September 2014 - Allocations-admins - mailman3.cels.anl.gov

[LCRC Accounts] Yearly Allocation Request for NEXTGENOPT
by accounts＠lcrc.anl.gov 10 Sep '14

10 Sep '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Naiyuan Chiang Project Name: NEXTGENOPT Division: MCS Project title: Next generation optimization Associated funding: Office of Science Other Systems: Fusion Science: Advances in algorithms for stochastic optimization (SO) and uncertainty quantification (UQ) are critical for the efficient design and real-time optimization of next-generation national infrastructures. Anticipating and mitigating uncertainty of weather, demands, and contingencies in a more integrated environment is necessary to mitigate market volatility and prevent cascading failures that can ultimately lead to catastrophic shortages of supply. The integrated monitoring and optimization of infrastructures systems will generate huge amounts of data that cannot be possibly processed without exploiting underlying probabilistic structures (e.g., sparsity in spatio-temporal correlation patterns) and without exploiting the properties of the particular decision-making objective at hand (e.g., stochastic, PDEs). It is thus at the heart of our current research the philosophy that a structure-oriented and integrated approach to UQ/SO is necessary. Project description: Sponsored by a Department of Energy (DOE) Early Career Award, we are currently developing new algorithms for UQ/SO that have a holistic view of the entire data-modeling-optimization process. A particular example of such algorithms is a new clustering-based interior-point strategy for stochastic programs in which scenarios are compressed based on their influence on the first-stage decision, not on data. This has shown to lead to drastic compression rates (and solution times) of above 80%, which cannot be achieved with data compression alone (see Figure 1). The compression is done adaptively, at the linear algebra level, by looking at the different contributions of the scenarios on the Schur complement. We are also currently developing new interior-point algorithms for nonconvex structured optimization capable of dealing with complex multi-scale physics (typically involving large networks of partial differential equations) and stochastic components. These problems arise, for instance, from natural gas inventory (line-pack) management (see Figure 2). This requires of a redesign of existing algorithms and modeling environments capable of conveying problem structure down to the linear algebra kernels. In addition, hybrid linear algebra kernels need to be designed to exploit different structural facets. Problems with billions of variables are being targeted. Industry partnership: Project URL: http://www.mcs.anl.gov/~vzavala/earlycareer.html Current FY Hours Used: undetermined amount New FY Requested allocation: 80000 Q1: 20000 Q2: 20000 Q3: 20000 Q4: 20000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Request: ML4Maxwell
by accounts＠lcrc.anl.gov 10 Sep '14

10 Sep '14

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: Alejandro Lopez Bezanilla Applicant's institution: ANL Applicant's division: MSD Project Name: ML4Maxwell Project title: Machine Learning to Solve a Maxwell's Equation Associated funding: DOE Other Systems: Science: A machine learning approach will be used to demonstrate that the first Maxwell's equation can be solved by means of statistical methods. A large amount of data describing the potential created by a charge distribution in multidimensional space with boundary conditions will be computed to obtain a training and validating set. An algorithm is expected to learn from the training data and compute new configurations. The accuracy will be compared with the validating data set. We performed a similar study on charge transport last year and successfully published on Phys. Rev. B 89, 235411 Project description: The numerical evaluation of the potential generated by punctual charges will be done by solving the Poisson's equation on a fine grid containing about 10^6 points. A sufficient large amount of training set will be generated with well-established methods such as the finite differences. A parallelized in-home C/C++ code with linear scaling over multiple nodes will be used. Also we are evaluating to perform calculations with the code MADNESS, specially for the challenging case of a 3D configuration involving 10^9 point and large amount of memory allocations. Industry partnership: Project URL: Requested allocation: 450000 Q1: 125000 Q2: 100000 Q3: 125000 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

1 0

[LCRC Accounts] Yearly Allocation Request for Comp_Soft_Catalysts
by accounts＠lcrc.anl.gov 10 Sep '14

