HP-SEE Pilot Call: Awarded Applications

HP-SEE Pilot Call: Awarded Applications

HP-SEE project gives access for 13 scientific projects to HPC resources in South East Europe through the 1st pilot call for access.  

HP-SEE (High-Performance Computing Infrastructure for South East Europe’s Research Communities) pilot call for proposals was organized between 7th September and 5thOctober 2012. In total 20 projects were submitted from 9 different countries in the fields of Computational Chemistry, Computational Physics, Life Sciences, Earth and Space Sciences, and Engineering. The submitted projects underwent both technical evaluation by technical reviewers-experts of the participating HPC centers (Centers offering resources to the plot call), and scientific peer review by selected scientific reviewers from all the participating countries in the region. Finally, 11 projects which satisfied the eligibility criteria are selected and will take access to the powerful national High Performance Computing systems available in the SEE region, for a period of 12 months, starting from December, 2012. Another 2 projects from the Eastern Mediterranean region gained access to limited HPC resources in collaboration with the LinkSCEEM-2 project.

The available systems to be used by the selected applications are:

  • A hybrid HPC cluster offering a 3.2 TFlop CPU (Intel Xeon X5560) sub-cluster and a 4.6 Tflop double precision GPU sub-cluster is being available in Bulgaria.
  • Two AMD Opteron HPC clusters with 5.4 and 14 Tflops performance respectively, an SGI Ultraviolet SMP machine with a performance of 10 Tflops, an SGI HPCsystem with Intel Xeon processors offering a performance of 18 Tflops being available in Hungary.
  • An Intel Xeon based HPC cluster with a performance of 6 Tflops being available from Serbia.
  • An Intel Xeon based HPC cluster with 2.72 Tflops performance, a GPU Cluster with 480 GPU cores, and a BlueGene/P Supercomputer with a performance of 13.4Teraflops begin offered by Romania.
  • A 7.76 Teraflop HPC cluster with Intel Xeon processors being offered by FYR of Macedonia.

The following paragraphs give further details for the selected applications.

Interface States in Organic Materials (ISTORM)

Scientific Computing Laboratory, Institute of Physics Belgrade

HPC system assigned: Szeged SC (Hungary)

Application description (Computational Physics):

Organic semiconducting materials based on conjugated polymers and small molecules hold great promise for applications in cheap solar cells, field-effect transistors and organic LEDs. Realistic materials exhibit a rather complicated structure polymer based materials contain a mixture of crystalline and disordered domains, while small molecule based materials contain crystalline domains with different orientations separated by domain boundaries. In both cases, it is crucial to understand the nature of electronic states that arise at interfaces between the domains. The presence of such states is expected to have a strong impact on the current through the material and consequently on the important device characteristics (such as solar cell efficiency, LED brightness, etc.) Up to now, there has been a lack of understanding of the properties of the states that occur at domain boundaries. Due to complex structure of the material, it is difficult to probe these states experimentally. On the other hand, theoretical studies have focused mostly on purely ordered structures and more recently also to completely disordered structures. In the past, PI of this application has developed a set of methods which can be used to calculate the electronic states of organic systems that contain thousands of atoms the charge patching method and the overlapping fragments method. These methods will be used to calculate the electronic states that arise at the boundaries of crystalline domains in naphthalene a representative of small molecule based organic semiconductors, as well as to study the states at the boundaries between crystalline and disordered domains in P3HT a representative of conjugated polymer based materials.

Mechanistic studies of the Arp2/3 complex activation (ARP23)

Molecular Modeling and Drug Design Group, Center of Basic Research I, Pharmacology-Pharmacotechnology Division, Biomedical Research Foundation of the Academy of Athens

Institute of Molecular Biology, Department of Chemistry, University of Oregon

HPC system assigned: Szeged SC (Hungary)

Application description (Life Sciences):

Actin-related-protein 2/3 (Arp2/3) complex is a seven subunit ATP-ase that plays a central role in the regulated actin assembly. The Arp2/3 complex nucleates the polymerization of a new actin filament that emerges from an existing filament and is required for many cellular processes. Importantly, tumor cell migration is thought to require Arp2/3 complex, and Arp2/3 overexpression contributes to pathogenesis, growth, and invasion of carcinomas. While inactive on its own, Arp2/3 is activated by interacting with nucleation promoting factor (NPF) proteins that contain the conserved VCA domains. Biochemical evidence suggests that two VCA domains bind Arp2/3 each delivering an actin monomer to the complex. The result of VCA/actin binding to Arp2/3 is the activation of the complex that is accompanied with a large conformational change. It is unclear how the VCA domains bring about the Arp2/3 activation. The aim of the project is to unravel

the mechanism that underlies the Arp2/3 complex activation using standard and biased molecular dynamics (MD) simulations. In collaboration with the experimental group of Dr Brad Nolen in the University of Oregon, this study will investigate the following questions: a) The location of the VCA-binding sites on Arp2/3 complex b) The effect of the bound VCA domains on the conformation of Arp2/3. The results of these studies will help us in the design of new small molecule inhibitors of Arp2/3 complex for basic research and biomedical applications.

