Topic Proposal 3 (Optimisation, Scalability and Porting of Codes) Technical content/scope Optimisation and scaling of application codes to thousands of cores including porting of codes to new (heterogeneous or homogeneous) multicore hardware architectures, using advanced methods, technologies, and tools. Examples include: use of new methods for mesh generation, new solver parallelisation, various forms of task and data parallelisation, utilization of specific accelerators, including GPU and FPGA. Scientific computing domains and application domains are focused on, but not limited to: CFD, molecular dynamics, electromagnetic, biology, seismic signal processing and remote sensing. Funding scheme: Small or medium-scale focused research projects. Expected impact: (i) The state-of-the-art in optimisation and scalability methodologies should be significantly advanced. Effective measurements of improved performance and comparison between various types of parallelisation will be valuable (ii) porting of codes to bigger number of cores (iii) increased cooperation between EU and Russian organisations. topic
Project expertise • Project need real experience on decision of large CFD problems with using high performance supercomputers with productivity 100 Tflops+. • Motherboards based on graphical processors are an example of successful use of multicore processors . • Modern trends of supercomputer technology will be connected with construction of multicore processors with number of cores about 100, 1000 etc. Computer systems based on “super-multi-core” architecture possess essentially smaller cost and power consumption. • Main direction of proposed work is finding of program algorithmic decisions based on supercomputers with supermulticore architecture.
Supercomputer technologies for efficiency increasing of geological survey and oil recovery. 2 Search for nonstructural deep laying hydrocarbon deposits 1 Oil recovery increasing with the help of passive microseismic monitoring 3
Project background Carbon nanostructures, such as nanotubes and graphens, have unique mechanical properties, which can be used for development of extremely sensitive devices. Carbon nanotubes Mass nanosensor nanorelay nanooscillator nanoactuator
HPC Activities Available HPC models HPC Continuum Models HPC Atomistic Models • High-precision numerical methods for modeling complex dynamic wave processes in heterogeneous medium • Optimization for massive parallel multicore platforms (up to 1000 cores) • Solvers on adaptive hierarchical grid • Parallel solvers based on domain decomposition for semi-empirical molecular dynamics • Hybrid atomistic methods • Linear scaling atomistic methods for calculation of electron transport
Research topics Development of hybrid atomistic and continuum models for modeling of nanoelectromechanical systems (NEMS) Scales Atomistic Continuum Mechanical Molecular Dynamics with semi-empirical potentials Shell theory • Development of numerical parallel methods for hybrid atomistic and continuum models for calculation of mechanical, electronic and emission properties of NEMS: • Integration of atomistic and grid mesh continuum methods • Interlevel data transfer • Optimization of processor loading for multilevel methods Electronic Quantum transport equation Electro-dynamics of continua Physical Properties Emission Quantum transport equation Quasi-hydrodynamic models
Project example multiscale and multiphysics simulation methods for NEMS modeling, based on combination of atomistic and continuum models. Noise and fluctuations Fokker-Plank equation Phase stability Langevin dynamics Dissipation Diffusion and drift coefficients Molecular Dynamics Electron transport Friction force, fluctuations, van der Waals force Ab initio Interatomic potentials