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

[LCRC Accounts] Yearly Allocation Request for Test
by accounts＠lcrc.anl.gov 05 Aug '15

05 Aug '15

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: John Blaas Project Name: Test Division: CLS Project title: Test Associated funding: Me Other Systems: Fusion Science: Test Project description: test Industry partnership: test Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 4 Q1: 1 Q2: 1 Q3: 1 Q4: 1 Justification: Storage requirements: None Thank You, The LCRC Accounts System

1 0

Re: [allocations-admins] 8/5 Outstanding Allocations Requests and Recommendations
by Bair, Raymond A. 05 Aug '15

05 Aug '15

Thanks, John. ----------------------------------------- Ray Bair Computing, Environment, and Life Sciences Argonne National Laboratory and the University of Chicago TCS Building 240, Room 4122 9700 South Cass Avenue Argonne, IL 60439 email: rbair(at)anl.gov Phone: (630)252-5751 On 8/5/15, 10:54 AM, "allocations-admins-bounces(a)lcrc.anl.gov on behalf of John Blaas" <allocations-admins-bounces(a)lcrc.anl.gov on behalf of wilshire(a)mcs.anl.gov> wrote: >> MD_FeCr >> Yinbin Miao >> 150K on Fusion is OK >> Recommend: Approve > >Done > >> MBM_CEES2 >> Hakim Iddir >> 150K on Fusion is OK >> Recommend: Approve > >Done > >> ElectrolyteRedoxFlow >> >> Cheng Lei >> 50K on Fusion is OK >> Recommend: Approve > >Done > >> XRayFuelSpray >> >> Daniel Duke >> 100K on Fusion is OK >> Recommend: Approve > >Done > >> Li-S_Sol >> >> KC Lau >> 150K on Fusion is OK >> Recommend: Approve > >Done > >- JB >_______________________________________________ >allocations-admins mailing list >allocations-admins(a)lcrc.anl.gov >https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins

1 0

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by John Blaas 05 Aug '15

05 Aug '15

Granted 150K on Fusion and sent note to PI. - JB On Tue, Aug 4, 2015 at 2:39 PM, accounts(a)lcrc.anl.gov <accounts(a)lcrc.anl.gov> wrote: > Hello, > > A change in allocation has been requested: > > Requester: kclau (Kah Chun Lau) > Project: Li-S_Sol > Title: Li-S battery Solution > Description: The project can be divided into two related problems: > (I) Solubility of polysulfides and sulfides > A systematic quantum thermochemistry calculation of free energy of solvation for lithium polysulfides (Li2S, Li2S2, and Li2Sn, n > 2) will be carried out using Gaussian09 code. For each polysulfides species, the effect of a range of solvent dielectric constant will be investigated systematically. This will give a set of various optimized configurations of solvated polysulfides as an input for initial configuration in AIMD bulk simulation using VASP code. > > (2) Dynamics of solvated polysulfides species > From the item#1, at least two different bulk solvent electrolytes will be chosen for AIMD NVT simulation to investigate the dynamic properties (e.g. ion diffusion) of the solvated polysulfides species. Overall, the calculations will be carried out either based on VASP or CP2K. > > > Calculation size > For a standard quantum thermochemistry calculations (item#1), we plan to simulate about at least about 20-30 configurations for the low energy configurations of polysulfides. To obtain a reasonable good accuracy with high level quantum chemistry calculations, about 100 K processor hrs is needed. > > The smallest cells for which the effects of interest can realistically be simulated for electrolyte AIMD simulation will contain about four hundred atoms. All numerical shortcuts (gamma-point only, soft pseudopotentials) will be exploited. We expect to run on 256 processors. Of the calculations proposed here, the AIMD simulations (item 1 and 2) will have the largest CPU-time requirements per unit MD-time step of 1 fs. The AIMD simulations performed in NVT-ensemble simulations may require about 1 processor-hr per time step, for a cell size of about 400 atoms (using perhaps 256 processors). Obtaining a reasonably good MD statistics calculation may require of the order of 100 k processor hrs. In our problem, at least two to three different electrolytes will be explored, therefore it requires 500k processor hours for AIMD simulation for both item#1 and #2. > Thus requirements for the entire project are thought to be 600,000 processor hours. > > > Current: undetermined amount > Justification: The smallest cells for which the effects of interest can realistically be simulated will contain about three to four hundred atoms electrolyte. All numerical shortcuts (gamma-point only, soft pseudopotentials) will be exploited. We expect to run on about 256 processors, a parallel efficiency of ~ 60% -75% can be achieved using 256 cores for VASP AIMD simulation. Thus to explore a moderately large systems (~ 300-400 atoms) with simulation time up to ~ 5ps in 1 fs time step, about 100 000 core-hour is sufficient to obtain to reasonably good thermodynamic statistics. > > Requested: 180000 > > A specific reason has been given: > To complete three ongoing Li ion diffusion in three different solvents (~360 atoms equilibrated for another 2 ps) for same solvated Li-salt AIMD calculations (which the total still need 100K core hours). Another 80K core hours are used for Gaussian09 calculations for implicit solvents, thermochemistry calculations for 4 different polysulfides species (~ 40 - 70 atoms) with partial configurations obtained from AIMD trajectories (with total ~ 12 configurations). > > This needs to be approved and the final allocation amount decided upon. > > 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

