Sciences and Exploration Directorate

Matthew A Thompson

(SR SOFTWARE ENGINEER)

 matthew.thompson@nasa.gov

 301.614.6712

Org Code: 610.1

NASA/GSFC
Mail Code: 610.1
Greenbelt, MD 20771

Employer: SCIENCE SYSTEMS AND APPLICATIONS INC

Brief Bio


Dr Matthew A Thompson is a Lead Scientific Programmer/Analyst in the Global Modeling and Assimilation Office (GMAO) at NASA Goddard Space Flight Center (GSFC). His current role is to support the GEOS GCM in regards to high-performance computing, focusing on maintaining and extending the infrastructure surrounding it including portability and building on various architectures.

Dr Thompson received his BA in Chemistry, summa cum laude, from The Colorado College in Colorado Springs, Colorado. He then moved to the University of Colorado at Boulder where he did his PhD work in Theoretical Chemistry under Dr Robert Parson at JILA. His thesis was "From Femtoseconds to Nanoseconds: Simulations of IBr- Photodissociation Dynamics in CO2 Clusters". After that he was a National Research Council Postdoctoral Scholar at the Naval Research Lab in Washington DC working with Dr Brett Dunlap on analytical density functional theory.

Research Interests


Emerging Computational Acceleration Strategies

Earth Science: Technology & Missions

Investigation and implementation of emerging computational acceleration strategies. Examples include the use of GPUs, MICs, and compiler optimization targeting to speedup calculation of global circulation models

Education


Ph.D. Chemistry, University of Colorado at Boulder, 2007.
Thesis: From Femtoseconds to Nanoseconds: Simulations of IBr- Photodissociation Dynamics in CO2 Clusters

B.A. Chemistry, The Colorado College, 1999.

Positions/Employment


Lead Scientific Programmer/Analyst

SSAI, Inc. (2009-Present) SAIC (2007-2009) - NASA/GSFC Global Modeling and Assimilation Office

June 2009 - Present

Support of the Global Modeling and Assimilation Office at NASA Goddard Space Flight Center

  • Successfully adapted the GEOS-5 global atmospheric model for use with NVIDIA GPUs by converting the existing Fortran code into CUDA Fortran.
  • Demonstrated speedups of 10-15x within the full model with radiation, moist, turbulence, and gravity wave drag physics routines. Excluding data transfer, achieved kernel speedups of 25-50x.
  • Successfully transitioned the GEOS-5 shortwave radiation kernel for use on the Intel Xeon Phi for use in both Offload and Native modes.
  • Experienced with building, setting up, running, and analyzing results from the GEOS-5 Global Circulation Model with various compilers and on various workstations and supercomputers.
  • Successfully built and maintained on various clusters and with various compilers the required libraries for the GEOS-5 GCM including HDF5, NetCDF, cURL, and ESMF.


National Research Council Postdoctoral Scholar

National Research Council (Advisor: Brett Dunlap) - Naval Research Laboratory

June 2007 - May 2009

Investigation on the use of genetic algorithms for optimization of analytic density functional theory (DFT)

  • Constructed and integrated a genetic algorithm in Python with existing analytic DFT code leading to better results than previous efforts.
  • Leveraged Parallel Python and MPI to rapidly parallelize genetic algorithm on a per-node and per-core basis for use on workstations and clusters.


Graduate Research Assistant

University of Colorado at Boulder (Advisor: Robert Parson) - JILA

June 2000 - May 2007

Nonadiabatic molecular dynamics and high-level (MRCI and CR-CCSD(T)) electronic structure calculations of IBr-(CO2)n and ICl-(CO2)n

  • Predicted the later-confirmed appearance and disappearance of a long-time absorption recovery process in near-infrared photodissociation of intermediate-size clusters.
  • Mentored fellow students and researchers in installation and use of Linux and quantum chemistry programs on local workstations and clusters.

Professional Societies


Association for Computing Machinery

(Member, Sigplan-Fortran Forum, SIGHPC)

2012 - Present

Awards


2017 - Performance Award, GMAO -- "For consistently and routinely going above and beyond expectations in software development and maintenance"

2014 - Outstanding Technical Support, GMAO -- "For providing superior software engineering and High-End Computing support that significantly enhances GMAO research and operations"

2013 - Outstanding Technical Support, GMAO -- "For groundbreaking work researching, testing, and documenting the use of Graphical Processing Unit (GPU) computing architecture with the GEOS-5 atmospheric general circulation model"

Publications


Refereed

Caraballo-Vega, J. A., N. S. Smith, M. L. Carroll, et al. L. Carriere, J. E. Jasen, M. T. Le, J. Li, K. Peck, S. L. Strong, G. S. Tamkin, M. A. Thompson, and J. H. Thompson. 2022. Remote Sensing Powered Containers for Big Data and AI/ML Analysis: Accelerating Science, Standardizing Operations IGARSS 2022 - 2022 IEEE International Geoscience and Remote Sensing Symposium [10.1109/igarss46834.2022.9883436]

