2010 Physical Sciences Review » Biographies » Mark Govett

Mark Govett

Mark Govett

Mark Govett

Chief, Advanced Computing Section

Global Systems Division


Biography

I manage the Advanced Computing Section, a software group that supports model development, parallelization, and porting to high performance computers, and explores advanced computing technologies for GSD, ESRL and NOAA. I have a background in high-performance computing, code parallelization and compiler development.

Education

  • Bachelors of Science, Computer Science, University of Colorado, Denver, 1991
  • Masters of Science, Computer Science, University of Colorado, Boulder, 1995

Research Interests

I have diverse interests in high performance computing including Graphical Processors (GPUs), Cloud Computing, Distributed Data Access, Web Portals, and software engineering to improve model performance, portability, and interoperability. I believe GPUs are the next major advance in HPC and that Cloud Computing can be used to effectively distribute data. I believe portals are needed to support the development, configuration and testing of increasingly complex modeling systems. Finally I am interested in improving software engineering practices that enhance sharing models across NOAA and with our collaborators.

Accomplishments

  • Developed the Scalable Modeling System (SMS) compiler to translate Fortran codes into performance portable code that can be run on most HPC systems.
  • Led the exploration of Grid Computing at NOAA. Deliverables included developing a prototype grid that linked multiple NOAA sites.
  • Led the development of WRF Domain Wizard, and WRF Portal that support model development and testing.
  • Wrote a Fortran-to-GPU translator and used it to parallelize NIM. This code runs 25 times faster than on the CPU.

Recent Publications

  • The Grid: an IT Infrastructure for NOAA in the 21st Century, Proceedings of the Eleventh ECMWF Workshop, Reading, UK, Oct 2004.
  • The Scalable Modeling System: Directive-based Code Parallelization for Shared and Distributed Memory Computers, Journal of Parallel Computing, August 2003.