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by John Blaas 05 Aug '15

05 Aug '15

Granted 100K for use on Fusion and sent note to PI. - JB On Mon, Aug 3, 2015 at 3:06 PM, accounts(a)lcrc.anl.gov <accounts(a)lcrc.anl.gov> wrote: > Hello, > > A change in allocation has been requested: > > Requester: dduke (Daniel Duke) > Project: XRayFuelSpray > Title: Fuel Injection and Sprays Studied Using X-Ray Diagnostics > Description: The cavitation modeling component of the project is being performed using the OpenFOAM framework. A number of state-of-the-art numerical solvers have been implemented in OpenFOAM; namely incompressible-liquid and fully compressible homogeneous relaxation cavitation models and more recently a fully compressible homogeneous relaxation cavitation model which includes non-condensable gas modeling (HRMFoam). These codes were developed by Schmidt et al at the University of Massachusetts-Amherst, and we continue to collaborate with them on code development and efficiency improvements. > > In the last year, we have made some substantial improvements to the cavitation model and the parallel efficiency of its implementation on Blues/Fusion. With the support of LCRC technical staff and our collaborators at the University of Massachusetts-Amherst, we have implemented a custom build of OpenFOAM-3.0 on Blues against the system native MPI, and we have made the necessary changes to the HRMFoam solverâ€™s source code in order to move from the older OpenFOAM-1.6 to the new 3.0 framework. This has led to substantial gains in parallel performance; up to 20% on 32 cores and 12% on 64 cores on Blues. A scalability study has been performed with the new code, and details are given in the Large Allocation Efficiency statement. > In the new OpenFOAM-3.0 build, we have successfully implemented both Large Eddy Simulation (LES) and Realizable k-epsilon (KE) turbulence models. Furthermore, we have implemented code which allows us to take virtual projections of the three-dimensional simulations so that they may be quantitatively compared with the x-ray experiments. > > In the following year, we plan to extend high-fidelity cavitation simulations by undertaking longer simulation times in order to capture lower-frequency hydrodynamic phenomena which we have observed in our experiments. Improvements in efficiency will allows us to pursue longer run-time simulations. We also intend to simulate the effects of varying pressure and flow rate on cavitation, in order to compare against new x-ray fluorescence data from APS. > The gas jet simulations which we plan to undertake in Q1 are performed with the CONVERGE CFD software, its use is already well established at LCRC. > > Tomographic reconstruction of gasoline direct injection spray measurements from APS will be conducted using the TomoPy software package developed at APS. We intend to run single-node jobs to efficiently perform multiple reconstructions of large (multi-GB) datasets. We have also successfully compiled a Paraview server on our Fusion dedicated nodes which allows us to rapidly visualize the results of our simulations and reconstructions. Using Blues and Fusion will allow us to achieve rapid turnaround of results during beamtime. This will be a significant aid to the experimental program. As the quantity of tomography work increases, we may consider purchasing more dedicated nodes. > Current: undetermined amount > Justification: Our requested allocation for next year will exceed 0.5M core-hours in order to encompass the gas jet simulations in Q1 and tomographic reconstruction work in Q2-4. > > The majority of core hours will still be dedicated to OpenFOAM simulations, which will be undertaken on both dedicated Fusion nodes and also on Blues, using OpenFOAM-3.0. We have demonstrated notable improvements in scalability on Blues as detailed above. > > A scalability study of OpenFOAM-3.0 performance with multi-million cell meshes has been undertaken, demonstrating excellent linear scalability. We observe 81% efficiency relative to single-CPU performance, scaling linearly up to 320 cores (20 nodes) with a 15-million cell mesh (a typical upper limit). We have implemented Scotch decomposition in order to ensure good load balancing & minimization of cross-processor communication boundaries. The present OpenFOAM simulations use static meshes, so load rebalancing at runtime is not a concern. We intend to run typical job sizes of 64 to 128 cores; smaller than the maximum job size for which good scalability has been demonstrated. Further details regarding scalability studies can be obtained by contacting the project PIs. > > Requested: 100000 > > A specific reason has been given: > We would like to request an additional 100k core hours to continue ongoing numerical simulations work for rapid publication, which we intend to complete before FY2016 allocations are assessed. Our prior allocation was exhausted due to the need to perform additional simulations (for conference papers) which were not originally planned, and to perform tomographic reconstructions of x-ray data from APS. We will use the additional hours to perform OpenFOAM simulations; scalability test data are available upon request (up to 300 cores using system MPI). We intend to run 100-200 core sized jobs. If 100k allocation is not possible, we would benefit from any additional allocation that can be made available. > > This needs to be approved and the final allocation amount decided upon. > > Thank You, > The LCRC Accounts System