A hybrid methodology High Performance methodology for the simulation of microcrystalline Silicon (High Performance Silicon - HPS)

Laboratory of statistical mechanics and macromolecules, Department of Chemical Engineering, University of Patras

Laboratory of statistical mechanics and macromolecules , Institute of Chemical Engineering and High-Temperature Chemical Processes

HPC system assigned: Debrecen SC (Hungary)

Application description (Engineering)

The last ten years have seen a dramatic change for the field of renewable energies. High conventional energy prices and the awareness of the impact of global warming have led, with the assistance of suitable supporting laws, to the creation of a strongly growing clean-tech industry. Photovoltaics are particularly attractive because can physically cover on the long term an important fraction of electricity needs. Materials which exhibit ordered morphology at the nano-scale have drawn considerable attention in the last two decades due to their unique combination of opto-electronic properties, ease of preparation and low cost manufacturing. Systems like micro-crystalline silicon have proved to be very promising candidates for applications in photovoltaics. Simulating these materials at the nanoscale is inefficient with Molecular Dynamic (MD) methods because of the problem of long relaxation times, especially when highly ordered structures are formed at low enough temperatures. In order to circumvent any MD drawbacks, its imperative to develop and implement efficient Kinetic MC Monte Carlo (KMC) techniques to effectively simulate systems with large scale nanophase-separated structures and overcome obstacles related with large system sizes and sluggish dynamics. This presents a considerable problem when one wishes to simulate large systems at microscale with atomistic details over a wide range of experimental conditions. To deal with such problem, we resort to parallel versions of the KMC technique. We propose here designing such a novel parallel kMC algorithm combined with MD simulations in order to simulate micro-crystalline silicon thin films.

Crosspoint Queued Switch simulator (CQS)

Center of Information Systems, University of Montenegro

Faculty of Electrical Engineering, University of Montenegro

HPC system assigned: PARADOX (Serbia)

Application description (Engineering)

The crosspoint queued switches have been evaluated in the late eighties, but mostly for uniform traffic and very small buffer lengths. Due to the technological limitations of that time, it was impractical to implement large buffers together with switching fabric. The crosspoint queued switch architecture has been recently brought back into focus since modern technology enables an easy implementation of large buffers in crosspoints. The advantage of this solution is the absence of control communication between linecards, where the buffers resides, and schedulers. In this software, the performances of four algorithms are analyzed and compared: longest queue first, round robin, exhaustive round robin and frame based round robin matching. Throughput, average cell latency, packet loss probability and instantaneous packet delay variance are evaluated under uniform and two nonuniform traffic patterns. The results obtained for the crosspoint queued switch are compared with the output queued switch.

Transport and dissipation in nano-devices - molecular dynamics and density functional theory simulations (TDN-SIM)

Group of Materials and Devices for Electronics and Optoelectronics (MDEO), Faculty of Physics, University of Bucharest

Group of Nanophysics and Emerging Materials, Department of Theoretical Physics, Horia Hulubei National Institute of Physics and Nuclear Engineering

HPC system assigned: Debrecen SC (Hungary)

Application description (Computational Physics)

In the current project we analyze the non-equilibrium properties of atomic scale systems using ab initio density functional theory (DFT) and molecular dynamics calculations. We shall investigate the eletronic structure, the charge and spin transport properties of low dimensional systems such as nanowires and nanoribbons as potential candidates for future electronic and spintronic devices. We shall also focus on thermal properties by calculating the phonon spectra and the vibration properties of the systems in contact with a thermostat. For the DFT calculations we employ the SIESTA code, which has the major advantage that the computational time scales linearly with the number of atoms in the simulation, due to the usage of localized basis sets. In this way, one may simulate systems of thousands of atoms. The SIESTA code already has a parallel implementation, which is afforded in the large scale calculations. The additional package TRANSIESTA is employed in transport calculations in the framework of non-equilibrium Green's functions. We shall also investigate the nature of dynamical defects and glassy properties in amorphous materials. For larger systems of this type we shall perform molecular dynamics simulations using the parallel code NAMD. We shall investigate the defects and the dissipation of elastic vibrations in nanoscopic objects like electro-mechanical resonators and nanodetectors.