10 Sep '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Juan Lopez-Encarnacion Project Name: Comp_Soft_Catalysts Division: CSE Project title: Soft Catalysts for Green Chemistry and Energy applications Associated funding: Strategic Initiative on Materials for Energy (LDRD-2013-060-R2) Other Systems: None Science: The objective of this project is to perform extensive modeling and computational studies that will aid in design and synthesis of new catalytic materials for efficient transformation of abundant C1 sources (carbon dioxide (CO2) and methane (CH4), in particular) into chemical feedstocks and fuels. The planned theoretical/computational work will be performed in close contact with the ongoing experimental effort on the subject at CSE. The goal is to design and synthesize a new type of catalytic materials that will combine the best characteristics of both homogeneous catalysts (high selectivity at low temperature, tunability) and heterogeneous catalysts (stability, ease of separation). Such catalytic systems are not available at present. Project description: In the FY2015, the computational plan will continue focus on molecular dynamics studies of pores sizes, apertures sizes, and accessible-solvent surface area at different temperatures of the crystalline and amorphous porous organic polymers (POPs). The POPs we are planning to generate are those with the building blocks monomers relevant to experimentalist in our project. We will continue the exploration of alkane metathesis reaction mechanisms promoted by different type of Ta-based/POP catalysts. One of the targeted processes is conversion of ethane into methane and propane. In addition, we started and will continue performing computations on chemical reaction mechanism for synthesis of ammonia from N2 and H2 gases (i.e., one of the challenging problem in catalysis, because the high temperature and high pressure required in the current industrial process). Based on our preliminary results, we expect the Ta-based reaction centers embedded on POP supports t o reduce the relevant reaction barriers as compared to those of the more common supports, e.g., silica or alumina based. The structural and electronic aspects of reactants, intermediates, transition states, and products will be explored and characterized using state-of-the-art density functional theory. The number of atoms treated at a time will vary from a few to a few tens (e.g., in organometallic reaction centers), from tens to about a hundred (e.g., in struts), and to a few hundreds (e.g., in POP networks). In the case of generated 3D amorphous POP networks, the studies in will require inclusion of thousands of atoms, and they will be performed using efficient and at the same time adequate semiempirical potentials and/or force fields to characterize their low-energy structures and dynamical aspects. Then single-point DFT treatments will be performed on these structures to characterize their electronic features using SIESTA or CP2K which allow to performed calculations for thousands of atoms. The dynamical and thermal behavior will be explored using both first principles based and semiempirical potentials/force fields. The kinetic aspects will be evaluated and characterized using transition state theory. In addition to the SIESTA and CP2K codes we will also use other pa ckages (e.g., LAMMPS, NWChem, and Gaussian) as appropriate. The requested time is based on estimates performed on Fusion (see below). The methods used to study the dynamical/kinetics aspects of our systems require considerably computational power/time than those methods used during the FY2014 and the test show below (e.g., crystal, atomic, and electronic structure computations). The expected number of project members is three (3) or more. Industry partnership: N/A Project URL: http://www.cse.anl.gov/catalysis_energy_conversion/homogeneous/index.html Current FY Hours Used: undetermined amount New FY Requested allocation: 980000 Q1: 245000 Q2: 245000 Q3: 245000 Q4: 245000 Justification: The scaling efficiency on Fusion was tested using the code SIESTA in the full optimization of the atomic structure and cell volume of a crystalline porous organic polymer, which contain a total of 584 atoms: 192 atoms of H (5 numerical basis function per each H), 360 atoms of C (13 numerical basis function per each C), and 32 atoms of O (13 numerical basis function per each O) in its unit cell; a total of 6,056 basis functions. Substantial speed-up is achieved with the increase in the number of cores: Cores Average elapsed time per optimization step (min) 16 4.38 32 2.70 64 2.34 128 1.68 Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for STOandPlus
by accounts＠lcrc.anl.gov 10 Sep '14