1 0

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by John Blaas 05 Aug '15

05 Aug '15

Granted 50K for use on Fusion and sent note to PI. - JB On Mon, Aug 3, 2015 at 2:12 PM, accounts(a)lcrc.anl.gov <accounts(a)lcrc.anl.gov> wrote: > Hello, > > A change in allocation has been requested: > > Requester: leicheng (Lei Cheng) > Project: ElectrolyteRedoxFlow > Title: discoveries of electrolyte for redox flow batteries > Description: In this renewal work, we will continue focusing on fundamental properties of transition metal complexes as redox active molecules for non-aqueous redox flow battery application. In collaboration with experimental group (Prof. Thompson, University of Michigan), we have previously studied redox windows of some novel compounds. We now study the solubilities of these compounds in non-aqueous solutions as the capacity of this type of battery is directly proportional to redox window and concentration of active compound in solution. Insight gained from these studies will greatly guide design and synthesis of the compounds experimentally. > In previously proposed work, we have successfully determined the B3LYP/6-31+G* and BP86/6-31+G* combinations as proper computational treatments for these system. The same methods will be used to calculate solvation energies of Cr, V, Mn and Fe based novel compounds. These energies will be calculated using both implicit continuum models such as PCM, CPCM and SMD methods, as well as explicit model with actual solvent molecules added to the quantum calculations. These solvation energies will then be compared with experimentally measured solubilities of the compounds to determine correlation between the two. Calculations with explicit solvent molecules will also be performed to further validate accuracy of the implicit model calculations as well as provide more insights on coordination details of the metal complexes by solvent molecules. Recently, interesting solubility trends of metal complexes in different binary organic solvents have been observed experimentally (private com > munication). We will also guide our research efforts towards providing physical understandings of the observed experimental results. > All calculations will be performed using Gaussian and NWCHEM. These two code packages are both considered very efficient and reliable for quantum chemistry calculations. Gaussian is not very scalable but it provides really robust geometry optimizer and the G3/G4 methods are built-in. NWChem is very scalable especially for higher level calculations such as coupled cluster methods and perturbation theories. After interesting molecule candidates and appropriate calculation methods are identified, derivatives of these molecules will be subject to high-throughput screening. The high-throughput screening will be performed using NERSC supercomputers. > Current: undetermined amount > Justification: The size of molecules we propose to study is very typical for the two computational software (Gaussian and NWChem) so they scale reasonably well for most DFT calculation. The higher level calculations will consume more memory as well as computational time. The NWChem code uses a globally-shared memory system, therefore one can request and utilize more memory through node parallelization. > > Requested: 50000 > > A specific reason has been given: > We have used up the computing time allocated because the systems (metal > organic complexes) of interest are a lot harder to converge than expected > due to the presence of transition metals with different possible charge and multiplicity states. In addition, our experimental collaborators have discovered some more interesting systems with different metal centers > that we are interested in investigating computationally. We are preparing > a manuscript based on the work conducted under this proposal. These > additional allocation will allow us to finish the extra computations > needed. > > I really appreciate your consideration! > Sincerely, > Lei Cheng > > This needs to be approved and the final allocation amount decided upon. > > 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

