Photo of Dr Martin Schreiber

Dr Martin Schreiber

Proleptic Lecturer (Education and Research)

Email:

Telephone: 01392 725280

Extension: (Streatham) 5280

Visit personal website »

Research interests:

  • Applications (among others): Weather, Climate, oceanic currents
  • Algorithms: development of new algorithms for scientific computing
  • Parallelization: new algorithms on HPC architectures, new parallelization concepts/models, Parallelization in Time
  • Architectures: CPU, GPU, XeonPhi, FPGAs
  • Mesh: Focus on dynamically adaptive grids based on space-filling curves
  • Compute resources: Dynamic resource scheduling
  • Realtime: Interactive simulations
  • Visualization: Efficient on- and offline processing

Possible PhD topics available (please get in touch via Email for further questions):

  • Parallelization in time algorithms for Exascale computing with focus on climate and weather
  • HPC hardware-aware optimization of the NEMO ocean simulation framework
  • Development of an regularization method for large-scale time-varying tomography framework on large-scale HPC systems (solar tomography)
  • Simulations on a dynamically adaptive grids

Martin Schreiber was appointed as a proleptic lecturer in 2015 at the University of Exeter. His current focus is in HPC in various areas: Biological parameter estimation on accelerator cards, parallelization in time methods, ocean simulations, etc.

In 2010, he wrote his Diploma thesis in Computer Science at the Technische Universitaet Muenchen (TUM) on simulation and visualization of the free surface Lattice Boltzmann Equation on GPUs (see Diplomathesis).

At the end of 2010, he joined the research group of Prof. Bungartz at TUM as a PhD student where he worked in the Invasive Computing Transregio Project (DFG funded). His work in this project was two-folded: In collaboration with other members of the project he redesigned algorithms to support dynamical resource management  on embedded systems. For high-performance systems, he developed a new cluster-based parallelization method for efficiently running simulations on dynamically adaptive triangular grids with MPI+X parallelization models and presented the benefits of dynamic resource management for Tsunami parameter studies (see PhD Thesis).