10 Sep '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Alejandro Lopez Bezanilla Project Name: STOandPlus Division: MSD Project title: Ab initio calculations for transition metal oxides Associated funding: DOE Other Systems: MIRA, 5M hours Science: Within this project we have simulated the magnetic and electronic properties of oxygen-deficient SrTiO3. A paper was submitted ( http://arxiv.org/abs/1408.3103 ) with a complete report on the importance of density of oxygen vacancies to observe localized magnetic states and polaronic effects. The next step will be to determine the effect of O di-vacancies computing the electronic structure of large STO supercells to observe the interplay between defects and magnetism when vacant sites are both isolated or coupled. Project description: We will intensively use DFT codes such as VASP, SIESTA and OpenMX. Due to the large size of the unit cells containing over 800 atoms, jobs requesting up to 32 nodes in Blues during several hours will be launched. The experience acquired during the previous similar study allows us consider Blues and Fusion as the most appropriate for this type of calculations. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 450000 Q1: 100000 Q2: 125000 Q3: 100000 Q4: 125000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for UNEDF_Optimization
by accounts＠lcrc.anl.gov 10 Sep '14

10 Sep '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Jason Sarich Project Name: UNEDF_Optimization Division: MCS Project title: Optimization for the NUCLEI SciDAC-3 Associated funding: DOE ASCR/NP/NNSA SciDAC-3 project NUCLEI DOE ASCR SciDAC Institutes FASTMath and SUPER Other Systems: Science: The NUCLEI (NUclear Computational Low-Energy Initiative) project will build upon recent successes in large-scale computations of atomic nuclei to provide results critical to nuclear science and nuclear astrophysics, and to nuclear applications in energy and national security. The large-scale computations we envision will transform the fields of low-energy nuclear physics and astrophysics. Physics topics to be addressed include nuclear interactions and their uncertainties, ab-initio studies of light nuclei and their reactions and of nucleonic matter and its astrophysical properties. We will also fundamentally advance the studies of neutron-rich nuclei and the fission of heavy nuclei, and the key nuclear physics issues in neutron stars and tests of fundamental symmetries. Project description: We are building upon our optimization successes under the SciDAC-2 UNEDF (Universal Nuclear Energy Density Functional) project in the successor project NUCLEI. In this project, optimizations over parameter spaces corresponding to energy density functional parameters will be performed using the Argonne-developed POUNDERS optimization algorithm. We will work with DFT codes HFBTHO and HFODD and begin work with ab initio codes such as MFDn. These physics codes will be run using parallel resources, in some cases with each of 100 nuclei being run on a separate node, in others (such as with MFDn), a simulating a single nucleus will require up to hundreds of nodes. In the past we have used OpenMP at the nucleus level in concert with library and compiler-based optimizations to bring the wall-clock time of a single by a factor 10. We will continue such code optimizations for the next-generation codes developed under NUCLEI. We anticipate using 40k core-hours for this task in FY15. For our parameter optimizations, we typically have 10-20 parameters and require on the order of 300 simulations. Based on the current simulation time, we would thus require 120 8-core nodes for up to 6 hours, or 6k core-hours for each version of the functional we want to optimize. We expect the number of core-hours will remain the same when using the new 16-core nodes. Past experience has let us to expect to do roughly 40 such optimizations in FY15, for a total of 240k core-hours for this task. POUNDERS is available in the Toolkit of Advanced Optimization (TAO), but we still prototype specialized versions of the algorithm using MATLAB/Octave on a single, master node. We will test updates to the TAO version on Blues and compare the performance and correctness against the MATLAB/Octave versions. As in the past, we will have roughly 4 users for this project. Industry partnership: Project URL: http://computingnuclei.org Current FY Hours Used: undetermined amount New FY Requested allocation: 280000 Q1: 70000 Q2: 70000 Q3: 70000 Q4: 70000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for ECG_LiAir
by accounts＠lcrc.anl.gov 10 Sep '14

