James M. Rosinski

SKILLS SUMMARY

• 35 years software engineering experience in C, C++, and Fortran.
• Distributed memory parallel software development using MPI.
• Threaded programming using OpenMP and pthreads under UNIX.
• Many years experience working on large codes in a team environment.
• Code optimization for cache and vector architectures.
• Object-oriented analysis and design.
• Scripting in Perl, bash, csh
• Operating System internals.
• Oral and written presentations.

EXPERIENCE

• 2010--present: Research Associate, NOAA/ESRL, Boulder, CO

  • Work with team developing next-generation data assimilation software for weather forecast models (https://www.jcsda.noaa.gov). Thus far have successfully implemented thread-based parallelism in radiation code providing up to 8X speedup on a 12-core socket.
  • Port my GPTL timing library (https://jmrosinski.github.io/GPTL) to GPU using OpenACC and CUDA to enable finer-grain timing than kernel-level, and include more information than provided by currently available GPU profilers.
  • Optimize performance of FIM, NIM, and FV3 forecast models on various supercomputing platforms, including Xeon Phi.

• 2008--2010: Climate Computational Scientist, Oak Ridge National Lab, Oak Ridge, TN

  • Port climate models and other large application codes to Cray XT5 and other on-site hardware.
  • Optimize performance of application codes on supercomputing platforms.
  • Study application scalability and offer improvements where possible.

• 2006--2008: Senior Applications Engineer, SiCortex, inc., Maynard, MA

  • Optimize performance of customer applications for start-up company (now defunct) which built high-performance Linux clusters from the silicon up.
  • Develop open-source performance analysis tools, and integrate them into a set of supported software. Many are based on PAPI.
  • Various customer-facing roles to sell and support the product line, including site visits.
  
  

• 2005--2006: Software Engineer, The Boeing Corporation, Kihei, HI

  • Develop TCP sockets-based software to parallelize application codes for optimal performance on Linux clusters and other distributed architectures.
  • Refine earlier work on general-purpose timing library (GPTL) to minimize overhead.
  
  

• 1986--2005: Software Engineer, National Center for Atmospheric Research, Boulder, CO

  • Part of a team building Community Atmosphere Model (CAM). This model is used by researchers worldwide for climate studies.
  • Parallelized CAM and its precursors, CCM2 and CCM3, for optimal execution on shared-memory and distributed-memory multi-processor platforms.
  • Devised and published a method to validate ports of atmospheric numerical models to new computing environments.
  • Wrote thread-safe general-purpose timing library (GPTL) for profiling application performance. It is callable from C, C++, and Fortran.
  • Presented tutorials on running climate models as part of CCM workshops.
  • Served on procurement teams for acquiring new supercomputing hardware.
  
  

• 1983--1986: Software Engineer, Applied Research Corporation, Landover, MD.

  • Implemented and modified atmospheric general circulation model on CDC Cyber 205 vector computer.
  • Vectorized code and developed post-processing software.

EDUCATION

Master of Science, Computer Science,
University of Colorado-Boulder                    December, 2001

• Wrote MPI-adherent message-passing library. It optionally uses TCP sockets or named pipes as the underlying communication mechanism.
• Modified Linux kernel to implement fair-share process scheduling.

Master of Science, Meteorology,
University of Wisconsin-Madison                   1983

• Thesis topic: diagnostic study of hurricane transition to an extra-tropical cyclone.
• Presented lecture material to large class of undergraduates.

Bachelor of Science, Meteorology,
McGill University, Montreal, Quebec               1980

• Obtained Minor in Computer Science.

PUBLICATIONS

• Govett, M., J. Rosinski, J. Middlecoff, T. Henderson, A. MacDonald, N. Wang, P. Madden, J. Schramm, A. Duarte (2017), Parallelization and Performance of the NIM Weather Model on CPU, GPU and MIC Processors. Bulletin of the American Meteorological Society, doi: https://doi.org/10.1175/BAMS-D-15-00278.1
• Carpenter, et. al. (2012), Progress towards accelerating HOMME on hybrid multi-core systems International Journal of High Performance Computing Applications http://hpc.sagepub.com/content/early/2012/11/15/1094342012462751November 16, 2012
• Hurrell, J., J.J. Hack, D. Shea, J. Caron, and J.M. Rosinski (2008), A new sea surface temperature and sea ice boundary data set for the Community Atmosphere Model. Journal of Climate vol. 21, pp 5145-5153.
• Williamson, D.L., and J.M. Rosinski (2000), Accuracy of reduced grid calculations, Quarterly Journal of the Royal Meteorological Society, vol. 126, pp. 1619-1640.
• Rosinski, J.M., and D.L. Williamson (1997), The accumulation of rounding errors and port validation for global atmospheric models, SIAM J. Scientific Computing, vol. 18, No. 2, pp. 552-564.
• Hack, J.J., J.M. Rosinski, D.L. Williamson, B.A. Boville, and J.E. Truesdale (1995), Computational design of the NCAR community climate model, Parallel Computing, vol. 21, pp. 1545-1569.
• Middleton, P., J.S. Chang, J. del Corral, and J. Rosinski (1988), Comparison of RADM and OSCAR Precipitation Chemistry Data, Atmospheric Environment, vol. 22, pp. 1192-1208.