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

New LCRC report submitted for COMPASSOPT
by support＠lcrc.anl.gov 03 Oct '14

03 Oct '14

Full project report at https://accounts.lcrc.anl.gov/reporting/project/2014/?project=COMPASSOPT

1 0

[LCRC Accounts] Yearly Allocation Request for Urban_Climate
by accounts＠lcrc.anl.gov 02 Oct '14

02 Oct '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Robert Jacob Project Name: Urban_Climate Division: MCS Project title: Urban Regional Climate Modeling Associated funding: LDRD Other Systems: Science: We will be using a weather-scale model with atmospheric chemistry to understand the interaction between cities and their regional climate. Urban areas change their own climate most famously by creating an urban heat island (UHI) and by introducing mechanical drag on air flow and moisture transport within the city. They also emit pollutants that have strong negative local effects. But these effects are not confined to the cities themselves. Pollution plumes can contribute to pollution levels hundreds of miles from a city and add to the hemispheric background pollution. Cities can modify storm systems by “splitting” storms as they approach or creating convection through the UHI or directing wind vertically due to the buildings in them. The extent of the impact of urban areas on the mesoscale and regional climate depends on the advection of heat and pollutants outside the city limits and the changes introduced by the UHI on the large-scale weather/climate pa tterns. Thus, there is a need to quantify and model both the urban heat island and its diurnal, seasonal and annual cycles and the flow, especially the vertical transport, of air through, over and around a city. Project description: We'll be using the Weather Research and Forecasting Model (WRF) with options for an urban surface and atmospheric chemistry. The WRF dynamical core is based on an Eulerian solver for the fully compressible nonhydrostatic equations, cast in flux (conservative) form, using a mass (hydrostatic pressure) vertical coordinate. Prognostic variables for this solver are column mass of dry air (mu), velocities u, v and w (vertical velocity), potential temperature, and geopotential. Non-conserved variables (e.g. temperature, pressure, density) are diagnosed from the conserved prognostic variables. The solver uses a third-order Runge-Kutta time-integration scheme coupled with a split-explicit 2nd-order time integration scheme for the acoustic and gravity-wave modes. 5th-order upwind-biased advection operators are used in the fully conservative flux divergence integration; 2nd-6th order schemes are run-time selectable. WRF-Chem is the WRF model coupled with Chemistry. The model simulates the emission, transport, mixing, and chemical transformation of trace gases and aerosols simultaneously with the meteorology. The model is used for investigation of regional-scale air quality, field program analysis, and cloud-scale interactions between clouds and chemistry. The WRF-Urban modeling system includes three urban parameterization schemes ranging from a simple bulk model to a sophisticated multi-layer urban canopy model. In FY15 we will continue to test runtime options and boundary conditions to develop the best set for Chicago. We will also begin setting up a forecasting system with Urban-WRF-Chicago. We will also explore scaling our configuration beyond the 18 Fusion nodes currently being used. 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: We require more but will use recently purchased disk for the output. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for WRF-CHEM-GVAX
by accounts＠lcrc.anl.gov 02 Oct '14

02 Oct '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Rao Kotamarthi Project Name: WRF-CHEM-GVAX Division: EVS Project title: Weather and Aerosol forecast products for the Ganges Valley Aerosol Project Associated funding: DOE-BER Other Systems: NERSC Science: The Ganges valley region is one of the largest and most rapidly developing sections of the Indian subcontinent. Impacts of changes in precipitation patterns, temperature, and the flow of the snow-fed rivers could be immense. Recent satellite-based measurements have indicated that the upper Ganges valley has some of the highest persistently observed aerosol optical depth values. The aerosol layer covers a vast region, extending across the Indo-Gangetic Plain to the Bay of Bengal during the winter and early spring of each year. This is one of the few regions showing a trend toward increasing surface dimming and enhanced mid-tropospheric warming. The consequences of aerosols and associated pollution for surface insolation over the Ganges valley and monsoons in particular are not well understood. Our proposed field study will provide critical data to address these hypotheses and will contribute to developing better parameterizations for tropical clouds, convection, an d aerosol-cloud interactions. The primary science questions for the mission are as follows: (a) What are the sources and sinks of the aerosols comprising the aerosol cloud over this region and what effects does it have on clear-sky radiation fields, as well as on cloud micro- and macrophysical properties? What effect does the diurnal PBL cycle in the Ganges valley have on the atmospheric distribution of aerosols? (b) What effect do increasing aerosols in the Ganges Valley have on the ISM? What effect do aerosols have on shortwave radiative forcing, mid-tropospheric heating and convective activity over continental India? To achieve these science objectives an implementation has been devised in collaboration with scientists in India. The primary anchor facility for the project is the ARM mobile facility. A deployment starting in approximately June 2011 and lasting to April 2012 at the ARIES observatory located in Nainital is anticipated for AMF ground operations. During this one-year deployment of the AMF there will be an intensive periods, extending from February 1st, 2012 to April 1st 2012 (winter intensive). Project description: We continue to analyze the measurements from GVAX using 3-D regional scale model simulations. These modeling results of air pollution and meteorological fields will become one of the central ‘data’ products for the experiments archived at the ARM websites. The expectation is that there will be a significant number of scientists from around the world who will be interested in GVAX data. In addition, the 3-D model will be a test bed for high performance computing data storage and visualization. The project this year will use the resources to perform diagnostic simulations for the year 2011-2012 and explore the data and model comparison issues arising from the GVAX project. Industry partnership: Project URL: http://www.arm.gov/sites/amf/pgh/ 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 Atmos_Chem
by accounts＠lcrc.anl.gov 02 Oct '14