10 Sep '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Mike Welland Project Name: ECG_LiAir Division: MSD Project title: Mesoscale modeling of electrochemical crystal growth for Li-air batteries Associated funding: Atomistically Informed Mesoscale Modeling for Advanced Electrical Energy Storage Systems Other Systems: Science: This work involves mesoscale simulations of oxide growth in Li-air batteries by modeling multicomponent mass transport, electrostatics, structural mechanics, and microstructure evolution. We are developing a new continuum scale model that considers large variations in constituent concentrations between phases, concurrent diffusion on substitution and interstitial lattices, and an original phase-field formulation. The study encompasses electrochemical crystal growth / dissolution as well as nanoscopic interfacial effects such as the electric double layer. In application to realistic systems, we seek to understand the factors influencing the performance, longevity and behavior of Li-air and similar advanced battery concepts. Project description: The model is derived from a thermodynamic description of the system as a functional of the concentration of ionic species (Li+, O2-, A- and e-) and a neutral solvent, the electrostatic potential, the local phase, and the local pressure. Working through the theory of irreversible processes, we can describe the electrochemical growth / dissolution process in a system of at least eight highly non-linear, coupled partial differential equations which must be solved numerically via the finite element method. We have selected the FEniCS package for this purpose, which is an open-source project developed by teams internationally and relies on the PETSc library for calculation, thus taking advantage of the latter's scalability and efficiency on computer clusters. We previously developed a phase-field formulation with a newly developed technique of calculating extreme concentration variations, on the order of 10^23 over a simulated interface width of 1Å, in a numerically robust fashion [1]. Figure 1 shows a simulations of the ideally polarized electrode which naturally produces the electric interfacial double layer, a distinctly mesoscopic phenomenon and an important interfacial phenomenon for electrical energy storage systems. For this work, it was required to build the desired FEM software package, FEniCS, on Blues. This was completed and is now being used in two projects by the current author (ECG-LiAir and Meso), and also by a collaborator. Qualitative experience shows good scaling on Blues up to 256 cores, which is the number used in the currently described simulations, and the software has been demonstrated to scale well to significantly larger systems due in part to its reliance on PETSc for linear calculations. The use of Blues greatly accelerated the development cycle for exploration of 2D and 3D systems. In collaboration with lower length scale investigations, a mechanism of oxide growth was identified as resulting from reaction of Li+ and O2- in the electrolyte to form solvated LiO2. The solvated LiO2 establishes equilibrium with particles of amorphous LiO2 and may lead to their growth. Meanwhile, a solid-state phase transformation is considered to occur inside the LiO2 particle to form Li2O2. Industry partnership: Project URL: http://cmcsn.phys.washington.edu/content/computational-microstructure-scien… Current FY Hours Used: undetermined amount New FY Requested allocation: 400000 Q1: 100000 Q2: 100000 Q3: 100000 Q4: 100000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for Meso
by accounts＠lcrc.anl.gov 09 Sep '14

09 Sep '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Olle Heinonen Project Name: Meso Division: PSE Project title: Mesoscale computational materials science Associated funding: LDRD, BES (FWP 59001-00-105) Other Systems: serafina cluster (MSD), BG/Q, CNM cluster Science: - Simulations of mesoscale coupled elastic systems - Oxygen defect motion in metal/oxide/metal heterostructures - Magnetization behavior and dynamics in nanostructured magnetic systems. Project description: In the first project, mesoscale coupled elastic systems, we are developing new methods and codes (within the MOOSE framework) as well as using the FEniCS code to more finite mesoscale systems in which elastic degrees of freedom are coupled to others, such as concentration or bound and free electric charges. Both the MOOSE and FEniCS codes use MPI and threading to parallelize and both have exhibited strong scaling up towards 100k cores. In these simulations we may use up to 512 cores for production runs. These codes both use PETSc, and MOOSE also uses LibMesh. We compile our own versions of these libraries. In the second project, oxygen defects, we are using a variety of first-principle electronic structure codes, principally SIESTA and FHI-aims, to calculate electronic properties and oxygen diffusion barriers (using nudged elastic band methods). For the heterostructures, we use the non-equilibrium transport code SMEAGOL to calculate the electronic transport. Our developer SIESTA/SMEAGOL version uses threading and MPI for parallelization and we have demonstrated scaling of up to ~64k cores. The FHI-aims code uses MPI only and is a very efficient codes. On blues, we expect to run up to 512 cores for the calculations; for much bigger systems, we are using Mira. We also have a collaboration with Subramanian Sankanayaranyan (CNM) who does MD simulations of interfaces that we feed into the electronic structure calculations. The third project, magnetization behavior, uses a home-grown code that uses OpenMP threading only, so this one runs only on up to 16 cores at a time per job but typically runs at close to 95% efficiency on 16 cores. The number of project members is five to six. 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: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for LiBattModeling
by accounts＠lcrc.anl.gov 09 Sep '14

