August 2011 - Allocations-admins - mailman3.cels.anl.gov

[LCRC Accounts] Yearly Allocation Request from ScalaGAUSS
by accounts＠lcrc.anl.gov 31 Aug '11

31 Aug '11

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Jie Chen Project Name: ScalaGAUSS Division: MCS Project title: Scalable Gaussian Process Data Analysis Associated funding: Scalable Gaussian Process Analysis of Spatio-Temporal Data (PI: Mihai Anitescu) Other Systems: MCS workstations Science: This project uses a maximum likelihood approach to perform Gaussian process analysis on both synthetic and real data. The observations considered for analysis are simulation and data from nuclear engineering and climate science. The resulting analysis is used to describe the spatial correlation of the sampled data, to predict and to understand uncertainty in very high fidelity phenomena. Project description: THIS PROJECT WAS STARTED CLOSE TO THE END OF FY2011 AND ONLY A SMALL NUMBER OF CORE-HOURS WERE ALLOCATED. WE ARE REQUESTING AN ADDITIONAL AMOUNT OF 500000 CORE-HOURS FOR FY2012. We model data as Gaussian process model using several covariance models, including a Matern covariance kernel. The goal of the computation is to estimate the parameters of the covariance kernel. We employ a maximum likelihood method, which is reformulated by using a sample average approximation approach. The novelty of the approach is that it allows for scalable analysis without compromise in the accuracy of the correlation. This reduces to solving linear systems that have the same size of the grid from which the data are sampled. The linear systems have multiple right-hand sides, and in the case of a regular grid, they are multilevel Toeplitz matrices. We employ a block CG solver with multilevel circulant preconditioners. Previous experiments indicate a linear scalability of computational costs for 2D grids of size up to $10^3times10^3$ on a single desktop machine in a Matlab environment. We have demonstrated that up to 1 million data points our approach scales virtually pe rfectly. We have developed parallel C++ codes based on the Trilinos project and will investigate the scalability of the algorithm for 4D data set with on the order of $10^9-10^{12}$ data points; beyond the $10^6$ data sets we can manipulate at the moment. For being able to compile our code, we need (in addition to standard compilers) Trilinos, and ideally also PETSC. The ScalaGauss code can be used in numerous other applications (chemistry, power grids, materials science) that require a Gaussian process statistical models. Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 499999 Q1: 125000 Q2: 125000 Q3: 125000 Q4: 124999 Justification: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Request: pnipam_agglomeration
by accounts＠lcrc.anl.gov 30 Aug '11