02 Oct '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Rao Kotamarthi Project Name: Atmos_Chem Division: EVS Project title: Process Scale Modeling of Atmospheric Chemistry and Transport Associated funding: DOE,DoD Other Systems: NERSC Science: The scientific aim of the proposed modeling activity is to further the understanding of the chemistry and distribution of aerosols over certain regional domains and calculate the heating rates as a result of increased aerosol loading primarily over India. A second objective is to use Ensemble Adjusted Kalman Filter technique for data assimilation of satellite data sets into global chemical transport models to estimate model biases in calculating trace atmospheric gas constituents and aerosols A third objective is to perform dynamic downscaling of climate model simulation results for regional scale applications. In particular, we will focus on the uncertainties in downscaled projections resulting from boundary conditions, the utility of spectral nudging and grid scale nudging for improving the energy balance of regional scale climate models over long integration times. Project description: We continue to investigate the regional scale climate modeling and data assimilation of aerosols in chemical transport models. The models employed for these studies include (a) WRF-CHEM (b) NRCM and (3) GEOS-CHEM. Each of these models has been ported to the fusion luster and uses to perform calculations by our group. All the codes use a mix of MPI and OpenMP programming models and can scale for as few as 4 nodes to several hundred. The GEOS-CHEM model doesn't scale beyond 32 nodes due to gridding constraints. The NRCM (Nested Regional Scale Climate Model) and the WRF-CHEM models are highly scalable and limited by job turnover times and queue limitations. The project will have approximately 4 members. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 200000 Q1: 50000 Q2: 50000 Q3: 50000 Q4: 50000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

New LCRC report submitted for entropy
by support＠lcrc.anl.gov 02 Oct '14

02 Oct '14

Full project report at https://accounts.lcrc.anl.gov/reporting/project/2014/?project=entropy

1 0

New LCRC report submitted for entropy
by support＠lcrc.anl.gov 02 Oct '14

02 Oct '14

Full project report at https://accounts.lcrc.anl.gov/reporting/project/2014/?project=entropy

1 0

[LCRC Accounts] Yearly Allocation Request for entropy
by accounts＠lcrc.anl.gov 02 Oct '14