09 Sep '14

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: Applied Battery Research Program, Office of Vehicle Technologies, U. S. Department of Energy Other Systems: NERSC (about 250K core hrs.) Science: First principles local-density-functional-theory methods are applied to calculate physical and electrochemical properties of lithium-ion-battery-electrode materials. This work complements an experimental program in the Battery Department of CSE to develop high capacity, high energy density cathode materials. Project description: We will perform simulations to evaluate substitutions such as Mo and Ru to partially replace Mn in the Li2MnO3 component of xLi2MnO3•(1-x)LiMO2 composites. The simplest choices of M are M=Co and M=Mn0.5Ni0.5. Relevant replacement fractions may lie in the range y=0.25 to 0.50; higher values may result in a structure significantly different from that of xLi2MnO3•(1-x)LiMO2. A. Stability. Simulations will be performed to elucidate the structure and stability of Li2(M’yMn1-y)O3, both separately and in xLi2(M’yMn1-y)O3•(1-x)LiMO2 composites. Of interest, for example, would be determination of the pseudo-binary phase diagram of Li2(M’yMn1-y)O3 in the monoclinic structure. Although a full phase diagram calculation (e.g., by the cluster expansion method) would require a huge computational effort, perhaps some features of the phase diagram can be established with only moderate effort. B. Migration of M’. Since the migration from the metal layer to the lithium layer in the activated standard material xLi2MnO3•(1-x)LiMO2 leads to degradation, stability of M’ against such degradation is desired. Simulation of the energy along a migration path between octahedral and tetrahedral sites will provide an indication of the stability of M’ candidates in tetrahedral sites. C. The performance of M’ in xLi2(M’yMn1-y)O3•(1-x)LiMO2 cathodes depends on the redox energies associated with M’. Simulations of the voltage as a function of delithiation for Li2(M’yMn1-y)O3, both as an isolated system and as part of a composite, will provide an estimate of the capacity and energy density as a function of upper charge voltage. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 800000 Q1: 200000 Q2: 200000 Q3: 200000 Q4: 200000 Justification: The VASP software employed in this work is known to provide relatively good scaling, for the types of simulations proposed, up to a number of cores roughly equal to the number of atoms in the periodic cell. Typically, our cells include up to about 400 atoms. Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for Lattice-QCD
by accounts＠lcrc.anl.gov 08 Sep '14