30 Aug '11

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: Sanket Deshmukh Applicant's institution: ANL Applicant's division: CNM Project Name: pnipam_agglomeration Project title: Agglomeration dynamics in thermo-sensitive polymers across the lower critical solution temperature: A molecular dynamics simulation study Associated funding: LDRD Other Systems: Carbon (CNM) Science: PNIPAM is a thermo-sensitive polymer that has a lower critical solution temperature (LCST) around 305K in aqueous solution.1 As the LCST of PNIPAM is close to human body temperature, it is a serious contender for a controlled drug delivery system.2,3 Coil-to-globule transition occurs in PNIPAM when temperature is raised above the LCST. It is believed that this occurs due to the change in hydrophilic interaction to the hydrophobic interaction.1 Concentration and molecular weight of PNIPAM affects the LCST by less than 1K (304–306K). Lower molecular weight samples appear to have slightly higher LCSTs. Presence of salts ions in the solution lowers the LCST significantly.1,4 It is believed that, when PNIPAM is dissolved in solvent like water, PNIPAM has to arrange it self in a specific orientation to form hydrogen bond with the already somewhat arranged water molecules.1 In the case of water molecules, which are near to hydrophobic group of polymer, -CH3 of the PNIPAM side chains, these water molecules must reorient being unable to form hydrogen bond with the non-polar groups in PNIPAM. This reorientation leads to formation of clathrate-like structures, this phenomenon is known as hydrophobic effect, results in decreased entropy upon mixing (negative ΔS).1 When temperature is raised, the entropy term dominates the otherwise exothermic enthalpy of the hydrogen bonds formed between the polymer polar groups and water molecules that is the initial driving force for dissolution. Once the free energy change (ΔG) becomes positive upon mixing, the consequence is phase separation above a LCST as seen in the case of PNIPAM.1 If the concentration of the PNIPAM is high enough, this replacement of PNIPAM-water contacts with PNIPAM-PNIPAM and water-water contacts is manifested by precipitation.1 Since the discovery, that PNIPAM changes its solubility in water as a function of temperature, numerous reports deal with the origin of the change from solubility to insolubility at elevated temperatures.5-7 Based on the study of diluted aqueous PNIPAM solutions by static and dynamic light scattering (DLS), a coil-to-globule transition caused by dehydration of polymer chains during heating was postulated.6,7 For concentrated aqueous PNIPAM solutions, phase separation at LCST occurs as a macroscopic manifestation of the coil-to-globule transition followed by aggregation.8 To validate this theory and to understand the exact atomistic scale origin of the phase transition at the LCST, we propose to use MD simulation techniques. Two PNIPAM chains, consisting of 30 monomer units each, will be placed in a simulation cell and will subsequently be solvated. We plan to carry out simulations of two 30-mer 20 Å and 50 Å apart from each other. Simulation cell will be solvated with ~24,000 water molecules. Simulations will be carried out below and above the LCST, namely at 278, 295, and 310K for 50ns. Simulated trajectories will be analyzed for structural and dynamical properties of PNIPAM (e.g. radius of gyration (Rg), radial distribution functions (rdf)) and diffusion properties of water. The agglomeration and coil-to-globule transition, if any, will be studied at 278, 295, and 310K. Project description: Accuracy of the model and MD simulation results mainly depends upon the method of structure generation and force-field used for MD simulations.9 In this work we propose to use the Polymer Consistent force field (PCFF) to understand the structural changes in the polymer brush at global as well as local level using an atomistic level model.10 The preliminary studies we carried out with PCFF on a single chain made up of 30 monomer units of PNIPAM in presence of pure water predicted the LCST close to 305K, which is in good agreement with experimentally observed value. To carry out MD simulations we will be using LAMMPS MD simulation package. Two 30-mer will be placed in simulation cell (20 Å and 40 Å apart from each other). Simulation cell will be solvated with ~24,000 water molecules (72,000 atoms). Simulations will be carried out below and above the LCST, namely at 278, 295, and 310K for ~50ns.Transport properties of water will be studied using Einst ein relation. Results will be compared with the ongoing experiments in the APS. The computational requirements for the proposed study will be substantial. For the proposed calculations, to study the phase transition across the LCST simulations should be run for ~50ns. From preliminary results of the simulations we carried out on the Fusion cluster, we observed a clear phase transition across the LCST, for polymer chains with 30 monomer units (572 atoms) in presence of 9,000 water molecules. This system with ~30,000 atoms required ~135 hrs of simulation time for 10ns on 96 cores. In the proposed work we plan to study, two 30-mer chains of PNIPAM in presence of ~24,000 water molecules (~72,000 atoms). All the simulations will be carried out over a temperature range across the LCST, namely 278, 295, and 310 (3*4 = 12 simulations). Each simulation will be run for ~50 ns (total simulation time = ~600ns). The estimated time for this particular work, to carry out the mentioned calculations, would be ~750,000 core hours. References: 1. Y. Tsujii, K. Ohno, S. Yamamoto, A. Goto, T. Fukuda, Adv. Polym. Sci. 2006, 197, 1-45. 2. D. I. Dimitrov, A. Milchev, K. Binder, J. Chem. Phy. 2007, 127, 084905-084909. 3. H. G. Schild, Prog. Polym. Sci. 1992, 17, 163-249. 4. D. Schmaljohann, Adv. Drug Del. Rev. 2006, 58(15), 1655-1670. 5. American Cyanamid Company 1963, N-Isopropylacrylamide, Brochure C3-1354-500-4/63. 6. K. Kubota, S. Fujishige, I. Ando, J. Phys. Chem. 1990, 94, 5154. 7. C. Wu, S. Q. Zhou, Macromolecules 1995, 28, 8381. 8. K. Chan, R. Pelton, J. Zhang, Langmuir 1999, 15, 4018. 9. M.P. Allen, D. J. Tildesley, Computer Simulations of Liquid, Clarendon Press, Oxford, 1987. 10. Accelrys, Inc. Cerius2 and Discover (programs) and Discover User Guide, version 96.0/4.0.0; Molecular Simulations, Inc.: San Diego, CA, 1996. Project URL: Requested allocation: 750000 Q1: 550000 Q2: 200000 Q3: 000000 Q4: 000000 Justification: To study the phase transition across the LCST in the case of single PNIPAM chain, simulations should be carried out for ~50ns. From preliminary results of the simulations we carried out on the CNM’s cluster Carbon, we observed a clear phase transition across the LCST, for polymer chains with 30 monomer units (572 atoms) in presence of 9,000 water molecules. This system with ~30,000 atoms required ~145 hrs of simulation time for 10ns on 96 cores. On the Fusion cluster same calculations took ~135 hrs. 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] Project Request: nachbac
by accounts＠lcrc.anl.gov 29 Aug '11