Hu, L., C. A. Keller, M. S. Long, et al. T. Sherwen, B. Auer, A. Da Silva, J. E. Nielsen, S. Pawson, M. A. Thompson, A. L. Trayanov, K. R. Travis, S. K. Grange, M. J. Evans, and D. J. Jacob. 2018. Global simulation of tropospheric chemistry at 12.5 km resolution: performance and evaluation of the GEOS-Chem chemical module (v10-1) within the NASA GEOS Earth system model (GEOS-5 ESM) Geoscientific Model Development 11 (11): 4603-4620 [10.5194/gmd-11-4603-2018]

Eastham, S. D., M. S. Long, C. A. Keller, et al. E. Lundgren, R. M. Yantosca, J. Zhuang, C. Li, C. J. Lee, M. Yannetti, B. M. Auer, T. L. Clune, J. Kouatchou, W. M. Putman, M. A. Thompson, A. L. Trayanov, A. M. Molod, R. V. Martin, and D. J. Jacob. 2018. GEOS-Chem High Performance (GCHP v11-02c): a next-generation implementation of the GEOS-Chem chemical transport model for massively parallel applications Geoscientific Model Development 11 (7): 2941-2953 [10.5194/gmd-11-2941-2018]

Freitas, S. R., G. A. Grell, A. Molod, et al. M. A. Thompson, W. M. Putman, C. M. Santos e Silva, and E. P. Souza. 2018. Assessing the Grell-Freitas Convection Parameterization in the NASA GEOS Modeling System Journal of Advances in Modeling Earth Systems [10.1029/2017ms001251]

Sheps, L., E. M. Miller, S. Horvath, et al. M. A. Thompson, R. Parson, A. B. McCoy, and W. C. Lineberger. 2011. Solvent-mediated charge redistribution in photodissociation of IBr- and IBr-(CO2) J. Chem. Phys. 134 184311 [10.1063/1.3584203]

Sheps, L., E. M. Miller, S. Horvath, et al. M. A. Thompson, R. Parson, A. B. McCoy, and W. C. Lineberger. 2010. Solvent-mediated electron hopping: Long-range charge transfer in IBr-(CO2) photodissociation Science 328 220–224 [10.1126/science.1184616]

Thompson, M. A., J. P. Martin, J. P. Darr, W. C. Lineberger, and R. Parson. 2008. A combined experimental/theoretical investigation of the near-infrared photodissociation of IBr-(CO2)n J. Chem. Phys. 129 224304 [10.1063/1.3033746]

Thompson, M. A., and B. I. Dunlap. 2008. Optimization of analytic density functionals by parallel genetic algorithm Chem. Phys. Lett. 463 278–282 [10.1016/j.cplett.2008.08.075]

Dribinski, V., J. Barbera, J. P. Martin, et al. A. Svendsen, M. A. Thompson, R. Parson, and W. C. Lineberger. 2006. Time-resolved study of the solvent induced recombination in size-selected IBr-(CO2)n clusters J. Chem. Phys. 125 133405 [10.1063/1.2217741]

Sanford, T., S.-Y. Han, M. A. Thompson, R. Parson, and W. C. Lineberger. 2005. Photodissociation dynamics of IBr-(CO2)n, n<15 J. Chem. Phys. 122 054307 [10.1063/1.1839178]

Sanford, T., D. Andrews, J. Rathbone, et al. M. Taylor, F. Muntean, M. Thompson, A. B. McCoy, R. Parson, and W. C. Lineberger. 2004. Time resolved solvent rearrangement dynamics Faraday Discuss. 127 383–394 [10.1039/b316330j]

Talks, Presentations and Posters


Invited

Continued Efforts in Adapting the GEOS-5 AGCM to Accelerators: Successes and Challenges
Programming Weather, Climate, and Earth-System Models on Heterogeneous Multi-core Platforms Workshop, Boulder, CO

2013


Progress in Adapting the GEOS-5 GCM to CUDA FORTRAN: Successes and Challenges
Programming Weather, Climate, and Earth-System Models on Heterogeneous Multi-core Platforms Workshop, Boulder, CO

2012

 


Lessons Learned from Adapting GEOS-5 GCM Physics to CUDA Fortran
Programming Weather, Climate, and Earth-System Models on Heterogeneous Multi-core Platforms Workshop, Boulder, CO
 

2011

 


Other

Extending and Exploring GEOS Performance Portability (with Amidu Oloso)

Supercomputing 2018, Dallas, TX

December 2018


Improving the Scalability of the GEOS-5 Earth System Model (with Amidu Oloso, Tom Clune, and William Putman)

Supercomputing 2016, Salt Lake City, UT

2016


Analytic Density-Functional Theory Optimization with a Parallel Genetic Algorithm
Poster, American Conference on Theoretical Chemistry 2008, Evanston, IL

2008


Nonadiabatic MD Simulations of IBr-(CO2)n Photodissociation
Poster, American Physical Society, March Meeting, Denver, CO

2007