08 Sep '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Donald Sinclair Project Name: Lattice-QCD Division: HEP Project title: Lattice simulations of Conformal and Walking Technicolor. Associated funding: DOE Other Systems: NERSC Cray XC30, "Edison" NERSC Cray XT6, "Hopper" NERSC Cluster, "Carver" Current NERSC allocation: 35,000,000 MPP hours TACC cluster, Stampede. Allocation 3,000,000 core-hours. Science: We are using our simulation methods developed for lattice QCD to study extensions of the standard model of High Energy Physics, in which the Higgs sector is strongly interacting, and the Higgs fields are composites. This has become all the more relevant with the discovery of a Higgs-like particle at the LHC at CERN. We are particularly interested in QCD-like theories whose pions play the role of the Higgs field in giving masses to the W and Z weak vector bosons. Such theories are called Technicolor theories. We are particularly interested in Walking Technicolor theories, whose slowly evolving couplings evade some of the technical difficulties associated with the phenomenology of Technicolor and Extended Technicolor theories. Our initial studies are of the thermodynamics of such theories, since these provide a simple way of measuring the scales of confinement and of chiral symmetry breaking. From such simulations, we can determine if confinement and chiral-symmetry b reaking survive the continuum limit. In this case, our chosen theory is QCD-like. If not the theory is a conformal field theory. We choose theories which 2-loop perturbation theory suggests have an infrared fixed point and are thus conformal. If chiral symmetry breaks spontaneously before this fixed point is reached, the fixed point is avoided and the theory is ultimately QCD-like. However, the close proximity of the fixed point means that there is a region where the running coupling constant evolves very slowly, i.e. the theory walks. The candidate Walking-Technicolor theory which we are studying is 'scaled up' QCD with 2 colour-sextet quarks. QCD with 2 sextet quarks has just the right number of Goldstone 'pions' (3) to give masses to the Ws and Z. It remains an open question as to whether it has a light Higgs. Because it is difficult to determine the nature of this theory, it is useful to study a closely related theory with known behaviour. We study QCD with 3 sextet quarks which is almost certainly conformal for comparison. We have a second project starting this year, namely simulations of Lattice QCD at finite quark-number chemical potential mu, using Complex Langevin methods. This is relevant to the understanding of the phases of nuclear-matter at zero and finite temperature. Cold nuclear matter is found in neutron stars. Hot nuclear matter (quark-gluon matter/plasma) is believed to be produced in relativistic heavy-ion colliders, and was certainly present in the early universe, Since the integrand of the partition function, which describes the quantum dynamics of QCD at finite mu, is complex, standard simulation methods which rely on importance sampling fail and we turn to complex Langevin simulations. Because these require using a non-compact extension of the gauge group, runaway solutions are common, and these caused earlier use of this method to fail. Recent advances indicate that at least some of these problems were associated with large gauge excursions from well-behaved gauge configura tions, and can be circumvented by judicious choice of gauge (gauge-cooling) after each field update. Project description: We perform simulations to evaluate the functional integrals of QCD-like theories formulated on a discrete space-time lattice, to enable determination of the non-perturbative aspects these theories. These include the properties of these theories at non-zero temperature, including the scales of confinement and chiral symmetry breaking, and such zero temperature properties as spectra (including the Higgs mass), decay constants and the running of the gauge coupling constant. We are particularly interested in those theories where the coupling constant evolves very slowly, since these are candidate 'Walking Technicolor' theories. Related to these are theories with an infrared fixed point (conformal field theories). Our first goal is to differentiate between these two different types of behaviour for candidate theories. We are performing simulations of QCD-like theories which are models for Walking or Conformal Technicolor. We have been studying theories which are essentially QCD but with colour-sextet rather than colour-triplet quarks. We hope to measure the running of the QCD coupling constant. For 2 or 3 flavours, 2-loop perturbation theory predicts an infrared fixed point. For 2 flavours, it is possible that a chiral condensate forms before this fixed point is reached. If so, the fixed point is avoided, the theory is confining, and chiral symmetry breaks spontaneously. However, there