29 Aug '11

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: Benoit Roux Applicant's institution: ANL Applicant's division: BIO Project Name: nachbac Project title: selectivity of the sodium channel Associated funding: NIH R01-GM062342 Other Systems: Science: Understanding quantitatively the microscopic factors giving rise to Na+ selective channels is mysterious. For a long while, we had only K+ selective channels (such as KcsA). Just recently, the X-ray structure of a Na+ selective channel was determined. It is related to the NachBAC channel (Na+ Channel from bacteria). Project description: The most informative strategy to address the issue of Na+ selectivity in NachBAC is to compute the multi-ion PMF corresponding to the microscopic process in question (Berneche and Roux, Nature 2001). In particular, the calculation of the PMF enables us to pin-point the location of the largest barrier opposing the passage of a single K+ ion relative to Na+ in the channel, thus explaining while the channel becomes selective from the free energy profile. We can carry those type of calculations using our new and highly scalable method called umbrella sampling with replica-exchange MD simulations. This is a competitive projects as others are racing to produce calculations. This is why our request is biggest at the early stage of the project. Project URL: Requested allocation: 1000000 Q1: 500000 Q2: 250000 Q3: 150000 Q4: 100000 Justification: The MD umbrella sampling with replica-exchange is a new method that was recently implemented by Wei Jiang. On BG we are able to run such computations up to 32 racks on similar systems. The scope of this project is smaller and on fusion, we plan to use the resource to carry out the calculation as efficiently as possible. 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] Project Allocation Request
by accounts＠lcrc.anl.gov 26 Aug '11

26 Aug '11

Hello, A change in allocation has been requested: Requester: zapol (Peter Zapol) Project: nanocatalysis Title: Catalysis in nanostructured materials Description: Molecular dynamics studies using many-body potentials as implemented in DL_POLY computer code and DFT studies in VASP will be used to simulate support structures covered by hydroxyl groups. Radial distribution functions will be compared to available experiments. The code uses 32 to 48 PEs for about 50,000 atom calculations. The geometries from molecular dynamics simulations will be used in studies of catalytic activity. We will study reactions on supported metal subnanometer clusters. We will also continue investigations on using thermodynamics to modify shape of catalytic particles, in particular in electroche4mical environment. Density functional studies will be performed using VASP, NWChem and GAUSSIAN codes to determine reaction energies and barriers. The VASP calculations typically use 24-64 processors for 24 hours to 48 hours. Expected number of members in the project is three to six. Current: undetermined amount Justification: We have used our allocation very effectively in the previous year. Requested: 80000 A specific reason has been given: We have received considerably less computing time than requested. In line with our estimates, we need additional time to complete the project. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 26 Aug '11