02 Oct '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Sunhwan Jo Project Name: entropy Division: LCF Project title: Quantifying protein-protein binding with greatly scalable multiple copy algorithms of NAMD Associated funding: None Other Systems: 0 (previously had 10M core hours in BG/Q, but is expired and submitted extension) Science: Protein-protein interactions (PPIs) represent a critical aspect of a wide range of biological processes, as most proteins need to associate to various binding partners to carry out their functions. While the basic principles of PPIs are generally understood, it is a grand challenge to accurately predict the binding affinity quantitatively for any given complex. Recently, this potential of mean force (PMF)-based methodology was extended to a much larger protein-peptide system, and we accurately determined the absolute binding free energy between the receptor protein and the ligand peptide. We plan to use our PMF-based methodology to probe effects of mutations in thermodynamics of protein-protein interaction, i.e., thermodynamic decomposition of barstar-barnase protein complex association. Project description: In PMF-based framework, a set of geometric and conformational restraints is applied to the proteins to reduce sampling noise and accelerate the convergence of the resulting free energy profile. These biases are rigorously accounted for to produce the absolute binding free energy. To enhance the PMF convergence, we will combine the umbrella sampling with Hamiltonian and temperature (ℋ and T) replica-exchange molecular dynamics (REM) algorithm. In the US/(ℋ,T)-REMD simulation approach, multiple copies of the molecular  systems are simulated concurrently, but  all with different  conditions, e.g., umbrella potential (Hamiltonians) or temperature. Attempts are periodically made to exchange configurations between different systems using a Metropolis-type Monte Carlo acceptance criterion, insuring Boltzmann-weighted statistics. Such REMD  methodology is extremely scalable, yet also very effective to sample a rugged free energy surface. We anticipate the simulation systems will be comprised of ~50,000 atoms and using at least 255 replicas concurrently. Because of its large size, we plan to use primarily Mira for production and Fusion will be used for initial system preparation and development of enhanced methodology in PMF-based framework. All planned simulations will be carried out by the open-source NAMD molecular dynamics software. The software engine used for the planned simulations is NAMD 2.10 with Charm++ 6.4.0. We have tested t he performance on BG/Q Mira. The scaling of parallel performance was linear up to 512 cores for our test system comprised of 100,000 atoms without using restraint potential and achieved the speed of about 0.025 s/step. Using restraint potential reduces the parallel performance significantly (~0.2 s/step). The expected number of project member is 3. I, Sunhwan Jo, recently joined LCF as a postdoctoral appointee and will be the main person performing code improvement and algorithm development under the supervision of Wei Jiang (LCF) and Benoit Roux (UChicago). 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 perfopt
by accounts＠lcrc.anl.gov 30 Sep '14

30 Sep '14

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Prasanna Balaprakash Project Name: perfopt Division: MCS Project title: Optimization Algorithms for Empirical Performance Associated funding: DOE ASCR Other Systems: Mira, ANL: 100000 core hours NERSC: 100000 core hours Science: Our scientific objectives are: i) To advance the state-of-the-art in empirical performance tuning and adaptation of scientific applications on extreme scale architectures and thereby significantly increasing scientific productivity ii) To improve our understanding of large-scale empirical performance tuning iii) To develop novel mathematical optimization algorithms that exploits the characteristics of the search problem in performance tuning Project description: There are dierent strategies for performance model construction. Analytical approaches for performance modeling are cheaper in the sense that they don't suer from the proling and model generation cost as in the case of machine learning techniques. But analytical models are often dicult to construct for complex programs and they require detailed understanding of the underlying hardware and implementation of communication functions. Machine learning techniques come into play when it is dicult to construct an analytical model for an application. Models generated using machine learning techniques may not be as highly precise as an analytical model but still they help to understand the performance of an application that in turn help in optimizations such as auto tuning. In this research, we first try to build analytical models for communication of various application kernels. We use architecture-independent parameters as described in well known communication models such as LogP family to build the analytical models. If the constructed analytical model is accurate enough in predicting the communication time, it saves the huge overhead in gathering training data for the application of machine learning algorithms. We will test the accuracy of the developed analytical models in predicting the communication time of several application kernels from the CORAL benchmarks. If the analytical model is not accurate enough (the error in prediction goes above acceptable range), we switch to gathering profile data and then generate models using machine learning techniques on that data. This hybrid approach gives us a balance between the complexity of building analytical models and the overhead of running expensive machine learning algorithms on huge amount of collected proling data. Also the machine learning models are limited in the sense that we can't do extrapolation. Fitted analytical models come into play in that case. On Blues, we will conduct an experimental study of communication models. A major benefit Blues brings to this research is the ability to test the impact of MPI implementations and interconnect techniques on Intel-based clusters. Eventually, we hope to deploy our techniques for improving the performance of scientific applications running on Blues. We estimate 100,000 core hours for our experimental study: evaluation of the strengths and limitations of LogGP models = 15,000 core hours; experimental analysis on multiple problem classes on different architectures = 30,000 core hours; empirical analysis for new approaches = 30,000 core hours; tests on parallel performance tuning approaches = 25,000 core hours; We will strongly involve in research dissemination activities such as writing scientific articles and giving seminars at Argonne and in conferences. In the the first and second quarter, our jobs will be 80% parallel. For the third and fourth quarter, we will have 100%. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 100000 Q1: 40000 Q2: 20000 Q3: 20000 Q4: 20000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0