submitted for DMDOHEMA_1
by support＠lcrc.anl.gov 23 Aug '17

23 Aug '17

Full project report at https://accounts.lcrc.anl.gov/reporting/project/2017/?project=DMDOHEMA_1

1 0

[LCRC Accounts] Yearly Allocation Request for LiBattModeling
by accounts＠lcrc.anl.gov 22 Aug '17

22 Aug '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Roy Benedek Project Name: LiBattModeling Division: CSE Project title: Lithium-Ion Battery Electrode Modeling Associated funding: Battery Materials Research Program, Office of Vehicle Technologies, U. S. Department of Energy Other Systems: NERSC (about 400K core hrs.) Science: We apply first-principles (density functional theory) techniques to investigate candidate lithium-ion-battery cathode materials. Our recent efforts address primarily lithium-rich transition metal (particularly Ni, Mn, and Co) oxides with layered structures. The objective is to identify high energy-density and low (voltage and capacity) fade compositions and structures. This modeling effort complements the experimental program led by Michael Thackeray and Jason Croy in the Battery Department of the CSE Division, ANL. Project description: In previous work, we performed first principles MD simulations to characterize the atomic transformations that lithium-rich layered oxide cathode materials undergo during the first charge. We model the material with a 384 lattice site periodic unit cell. We found that after removal of a substantial fraction (about 2/3) of the Li ions from the material (at a state-of-charge near the middle of the Voltage Plateau), the oxygen sublattice becomes unstable with respect to dimerization, and after the creation of oxygen defects, transition metal ions can migrate from the metal layer to the lithium layer. In a sequel to this earlier work, we are now performing simulations of the first discharge (This work is about 1/3 complete). The discharge addresses compositions Li0.2+xMn0.4Co0.6O2.4, as a function of relithiation constant, x. The main objectives are to (a) determine to what extent the disordering introduced during the first charge is reversible during discha rge, (b) estimate the loss of capacity associated with the disorder induced during the first charge, and (c) estimate the amount of trivalent Mn created in the discharge. A general goal of the work is to provide a better fundamental understanding of materials that have been the subject of a huge amount of experimental work. The first principles MD simulations have been performed up to now using version 5.4 of VASP. Argonne is serving as a beta test site for version 6 of VASP, which has been implemented by John Low on Blues, and I am currently switching my work to this version of the code. I expect that will enable a higher throughput, and expedite progress on this project. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 800000 Q1: 200000 Q2: 200000 Q3: 200000 Q4: 200000 Justification: I am expecting that the beta-test version of VASP 6.0 will enable better scalability (up to about 1024 processors) than previous versions. Storage requirements: Thank You, The LCRC Accounts System

1 0