26 Aug '11

Hello, A change in allocation has been requested: Requester: zapol (Peter Zapol) Project: compnano Title: Computational Nanoscience Description: Computational studies aimed at materials by design will be performed using molecular orbital theory, density functional and molecular dynamics methods as implemented in Gaussian09, VASP, and NWChem computer codes. We have successfully performed NWChem, G09 and VASP calculations on Fusion last year. G09 scales well up to 16 processors. VASP shows good scalability from 16 to 256 PEs as well as NWChem which is highly suitable for massively parallel computations. We will also investigate how other codes scale on Fusion and benchmark performance. We will continue our studies of chemical reactivity on the surfaces and interfaces of nanomaterials in FY11. In particular, we will perform DFT calculations coupled to thermodynatic treatments to investigate properties of nanoparticles as a function of their size and shape for metals and metal oxides. We have performed work on computational screening of oxide-based materials for light to energy conversion that resulted in useful predictions for surface modification. We plan to extend these studies to doping with light elements for oxide band gap engineering and transition metals for modifications of catalytic properties. We will extend these studies to different size and shape of oxide particles. In addition, we will work on carbon nanomaterials using them both as support for heterophase nanostructures with novel chemical properties to investigate their potential for electrochemical applications in energy conversion and storage. A typical Gaussian calculation will take for 4 to 16 nodes anywhere from 1-2 hours to several days. VASP and NWChem calculations may involve 16 to 256 nodes and up to several days depending upon the size of the problem. This project will involve 5-7 members of the Molecular Materials Group. Current: undetermined amount Justification: The allocation is used by several postdoctoral and staff scientists to perform calculations on DOE-funded projects. From previous year's experience, we have used efficiently a large allocation, about a third of time for Gaussian and NWChem, and the rest for VASP calculations. Requested: 80000 A specific reason has been given: We have received considerably less computing time than requested and run out of computing time. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 26 Aug '11

26 Aug '11

Hello, A change in allocation has been requested: Requester: wild (Stefan Wild) Project: UNEDF_Optimization Title: Optimization of nuclear-energy DFT functionals Description: This DFT parameter optimization is expensive: each simulation of a single nucleus at a particular parameter set using HFBTHO can consume up to .3 CPU-hours. We have recently used OpenMP at the nucleus level, in addition to compiler-based optimization, to bring the wall time per simulation down to 1.5 minutes on a single node of Fusion. Our current optimizations over 13 parameters typically require 200-300 evaluations of about 100 nuclei, which can be done on Fusion in 5-7.5 hours using 100 8-core nodes. This gives us up to 6K CPU-hours for each version of the functional we want to optimize. We note that our current computational needs for a single optimization, 100 nodes for 7.5 hours, yields a significant improvement in wall time over our previous needs of 12 nodes for 90 hours. These optimizations are being done using the derivative-free optimization software POUNDerS developed at Argonne. For prototyping purposes, this software currently requires MATLAB on one master node. Our team of 3 users anticipate doing 40 optimization runs during this project cycle. We also expect to analyze functionals based on more nuclei and depending on more parameters, potentially resulting in further increases to the number of nodes we require. Current: undetermined amount Justification: Requested: 15000 A specific reason has been given: We had a productive FY11 and have dipped into a negative balance with 6 crucial weeks remaining. User sarich is currently solving calibration problems for a SciDAC project paper. User gawlik is a CSGF student who is testing a new open source PDE-constrained optimization solver and generating scaling results for a paper to be submitted before the close of FY11. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 25 Aug '11