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by John Blaas 05 Aug '15

05 Aug '15

Granted 150K for Fusion and sent note to PI - JB On Fri, Jul 31, 2015 at 3:58 PM, accounts(a)lcrc.anl.gov <accounts(a)lcrc.anl.gov> wrote: > Hello, > > A change in allocation has been requested: > > Requester: iddir (hakim iddir) > Project: MBM_CEES2 > Title: Modeling Battery Materials CEES2 > Description: In this work we will investigate the particles shape for non-stoichiometric as well as with dopants. We will perform theoretical work based on density functional theory using “VASP”, to simulate the equilibrium particle shapes of non-stoichiometric and doped LiMn2O4 particles, with the goal being an optimized particle shape that will provide both stability and good rate capability. Total energy calculations will be performed for different LiMn2O4 stoichiometries and compositions, in which Mn will be substituted with other candidate transition metals such as Ni, Mo, and Ti. This will require total energy calculations of surfaces (slab geometry) as well as corresponding bulk compositions. > A second focus will be on investigation of the effect of the support and hence the interface/strain of epitaxial thin films of LiMn2O4 grown on single crystals, such as SrTiO3 and MgO for three crystallographic orientations (111) (100) and (110). Here we will focus on the interplay between Jahn-Teller distortion, Mn dissolution and Strain. > > Current: undetermined amount > Justification: > > Requested: 200000 > > A specific reason has been given: > The systems investigated in this project seem to show some instabilities with respect to the magnetization of the elements such Mn, Co and Ni. > I am currently investigating a new approach in getting the lowest electronic ground state of these compounds, using the U-ramping method (DFT+U). With this technique a calculation that used to require a converged state with a specific U value, followed by single point calculation with no spin and with spin, now would require several calculations with U going from 0 to the desired U terminal value with small increments. Hence making these calculations more costly. > The AIMD simulations are very demanding and the structures under investigation are still not converged. > > This needs to be approved and the final allocation amount decided upon. > > Thank You, > The LCRC Accounts System

1 0

Re: [allocations-admins] [LCRC Accounts] Project Request: MD_FeCr
by John Blaas 05 Aug '15

05 Aug '15

Granted 150K for Fusion and sent note to PI - JB On Thu, Jul 30, 2015 at 12:24 PM, 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: Yinbin Miao > Applicant's institution: ANL > Applicant's division: NE > Project Name: MD_FeCr > Project title: Molecular Dynamics Simulation of Radiation-induced Microstructural Modifications in Fe-Cr Alloys > Associated funding: DoE > Other Systems: None > Science: Fe-Cr based alloys are regarded as the most important structural materials in both current and prospective nuclear reactors, mainly due to their exceptional radiation tolerance and high-temperature properties. In the extreme conditions such as elevated temperatures and high neutron flux, materials experience a series of modifications at microstructural level. As a result, the macroscopic mechanical and thermal properties are compromised. In order to better understand the fundamentals and mechanisms of these radiation-induced microstructural modifications in Fe-Cr based alloys, Molecular dynamic simulations are proposed to be performed to investigate the defects caused by radiation at atomic level. This effort will provide important information that would guide the development of advanced Fe-Cr based alloys that can be applied for next-generation nuclear systems. > Project description: In this proposed project, molecular dynamics simulation will be used to investigate the microstructural modifications induced by irradiation. LAMMPS code will be used to realize the MD simulation. The results will be compared with the experimental post-irradiation examination (PIE) data I or my group members collected or will collect in another DoE funded project. LAMMPS is based on the fundamental Newton mechanics to predict the behaviors of each atom in a system containing up to millions of atoms. The positions and velocities of all the atoms along with their interactions are captured by the code. MPI was used to enhance the parallel efficiency for LAMMPS code. The same code has been run on both LCRC clusters and the Taub cluster in University of Illinois, for similar objectives in other class of materials. For the calculation, a typical run will need around 4 to 8 nodes . This project will only involve one user. > Industry partnership: > Project URL: > Requested allocation: 450000 > Q1: 0 > Q2: 0 > Q3: 0 > Q4: 450000 > Justification: > Storage requirements: Enough > > > 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