Molecular dynamics study of pMHC/TCR complexes (TCR-MHC)

Department of Economics, University of Ioannina

Department of Informatics, TEI of Epirus

HPC system assigned: HPCG (Bulgaria), NIIFI SC (Hungary), PARADOX (Serbia)

Application description (Life Sciences):

This application intends to clarify the intermolecular interactions between peptide/MHC and T cell receptors, a key step in immune response in many diseases. Antigenic molecules are presented by the Antigen Presenting Cells (APCs) via the Major Histocompatibility Complex (MHC) to these T Cell Receptors (TCRs). TCRs have regions that are very polymorphic, that are responsible for the recognition: the Complemetary Determining Regions (CDRs). Despite the work done until now, mainly by x-ray studies, the detailed mechanism of pMHC/TCR complexation is not known and not fully understood. The complexation is very sensitive in subtle differences at the protein/protein interface and the paradigms we have do not corroborate in a single and straightforward mechanism. During the past decade, along with x-ray studies, several researchers have employed theoretical methods, mainly molecular dynamics simulations, in order to obtain a better insight into the interaction between the pMHC and TCR. This technique can provide vast information about the spatial arrangement of such protein complexes and also offers a framework to study the dynamics of the system. The proposed project aims to study several pMHC/TCR related to cancer, AIDS or other autoimmune diseases (melanoma, diabetes, EB virus, etc) to an unprecedented time scale and spatial resolution in order to get cutting-edge detailed information about the dynamics such important protein complexes.

The main focus of the project is in three major areas:

1. To describe analytically the specific molecular interactions of the pMHC/TCRcomplexes

2. To determine the energetics (free energy analysis) that govern the pMHC/TCR complexation and to examine the role of specificc residues at the protein/protein interface

3. To test the spatial arrangement of hypothetical complexes, that have not yet been examined experimentally, or even they do not exist, in order to determine the possible contribution of specific interactions to pMHC/TCR engagement.

With this project it is hoped to publish 3-5 papers in well known international journals.

Foreseen social impact

This project, if successful will have two main impacts:

1. Will facilitate the research for better vaccines to a number of diseases

2. Will provide a basis for personalized medicine treatment, where this is needed in special sub-groups of the population

Dynamics Analysis of Parallel Simulations of Biological Neural Microcircuits (DynAPSNeur)

Advanced Computing Application Lab, Department of Computer Science, Technical University of Cluj-Napoca

Department of Mathematics, Technical University of Cluj-Napoca

HPC system assigned: Pecs SC (Hungary)

Application description (Life Sciences):

The proposed topics belong to the Computational Neuroscience research field. More precisely, the project investigates the impact of parallelization strategies on the dynamic behavior of biological neural microcircuit simulation. The main components of the research in this field are:

·       development and investigation of parallel simulation models of spiking neurons and synapses (such as Hodgkin-Huxley, Izhikevich, Spike-Response-Model (SRM) etc;

·       design of a methodology for the analysis of the dynamic behavior of the simulations (based on dimensionality reduction, ;

·       application of this methodology to the parallel simulation of a wide range of biological neural microcircuits.

The ultimate goal of this research is to provide reference parallel implementations of the investigated biological models, whose simulation exhibit the same dynamics as the sequentially ones. So far, we have developed parallel implementations of the Hodgkin-Huxley and of the SRM models, we have built in simulations neural microcircuits based on these models, and computed various small and medium-size scenarios. The simulation code we developed for this research is based on OpenMP and MPI.

Implementation of parallel algebraic algorithms for solving problems in engineering (IPA2SPE)

 

Mathematical foundation of Informatics, Institute of Mathematics and Informatics

Department of Mathematics, Construction University "Lyuben Karavelov"

HPC system assigned: HPCG (Bulgaria)

Application description (Engineering)

New tools for solving problems in civil engineering by Monte Carlo methods and for solving problems in computer algebra (e.g. integer factorization) will be developed and tested. An attempt to find new codes over finite rings which to be applied to construction of communication systems will be made.

Problems solved by project

More effective tools for solving problems in civil engineering by Monte Carlo methods More effective tools for solving problems in computer algebra (e.g. integer factorization) New codes over finite rings which will be applied to construction of communication systems.

High Gain Reach in Fusion by Relativistic Compression Ion Bunches in Radiation Pressure Acceleration Regime (HGRF)

Programming Group, IT Department, Tbilisi State University

Group of Plasma Physics, Institute of Physics, Tbilisi State University

HPC system assigned: Debrecen SC (Hungary)

Application description (Computational Physics)

This project investigates the acceleration of low-Z ion high dense relativistic ion bunches by ultraintense circularly polarized laser pulses. Highly monoenergetic low-Z ions beams with GeV energies can be produced. The accelerated relativistic ion bunches have a very small divergence angle. A few GeV ion bunches can be produced although spatial conditions for the ultrashort laser pulse and the overdense plasma, such as extremely high laser intensity very sharp rising front and sharp inhomogeneity at the thin target surface. The project tries to find optimal parameters of laser pulses and targets for getting the possible highest fusion reactivity. The thickness of the compressed target and the bunches concentrations must be as more as possible. The project also studies the velocity regulation mechanisms of relativistic bunches for generation both blueshifted and redshifted radiation and the influence of thermal expiation on the dynamics of the plasma acceleration and EM pulse frequency modification.