25 Aug '11

Hello, A change in allocation has been requested: Requester: balaji (Pavan Balaji) Project: radix Title: Scalable Parallel System Software Description: This is a Computer Science request for 450,000 cpu-hours on Fusion for development and evaluation of a family of tools and libraries with a direct benefit to all users of high-performance computing resources. The Radix group in Argonnne's Mathematics and Computer Science division conducts research and development on facilities to better enable efficient Parallel Programming, I/O, and Scientific Understanding. The Fusion system represents an attractive milestone for our software development: projects start out on our breadboard cluster and can further explore scalability in the presence of a high speed network on Fusion before making the jump to BlueGene-scale parallelism. Radix experiments often are concerned with latency. The infiniband network on Fusion provides an excellent low-latency environment for message passing and storage management research while also providing ample bandwidth for I/O and message bandwidth testing. For FY2011 our Messaging middleware experiments include the usual MPICH2 development as well as application-oriented libraries like ADLB, a load-balancing library. Fusion hosts two quality parallel file systems, GPFS and PVFS. The wider HPC community often asks the Radix file system researchers to compare these file systems head-to-head. Fusion provides us the opportunity to do just that. For FY 2011 our file system research extends into data storage approaches suitable for exascale. Our research into object storage protocols, including high performance access and replication, fits well on Fusion, with the large amount of local storage on compute nodes. Hosting two parallel file systems allows us to do more than just drag-race (so to speak). The two file systems have distinct characteristics, and act differently in the broader role of the I/O software stack. On Fusion, we can utilize a single software stack (application, parallel-netcdf, ROMIO), and compare the impact of a change in the underlying file system on overall application behavior. We know from other research, for example, that the two-phase optimization in ROMIO needs to adapt to the underlying file system, and that aligning variables to file system boundaries can yield performance improvements for parallel-netcdf. In FY 2011 we are beginning research into new high level I/O libraries. We anticipate Fusion and Surveyor will be our two main test platforms. Fusion compute nodes actually look fairly attractive to active storage research. These nodes have enough disk space to store non-trivial datasets, and powerful enough processors to carry out complex computations without much impact on storage performance. Specific goals for 2011 with respect to Active Storage include deploying a Hadoop (MapReduce) layer on top of PVFS. Radix projects also focus on deriving insight both about application behavior as well as scientific insight. 8 core nodes make efforts like in-situ visualization, where for example an application renders a frame of a movie, ever more feasible. Even without fancy graphics accelerators, the processing power on Fusion nodes makes such analysis and visualization feasible, especially if the visualization can take advantage of multiple cores and multiple nodes. Naturally, the radix visualization tools have demonstrated scalability to Fusion sizes and beyond on BlueGene, so making full utilization of Fusion should not be a concern. As should be evident from the variety of planned experiments, the Radix group plans to make full use of Fusion. With Argonne employees and student collaborators, we expect around 20 members in FY2011. We plan on about 0% of our jobs being single core jobs. It's crazy that you even have to ask! While Fusion may contain a modest number of compute nodes, the machine still represents an attractive platform for testing: very fast, low latency interconnect; 8 cores per node; .25 TB of storage per compute node. The testing and experiments the Radix group can carry out on Fusion ensures the tools the group develops today will remain relevant even as supercomputers grow. Quite a few radix-developed projects are part of the Fusion software stack: CPU hours for this project yield improvements and benefits not just for the radix group, but for all users of Fusion and indeed users of high-end computational resources worldwide. Current: undetermined amount Justification: Requested: 300000 A specific reason has been given: We just received a reasonably high-profile request to optimize MPICH for graph problems for the upcoming graph500 competition. We have started optimizing it, but will need a fusion allocation to test our changes, profile them, and identify what needs to be further optimized. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 24 Aug '11

24 Aug '11

Hello, A change in allocation has been requested: Requester: sinclair (Donald Sinclair) Project: Lattice-QCD Title: Lattice simulations of Conformal and Walking Technicolor. 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, decay constants and the running of the gauge coupling constant. We are particularly interested in those theories where the coupling constant evolves particularly slowly, since these are candidate 'Walking Technicolor' theories. Related to these are theories with an infrared fixed point. Of special interest is to differentiate these two different types of behaviour. 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 does indeed walk. This is an ongoing program at NERSC. We wish to run small-lattice simulations of the 3-flavour theory, which we expect to have conformal, rather than walking behaviour on Fusion, for comparison. If we are correct, the properties of the 2 and 3 flavour theories should look rather different. 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 t rajectory 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. Benchmarks of this code on Fusion for the 24^3*12 lattice planned for some of our FY2011 runs give: 24 cores = 46 Gflops 48 cores = 95 Gflops 72 cores = 148 Gflops 96 cores = 220 Gflops 144 cores = 369 Gflops 288 cores = 784 Gflops We plan to run the 24^3*12 lattice jobs on 288 cores (36 nodes). Our smaller lattice 16^3*8 runs would be run on 128 cores (16 nodes). Current: undetermined amount Justification: This allocation request is based on having one 288 core job or two 128 core jobs running at all times. Our project is expected to use time at NERSC and NICS as well. LCRC has the advantage that it provides facilities for jobs which have medium levels of parallelism such as the 128 and 288 core jobs indicated here, which are discriminated against on larger machines such as Franklin and Hopper(stage 2) at NERSC, and Kraken at NICS. Carver at NERSC provides a similar platform to Fusion, but the large number of users and potential users mean that we cannot rely on it to provide the bulk of our computing at NERSC. Numbers given in the description section indicate that scaling is excellent. We have plans to increase the level of parallelism for this code, but this is probably irrelevant to our Fusion use. For codes which have smaller portions of the lattice on each core, such as our 16^3*8 code, performances of around 3 Gflop/core are sustained. Requested: 20000 A specific reason has been given: Amount of time to run last job of the current series (96 wallclock-hours on 192 cores). This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 22 Aug '11