is a region where the coupling constant evolves very slowly. These are the properties required for a walking technicolor theory. Simulations we have performed so far at finite temperature suggest that this theory might walk. So far the results are inconclusive. For 3 flavours we know that the theory should be conformal. Simulations at N_t=6 (12^3 X 6 lattice) and N_t=8 (16^3 X 8 lattice), some of which used Fusion, did not yet show evidence of conformality. We are thus moving to N_t=12 (24^3 X 12 lattice). We started these simulations on Blues, and intend to continue them in FY2015 on Blues, while performing lower mass simulations on Edison at NERSC. Our sextet quark codes are based on our earlier triplet quark codes and use the RHMC simulation method. The Rational Hybrid Monte Carlo (RHMC) is a stochastic molecular dynamics algorithm. The functional integral of QCD is written as a partition function of a classical field theory evolving in a fictitious time. The determinant of the Dirac operator raised to a fractional power is calculated by introducing bosonic fields (pseudofermions), and sandwiching this Dirac operator raised to minus said fractional power between them. This fractional power of the Dirac operator is approximated to machine accuracy by a rational approximation. After defining this this theory on a discrete space-time lattice, the inversions required by the partial-fraction expansion of the rational approximation are performed using Krylov space methods, in particular a multi-shift extension of the conjugate gradient algorithm. A global Metropolis Monte-Carlo accept/reject step applied at the end of each trajectory removes discretization errors introduced by the numerical integration of these stochastic equations of motion. We parallelize the code by assigning a fixed number of adjacent lattice sites to each MPI task. Network bandwidth ultimately limits how small a chunk of the lattice can be assigned to each task. Our 24^3 X 12 runs will be performed on 288 cores (18 nodes) of Blues. A short benchmark run of the 24^3 X 12 code yielded 667 Gflops or 2.3 Gflops/core. Since we have seen performances of > 3 Gflops/core on machines with similar processors, we suspect that at least 3 Gflops/core should be achievable on this machine. Since we have exhausted our allocation on Blues, we cannot perform more extensive tests. Scaling tests on Fusion and on Edison at NERSC for numbers of cores from 24 up to 288 cores indicate that the per core performance increases up to 144 or 288 cores, which we believe to be related to cache usage. Our new project, simulating QCD at finite mu using Complex Langevin with Gauge Cooling, will require some code development. Our old codes do not as yet have gauge cooling and our parallel code uses SHMEM rather than MPI for communication. Since the real Langevin is the limit of our Hybrid Molecular Dynamics code (the forerunner of the RHMC method) in which there is a single update per trajectory, most of what we have discussed for the RHMC applies, and will not be repeated here. We plan to use Blues and Carver at NERSC for the early, small-lattice simulations, 8^3 X 4 on 32 cores and 8^4 on 64 cores. Because of the smaller memory requirements, allowing more efficient cache usage, we expect to get a minimum performance of 3-4 Gflops/core for this code. Our requested allocation is based on running 1 288 core job half of the time at times when the NERSC machines are experiencing slowdowns for amall jobs. Industry partnership: Project URL: http://www.hep.anl.gov/dks Current FY Hours Used: undetermined amount New FY Requested allocation: 1200000 Q1: 300000 Q2: 300000 Q3: 300000 Q4: 300000 Justification: We have not yet performed a detailed scaling analysis on Blues, however, we have such analyses performed on Fusion as well as Edison at NERSC. For Fusion, for our QCD with sextet quarks code on a 24^3 X 12 lattice running on Fusion we observed the following performances: 24 cores = 45 Gflops = 1.9 Gflops/core 48 cores = 95 Gflops = 2.0 Gflops/core 72 cores = 148 Gflops = 2.1 Gflops/core 96 cores = 220 Gflops = 2.3 Gflops/core 144 cores = 369 Gflops = 2.6 Gflops/core 288 cores = 784 Gflops = 2.7 Gflops/core For the same code and lattice size running on Edison at NERSC we observed the following performances. 24 cores = 53 Gflops = 2.2 Gflops/core 48 cores = 115 Gflops = 2.4 Gflops/core 72 cores = 191 Gflops = 2.7 Gflops/core 96 cores = 245 Gflops = 2.6 Gflops/core 144 cores = 510 Gflops = 3.5 Gflops/core 288 cores = 959 Gflops = 3.3 Gflops/core We use a custom assignment of tasks to nodes in order to minimize communications. Earlier recoding reduced the number of global reductions in the routines, which use most of the CPU time, by a factor of 2. The node codes are transparently vectorizable, to allow use of SSE instructions. Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Request: POM
by accounts＠lcrc.anl.gov 08 Sep '14

