Work Experience

Post-Doctoral Research Scientist
IBM Research, Almaden, CA, 2017
  • Develop software frameworks and applications for modern CPU/GPU supercomputers
  • Optimize benchmark codes and co-design future hardware based on workloads
  • Write code to stress features of upcoming compilers and influence design and performance
  • Design math library algorithms and frameworks with Lawrence Livermore National Laboratory
Graduate Student Research Assistant
University of Michigan, Ann Arbor, MI, 2012-2017
  • Developed novel high-order discontinuous Galerkin (DG) methods and implemented in a large code
  • Advanced output-driven adjoint-based algorithms to target desired simulation metrics
  • Received several awards; results have been presented at major conferences and in papers
  • Demonstrated capability of new methods using curvilinear unstructured meshes distributed over hundreds of processors
Visiting Research Assistant
Lawrence Livermore National Laboratory, Livermore, CA, 2012
  • Implemented an adaptive parallel code using a newly-developed approach for general unsteady error estimation
  • Demonstrated this to adaptively solve the Porous Medium Equation, a nonlinear diffusion-like PDE
  • Developed interoperability betwen SUNDIALS time integration and Chombo strucutred adative mesh refinement libraries
  • Received funding from Predictive Science Academic Alliance Fellowship
Visiting Research Assistant
Naval Research Laboratory, Washington, DC, 2010
  • Designed and implemented parallel CFD algorithms for GPUs
  • Developed novel method and code for air-traffic control based on a technique mapping nodes back to a regular Lagrangian grid
  • Work led to paper that won the Alan Berman Research Publication Award at the lab
Research Assistant
Center for Information Technology Integration, Ann Arbor, MI, 2008-2012
  • Summer internships working on isolating bugs in NFSv4

Open Source HPC Contributions

Umpire and Simpool @ IBM
  • Developed efficient fixed- and dynamic-type memory pools templated on standard C++ allocators to accept any memory space on heterogeneous systems
  • Has become a key component of the modern C++ HPC toolchain at Lawrence Livermore National Laboratory and integrated into many of the applications and frameworks for large speed-ups
  • Enables new, optimal behavior: Multi-physics packages can query memory usage by other compo- nents and decide how best to allocate compute and memory resources
MFEM and Laghos @ IBM
github.com/mfem/mfem, 2017-current
  • Ported large CPU-only C++ finite element framework to a heterogeneous MPI parallel environment with an array of compute devices (MPI+X)
  • Replaced non-scalable parallel sparse matrix formulation by efficient matrix-free solvers that take advantage of the tensor-product basis structure - resulted in speedups of over 50x for high orders
  • Developed a GPU-enabled version of the Exascale Computing Project’s Lagrangian hydro mini-app Laghos and run at scale on Sierra supercomputer
  • Designed a modern C++ framework for expressing finite element operators and solvers to enable efficient computation
XFLOW @ University of Michigan
  • Developed a MPI-parallel finite element framework with a novel hybrid discontinuous Galerkin (HDG) algorithm that greatly reduces the size of the parallel sparse matrix and adds a level of parallelism
  • Designed and developed an adaptation framework for the framework that takes into account new information that tells where the error occurred
  • Demonstrated the new HDG solver together with the adaptive framework on a series of very large- scale test problems that were previously unable to be solved with the framework

Education

Ph.D. - Aerospace Engineering and Computational Science
January, 2017
University of Michigan, Ann Arbor, MI

M.S.E. - Aerospace Engineering
September, 2012
University of Michigan, Ann Arbor, MI

B.S.E. - Aerospace Engineering
April, 2011
University of Michigan, Ann Arbor, MI

Awards

Beyster Computational Innovation Fellowship
2016
XSEDE High-Performance Computing Allocation Award
2014
Naval Research Laboratory Alan Berman Research Publication Award
2013
National Science Foundation Graduate Research Fellowship
2012
Engineering Graduate Symposium Best Poster Award
2012
Predictive Science Academic Alliance Fellowship
2011
Johann Dahm © 2019
rss facebook twitter github gitlab youtube mail spotify lastfm instagram linkedin google google-plus pinterest medium vimeo stackoverflow reddit quora quora keybase paperclip