22 Aug '11

Hello, A change in allocation has been requested: Requester: walkosz (Weronika Walkosz) Project: Metastable_Materials Title: First-principles studies of growth mechanisms for metastable materials Description: We will perform first-principles calculations for the adsorption and diffusion of NH3 on GaN, InN, and In surfaces through reaction intermediates such as NH2, NH, N, and H. The simulations will be performed with Vienna Ab-initio Simulation Package (VASP), using plane wave basis sets and PAW potentials. We will first model surfaces of GaN, InN, and In metal in the presence of NH3 to determine the most stable surface sites/configurations for the absorbed NH3 and its decomposed species. Next, diffusion and dissociation pathways of NH3 on various surfaces will be studied by calculating energy barriers using the Nudged Elastic Band Method. The scale of the proposed calculations exceeds small computer cluster capacity. Therefore, we are requesting the LCRC resources to perform a computational study of NH3 on GaN, InN, and In metal. The speed and reliability of the Fusion cluster will allow us to finish the work in a timely manner. The calculations of adsorption a nd decomposition of NH3 on different surfaces will take approximately 48 hours each when ran on 64 processors. At least 50 different structures will be calculated. The calculations of the reaction energy barriers (NEB), on the other hand, will require about 96 hours each when ran on 80 processors as determined by on our previous experience with these types of calculations on semiconductors. At least 42 reaction barriers need to be calculated to validate a diffusion model. Therefore, a total allocation of ~ 475,000 processor hours will be needed to perform the proposed calculations. Current: undetermined amount Justification: Requested: 300000 A specific reason has been given: I ran out of units. I am in the middle of my project (which involves a collaboration with the experimental group) and it is crucial that I run jobs systematically and not fall behind the work. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 12 Aug '11

12 Aug '11

Hello, A change in allocation has been requested: Requester: norris (Boyana Norris) Project: pflotran Title: PFLOTRAN Performance Analysis and Tuning Description: PFLOTRAN is a massively parallel code developed at LANL for the accurate prediction of multiscale subsurface processes. PFLOTRAN has many applications and guides safer industrial techniques; for example, it can help geoscientists determine how carbon dioxide flows through aquifers and deep geologic formations during carbon-capture and storage. PFLOTRAN's performance has been studied extensively on Intrepid and Jaguar (Cray XT/5), but no performance studies on Intel-based clusters have been performed. The efficiency of the code is a small fraction of the peak on both BG/P and Cray XT/5. In the first year of this project, we focused mainly on analyzing and improving the performance of the sequential portions of key computational kernels (matrix-vector product, triangular linear solution, and some of the chemistry computations), which required many single-core runs of these kernels. During the next year, we plan to study and possibly improve the scalability of the Jacobian computation and the linear system solution. We'll also investigate approaches for managing the load imbalance between the flow and transport stages. This will require a number of parallel scaling runs. We will use input decks that can run on different ranges of cores, from one node to the whole machine. We'll target runtimes of less than an hour in most cases (by changing the number of time evolution steps). Current: undetermined amount Justification: Requested: 200000 A specific reason has been given: We request additional time to continue our research on automatically tuning key computations in applications such as PFLOTRAN. Our goal is to advance the state-of-the-art in automatic empirical tuning and adaptation of scientific applications on high performance architectures and thereby significantly increasing scientific productivity. Our current focus is on computational kernels that occur in PFLOTRAN. We have systematically evaluated the strengths and limitations of autotuning techniques using our growing collection of kernels on Fusion. This study has improved our understanding of large-scale auto tuning and also helped us to identify open issues that need to be addressed, which requires developing novel, flexible algorithmic approaches. During the current allocation period, this research has produced two conference/workshop articles. – P. Balaprakash, S. M. Wild, and P. Hovland. Can search algorithms save large-scale automatic performance tuning? in Proceedings of the International Conference on Computational Science, ICCS 2011, Procedia Computer Science, Vol. 4, pp. 2136-2145, 2011. – P. Balaprakash, S. M. Wild, and B. Norris. SPAPT: Search Problems in Automatic Performance Tuning, Preprint ANL/MCS-P1872-0411, April 2011. (will be submitted to ISPASS 2012) This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0