08 Sep '14

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: Baofu Qiao Applicant's institution: ANL Applicant's division: CSE Project Name: POM Project title: Macroions Aggregation on the Water/Air Interface: Atomistic Studies Associated funding: DOE-BES Other Systems: There is one additional computer cluster available, i.e., the TARDIS at Prof. Monica Olvera de la Cruz’s group at Northwestern University. Dr. Qiao had been working at Prof. Olvera de la Cruz's group for around 3 years, and has used more than 1,000,000 CPU hours on TARDIS. The current project on the interfacial aggregation behavior on water/air interface will be under continuous collaboration with Prof. Olvera de la Cruz. Science: Studying interaction between ions near air-liquid interfaces is of immense importance for the understanding of numerous processes in biology, chemistry, atmospheric and geological sciences. Recent advances in experimental techniques in probing liquid surfaces like Liquid surface X-ray scattering using synchrotron X-rays, non-linear optical measurements and photoelectron spectroscopy are providing new insights about the distribution of ions near air-liquid interfaces. Molecular dynamics simulations are also adding new dimensions which complement the experimental observations. We recently performed some experiments on looking at the distributions of hetero-polyanions at air-water interfaces. The experiments show counter-intuitive phenomena of charge dependent condensation of hetero-polyanions near air-water interface to form an ordered monolayer. Comparing the monolayer structure with the bulk crystal structure of the hetero-polyanions, we inferred tha t monolayer is stabilized near the interface by network of hydrogen bonds with the protons (H9O4) with a planar structure. These observations lead us to ask some fundamental questions like: a) What is the pH near air-water interface? b) How protons approach air-water interface? c) Effect of ions on the proton structure near the interface? These questions are still in a state of controversies. Answering the questions like these requires a holistic approach in viewing the structure near air-liquid interfaces by combining time-averaged experimental results with atomistic molecular dynamics simulations. Experimentally it is very difficult to probe protons due to their low electron density contrast but the interaction mediated by them to form ordered structures by ions with high electron density can be measured precisely. We propose to simulate the ordered structure of the ions in the presence of protons with various known structures like H3O, H5O2, H9O4, etc. and match with experimentally observed structures. This approach of combining the experiments with simulations would help us to get the complete understanding of interaction between the ions near air-water interface from atomistic to macroscopic scale. Apart from the protons, in-general the interaction of the ions with their counter-ions play crucial role in their ordering both in the bulk as well as near the interfaces of ionic-solutions. Starting with the protons as counter-ions we propose to study the ionic-environment with the solutions of 1) Hetero polyacids and their salts 2) Rare-earth and transition metal salts The understanding from these systems would lead us to take the first step towards answering the fundamental issues in solvent extraction and separation sciences in general. Project description: In this project, we will investigate the aggregate structures of POM macroions on water/air interface. In particular, we are planning to investigate the influences of POM concentration and the different proton models (H3O+, H5O2+, H9O4+). We are aiming to unveil the mechanism of the aggregation behaviors of macroions on the water/air interface. Such studies are expected to provide significant insights for the separation of metal ions. In this project, classical molecular dynamics (MD) simulations will be performed at the all-atom resolution, complementing the corresponding experiments by Dr. Mrinal Kanti Bera at the CSE division at ANL. All-atom MD simulations have been well proven in predicting supra-molecular aggregate structures. The open source package GROMACS will be employed for the simulations. GROMACS is one of the highly recommended packages for MD simulations. GROMACS has been reported to be able to scale well up to around 1,000 CPU processors. In this project, system size up to 20 × 20 × 10 = 4000 nm3 will be simulated, which will include around 400,000 atoms. Similar systems have been previously performed using the TARDIS cluster, which showed calculation speed of around 500 CPU hours per nanosecond simulation time using a single CPU. GROMACS shows very good scalability on TARDIS, where a peak number of CPU of around 100 CPU has been used due to the limited size of the TARDIS cluster. Two users will use this allocation time, Dr. Baofu Qiao and Dr. Mrinal Kanti Bera, both at the CSE division. Industry partnership: Project URL: Requested allocation: 400000 Q1: 100000 Q2: 100000 Q3: 100000 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

1 0