Numerical experiments on the natural convection of the fluids between coaxial cylinders and concentric spheres (NUM-EXP-NAT-CONV)

Group of Computational Physics, Department of Physics, Faculty of natural sciences - Tirana University

HPC system assigned: PARADOX (Serbia)

Application description (Computational Physics)

Natural convection of a fluid between rigid boundaries kept at constant surface temperature received much attention because of the theoretical interest and the wide engineering applications. The fluid flow in a cylindrical annulus shows a multiplicity of solutions (bifurcation phenomenon). The problem of the stability of the solutions in different geometries (cylindrical or spherical) is at the uttermost interest of several theoretical studies. In the case of two coaxial cylinders, using the Oberbeck-Boussinesq approximation[4],the partial differential equations governing conservation of mass, momentum and energy are written into non dimensional form of cylindrical coordinates introducing Prandtl and Rayleigh number. Running ANS YS-Fluent software package (noncommercial version) in a PC, we will study the stability of solutions for different values of Prandtl and Rayleigh parameters and for different temperature difference between coaxial cylinders. To receive more precise results (by better meshing and more time steps), we need to run the OpenFOAM package in a supercomputer or cluster. The received numerical results will be confronted with results of the theoretical studies. A special part of the project will concern to the natural convection of the fluid between two concentric spheres. In this case, a radial non-uniform gravitational field will be considered.

Phylogenetic Tree Analysis (PhyloTree)

Institute for Animal Husbandry, University of Banja Luka Faculty of Agriculture

Department of Biology, University of Tuzla Faculty of Natural Sciences

HPC system assigned: Debrecen SC (Hungary)

Application description (Life Sciences):

This application will be based on RAxML application from The Exelixis Lab. It will try to compare phylogenetic trees within two different species of fish (Oncorhynchus mykiss, Abramis brama) trying to establish the preferred genetic lineage for each of the species. Aim of the application is gentypization of named species through the use of modern tools providing us with high quality base for future research.

Problems solved by project

These species are among the most important aquacultural species in the world. These areas are domminated by commercially grown fish, especially trout, among which Oncorhynchus mykiss is very resistant to varying quality of water and surroundings.

Running Coupling and Anomalous Dimension of SU(3) Gauge Theory with Adjoint Fermions – (SU3-GT)

Tel Aviv University

HPC system assigned: BlueGene/P SC (Romania)

Application description (Physics):

Technicolor theories are extensions of the Standard Model of high energy physics and are specified by the gauge fields and the fermion species they contain. According to the number and types of the fermions, a given theory may be confining or conformal. To be of interest for phenomenology, a theory should be a "walking" theory - on the borderline between confining and conformal,

Applying lattice methods to technicolor models with fermions in the two-index symmetric representation based on the SU(2), SU(3), and SU(4) gauge groups, it was found that it is difficult to confirm the desired walking behaviour. The anomalous dimension of the mass operator was evaluated and it was found that a value very close to 1 is required for a walking theory to be viable. We found it to be less than 1/2 in all three theories, effectively eliminating them from consideration.

Problems solved by project

We propose to apply the same methods to the SU(3) gauge theory with fermions in the adjoint representation. This is also a candidate for a borderline walking theory.

A Petascale Approach to Large-Scale Computational Atomic and Molecular Collisions – (PALSC-AMC)

Jordan University of Science and Technology, Queen's University, Belfast, Institut des Sciences Mol. Auburn University

HPC system assigned: BlueGene/P SC (Romania)

Application description (Physics):

We focus on the development of computational methods to solve the Schrodinger and Dirac equations for a variety of atomic and molecular collision processes using the R-matrix and R-matrix with pseudo-states (RMPS) methods. Access to leadership-class computers allows us to benchmark our theoretical solutions against dedicated collision experiments and to provide atomic and molecular data for ongoing research in laboratory and astrophysical plasma science. Benchmarking our theoretical solutions against dedicated collision experiments remains one of our primary goals.

Problems solved by project

We provide theoretical support to ongoing experiments carried out at a variety of light sources. Theoretical calculations for the relevant cross sections are required involving millions of energy points for an equal comparison and provides an extremely precise test of our theoretical and computational methods.