Re: [allocations-admins] 8/5 Outstanding Allocations Requests and Recommendations
by John Blaas 05 Aug '15

05 Aug '15

> MD_FeCr > Yinbin Miao > 150K on Fusion is OK > Recommend: Approve Done > MBM_CEES2 > Hakim Iddir > 150K on Fusion is OK > Recommend: Approve Done > ElectrolyteRedoxFlow > > Cheng Lei > 50K on Fusion is OK > Recommend: Approve Done > XRayFuelSpray > > Daniel Duke > 100K on Fusion is OK > Recommend: Approve Done > Li-S_Sol > > KC Lau > 150K on Fusion is OK > Recommend: Approve Done - JB

1 0

8/5 Outstanding Allocations Requests and Recommendations
by Bair, Raymond A. 05 Aug '15

05 Aug '15

These are all requests for additional time. MD_FeCr Yinbin Miao 150K on Fusion is OK Recommend: Approve MBM_CEES2 Hakim Iddir 150K on Fusion is OK Recommend: Approve ElectrolyteRedoxFlow Cheng Lei 50K on Fusion is OK Recommend: Approve XRayFuelSpray Daniel Duke 100K on Fusion is OK Recommend: Approve Li-S_Sol KC Lau 150K on Fusion is OK Recommend: Approve

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 04 Aug '15

04 Aug '15

Hello, A change in allocation has been requested: Requester: kclau (Kah Chun Lau) Project: Li-S_Sol Title: Li-S battery Solution Description: The project can be divided into two related problems: (I) Solubility of polysulfides and sulfides A systematic quantum thermochemistry calculation of free energy of solvation for lithium polysulfides (Li2S, Li2S2, and Li2Sn, n > 2) will be carried out using Gaussian09 code. For each polysulfides species, the effect of a range of solvent dielectric constant will be investigated systematically. This will give a set of various optimized configurations of solvated polysulfides as an input for initial configuration in AIMD bulk simulation using VASP code. (2) Dynamics of solvated polysulfides species >From the item#1, at least two different bulk solvent electrolytes will be chosen for AIMD NVT simulation to investigate the dynamic properties (e.g. ion diffusion) of the solvated polysulfides species. Overall, the calculations will be carried out either based on VASP or CP2K. Calculation size For a standard quantum thermochemistry calculations (item#1), we plan to simulate about at least about 20-30 configurations for the low energy configurations of polysulfides. To obtain a reasonable good accuracy with high level quantum chemistry calculations, about 100 K processor hrs is needed. The smallest cells for which the effects of interest can realistically be simulated for electrolyte AIMD simulation will contain about four hundred atoms. All numerical shortcuts (gamma-point only, soft pseudopotentials) will be exploited. We expect to run on 256 processors. Of the calculations proposed here, the AIMD simulations (item 1 and 2) will have the largest CPU-time requirements per unit MD-time step of 1 fs. The AIMD simulations performed in NVT-ensemble simulations may require about 1 processor-hr per time step, for a cell size of about 400 atoms (using perhaps 256 processors). Obtaining a reasonably good MD statistics calculation may require of the order of 100 k processor hrs. In our problem, at least two to three different electrolytes will be explored, therefore it requires 500k processor hours for AIMD simulation for both item#1 and #2. Thus requirements for the entire project are thought to be 600,000 processor hours. Current: undetermined amount Justification: The smallest cells for which the effects of interest can realistically be simulated will contain about three to four hundred atoms electrolyte. All numerical shortcuts (gamma-point only, soft pseudopotentials) will be exploited. We expect to run on about 256 processors, a parallel efficiency of ~ 60% -75% can be achieved using 256 cores for VASP AIMD simulation. Thus to explore a moderately large systems (~ 300-400 atoms) with simulation time up to ~ 5ps in 1 fs time step, about 100 000 core-hour is sufficient to obtain to reasonably good thermodynamic statistics. Requested: 180000 A specific reason has been given: To complete three ongoing Li ion diffusion in three different solvents (~360 atoms equilibrated for another 2 ps) for same solvated Li-salt AIMD calculations (which the total still need 100K core hours). Another 80K core hours are used for Gaussian09 calculations for implicit solvents, thermochemistry calculations for 4 different polysulfides species (~ 40 - 70 atoms) with partial configurations obtained from AIMD trajectories (with total ~ 12 configurations). This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0