Researcher at national supercomputing center IT4Innovations. He is interested in the fields such as finite element method, linear solvers, and linear algebra generally. He works on domain decomposition method for massive parallelization known as Hybrid TFETI.
02/2013 - 08/2013
Working on Krylov subspace methods based on multiple search direction technique. Parallel implementation of block conjugate gradient methods, and GMRES.
Month internships supported by BULL and hosting at university at UPMC LJLL, collaboration with prof. F.-X. Roux.
2006 - present
Mechanical engineering, FEM and Linear Algebra programming courses in Matlab or Python
Thesis: Scalable Domain Decomposition Methods for Solution of Statics Problems
Thesis: The assessment of possibility of initiation of parametric vibration at winding cages
High Performance Computing, Domain Decomposition Method, Krylov subspace methods, Applied Linear Algebra
Main activities of the Intel® PCC at IT4I are divided into two pillars: The Development pillar of highly parallel algorithms and libraries focuses on the development of the state-of-the-art sparse linear iterative solvers combined with appropriate preconditioners and domain decomposition methods, suitable for solutions of very large problems distributed over tens of thousands of Intel® Xeon Phi™ coprocessors accelerated nodes. Developing solvers will become part of the IT4I in-house ESPRESO (ExaScale PaRallel FETI SOlver) library. The support of HPC community codes includes creating interface between ESPRESO and existing community codes Elmer and OpenFOAM Extend Project.
- Project Spomech – Creating a multidisciplinary R&D team for reliable solution of mechanical problems, reg. no. CZ.1.07/2.3.00/20.0070 within Operational Programme ‘Education for competitiveness’ funded by Structural Funds of the European Union and state budget of the Czech Republic.
- Development and implementation of scalable the algorithms for the solution of quadratic programming contact problems.
- Computationally intensive simulations and optimization provided by Ministry of Education, MSM6198910027, http://comsio.vsb.cz/.
- International project PRACE (see http://www.prace-project.eu).
- Scalable algorithms based on the domain decomposition methods for the solution of transient contact and impact problems with friction, (Czech Science Foundation, Post-doc project 2013).
- Opportunity For Young Researchers | CZ.1.07/2.3.00/30.0016
ESPRESO is an ExaScale PaRallel FETI SOlver developed at IT4Innovations. Main focus of the development team is to create a highly efficient parallel solver which contains several FETI based algorithms including new Hybrid Total FETI method suitable for parallel machines with tens or hundreds of thousands of CPU cores. The solver is based on highly efficient communication layer on top of pure MPI. The layer was developed specifically for FETI solvers and uses several state-of-the-art communication hiding and avoiding techniques to achieve better scalability.
2013 - 2015
PermonCube is a simple parallel mesh generator over cubical domain allowing the generate 3 dimensional elastoplasticity problem over billions of unknowns. For nonlinear plasticity problem It contains nonlinear solver based on Newton method, and for solving of linear system it cooperates with several solvers (initial ver. of ESPRESO, Permon, DDSolv). Improved version of PermonCube part covering mesh generating is nowadays component of ESPRESO.
The algorithms implemented in MatSol library are based on our long-term research focused on the development of scalable algorithms for multibody contact and contact shape optimization problems, 2D or 3D, with or without friction. Isotropic and anisotropic, Tresca and Coulomb friction models are supported.
Subject covers partial differential equations, basic methods of approximation of the solution of the boundary value problems, variational equalities and inequalities.
03 / 2009 - 04 / 2009
Prof. Peter Wriggers’ group, Theme: Contact problems, nonlinear mechanics.