UO-logo
Home
Contact
Research
About
TAU-logo
APEX-logo
SOS-logo
Kevin Huck headshot

Kevin A. Huck

Faculty Research Associate
Oregon Advanced Computing Institute
for Science and Society (OACISS),
University of Oregon

  Email:
CV: PDF
Publications: List in Google Scholar (Contact me for PDF copies)
ORCID: 0000-0001-7064-8417
GitHub: Kevin Huck on GitHub
LinkedIn: Kevin Huck on LinkedIn
 
Research Projects:
  • RAPIDS: SciDAC Institute for Computer Science and Data
  • PHYLANX: A Distributed Array Toolkit
  • ECP PROTEAS-TUNE: Programming Toolchain for Emerging Architectures and Systems
  • ECP CODAR: Co-Design Center For Online Data Analysis and Reduction at Exascale
  • ECP SDK: Software Development Kit
  • Apollo 2.0: Online Machine Learning for Performance Portability

Software Projects:
Research Interests: I am currently doing research in the areas of data mining, parallel performance analysis, software tool design and software engineering. I am interested in the unique problems of intelligent, automated parallel performance analysis, specifically analysis of very large datasets (high processor counts, high event counts, multiple metrics). I am also interested in automated methods for diagnosing and treating single-node performance problems. Recently, I have become more interested in making complex HPC tools more usable.
 
Education:

Current & Former
Employers & Clients:

Useful Links:
TAU data rendered by Ascent during 64 process run of LULESH
Figure: LULESH 2.0.3 executed with 64 MPI ranks, measured by TAU. Time spent in main loop rendered every 100 timesteps by Alpine-Ascent. Clockwise from upper left: Computed energy, accumulated time in main loop, time in main loop during last 100 timesteps, Δ time in main loop from previous 100 time steps.

Valid XHTML 1.0 Strict