Talks:
  • Nov. 2016: PinTing oscillatory problems with a massively parallel rational approximation, Fifth Parallel-in-time Integration Workshop, Banff, CA
  • Sep. 2016: Dynamic Adaptive Mesh Refinement with RLE-clustering vs. Parallelization-in-Time with REXI, Advances in HPC for geoscience applications, Milano, IT
  • Sep. 2016: Beyond scalability limitations: Massively parallel rational approximation of oscillatory problems & more (invited talk), NOAA, USA
  • Mai 2016: Massively parallel rational approximation of oscillatory problems, STFC Daresbury Laboratory, Scientific Computing Department, Daresbury, UK
  • Apr. 2016: Beyond scalability limitations: Massively parallel rational approximation of oscillatory problems (invited talk), Mini-symposium, HPC days in Lyon, Lyon, FR
  • Apr. 2016: HPC & accelerator cards (guest lecture), University of Bath, Bath, UK
  • Feb. 2016: Beyond scalability limitations: Massively parallel rational approximation of oscillatory problems (invited talk), University of Bath, Bath, UK
  • Dez. 2015: PinT it! Parallelization-in-Time for Climate and Weather (invited talk), Technical University of Munich, Munich, GER
  • Nov. 2015: Parallelization in time for climate and weather simulations (invited talk), University of Stuttgart, Stuttgart, GER
  • Oct. 2015: Parallelization in time with application for climate and weather, Partial differential equations on the sphere (PDEs), Seoul, South Korea
  • May 2015: Parallelization-in-time for climate and weather, Mathematics of the Weather Workshop 2015, Erquy, FR
  • Mar. 2015: Towards cluster-based parallelization for high-dimensional Cartesian grids, Galerkin methods with applications in weather and climate forecasting, ICMS, Edinburgh, UK
  • (hiking break and change of research topics)
  • Nov. 2013: Parallelization of dynamically changing grids with a cluster-based approach and invasion, University of Exeter, Exeter, UK
  • Nov. 2013: Parallelization of dynamically changing grids with a cluster-based approach and invasion (invited talk), Imperial College London, London, UK
  • Sep. 2013: SFC-based Communication Metadata Encoding for Adaptive Mesh, ParCo 2013, Garching, Germany
  • Aug. 2013: Cluster Optimization and Parallelization of Simulations with Dynamically Adaptive Grids, Euro-Par 2013, Aachen, Germany
  • Dec. 2012: Shared Memory Parallelization of Fully-Adaptive Simulations Using a Dynamic Tree-Split and -Join Approach, IEEE International Conference on High Perf. Computing (HiPC), Puna, India
  • Sep. 2012: Invasive Computing on High Performance Shared Memory Systems, Facing the Multicore-Challenge III, Stuttgart
  • Feb. 2012: Space-Filling Curves Continued: Sierpinski, Leogang 2012, Leogang, Austria
  • Jan. 2012: Let's play - building a game physics engine, WEP 2012, KAUST, Saudi Arabia
  • Jun. 2011: Free-Surface Lattice-Boltzmann Simulation on Many-Core Architectures, ICCS 2011, Singapore
Publications:
  • Journals:
    • 2016: M. Schreiber, T. Neckel, H.-J. Bungartz, Evaluation of an efficient stack-RLE clustering concept for dynamically adaptive grids (accepted), SIAM Journal on Scientific Computing
    • 2016: M. Schreiber, P. S. Peixoto, T. Haut and B. Wingate, Beyond spatial scalability limitations with a massively parallel method for linear oscillatory problems (IN REVIEW since Feb. 2016) International Journal of High Performance Computing Applications
    • 2014: M. Schreiber, C. Riesinger, T. Neckel, H.-J. Bungartz und A. Breuer, Invasive Compute Balancing for Applications with Shared and Hybrid Parallelization In International Journal of Parallel Programming, Springer
  • Proceedings (peer reviewed):
    • 2014: M. Schreiber und H.-J. Bungartz, Cluster-based communication and load balancing for simulations on dynamically adaptive grids, In Proceedings of the International Conference on Computational Science (ICCS'14),  Elsevier
    • 2014: M. Schreiber, A. Atanasov, P. Neumann und H.-J. Bungartz, Rendering of Feature-Rich Dynamically Changing Volumetric Datasets on GPU, In Proceedings of the International Conference on Computational Science, ICCS 2014
    • 2014: C. Tradowsky, M. Schreiber, M. Vesper, I. Domladovec, M. Braun, H.-J. Bungartz und J. Becker, Towards Dynamic Cache and Bandwidth Invasion, Springer, April 2014
    • 2013: M. Schreiber, C. Riesinger, T. Neckel und H.-J. Bungartz, Invasive Compute Balancing for Applications with Hybrid Parallelization, In Proceedings of the 25th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD'13)
    • 2013: M. Schreiber, T. Weinzierl und H.-J. Bungartz, SFC-based Communication Metadata Encoding for Adaptive Mesh, Proceedings of the International Conference on Parallel Computing (ParCo)
    • 2013: H.-J. Bungartz, C. Riesinger, M. Schreiber, G. Snelting und A. Zwinkau, Invasive Computing in HPC with X10, In X10 Workshop (X10'13)
    • 2013: M. Schreiber, T. Weinzierl und H.-J. Bungartz, Cluster Optimization and Parallelization of Simulations with Dynamically Adaptive Grids, Euro-Par 2013, Lecture Notes in Computer Science, Springer, Berlin
    • 2012: M. Schreiber, H.-J. Bungartz und M. Bader, Shared Memory Parallelization of Fully-Adaptive Simulations Using a Dynamic Tree-Split and -Join Approach, IEEE Xplore, Puna, India
    • 2012: M. Bader, H.-J. Bungartz und M. Schreiber, Invasive Computing on High Performance Shared Memory Systems, In Facing the Multicore-Challenge III, Lecture Notes in Computer Science
    • 2012: M. Gerndt, A. Hollmann, M. Meyer, M. Schreiber und J. Weidendorfer, Invasive computing with iOMP, In Specification and Design Languages (FDL)
    • 2011: M. Schreiber, S. Zimmer, P. Neumann und H.-J. Bungartz, Free-Surface Lattice-Boltzmann Simulation on Many-Core Architectures, In Proceedings of the International Conference on Computational Science (ICCS) 2011

 

 

Interactive fluid simulation (based on Lattice Boltzmann method) and visualization (volume tracing, photon mapping with OpenGL).


 


Breaking dam simulation.


Simulation of the Tohoku Tsunami on dynamically adaptive grids, see website for Sierpinski framework for further information
Google+