Kevin A. Brown, Ph.D.

Kevin A. Brown, Ph.D.

Argonne Scholar

Argonne National Laboratory

Biography

I am the inaugural Walter Massey Fellow at Argonne National Laboratory, a US Department of Energy national research laboratory.

I works in the area of network and I/O performance analysis as the PI for the project Improving Data Movement Performance for Emerging AI and Climate Science Workloads on Future Supercomputers, a 3-year Laboratory Directed Research Development (LDRD) Program funded project.

My profile page at Argonne is here.

Interests

  • Supercomputing
  • Performance measurement and analysis
  • Network interconnects and I/O
  • Artificial Intelligence
  • Climate Science

Education

  • PhD in Mathematical and Computing Sciences, 2018

    Tokyo Institute of Technology

  • MSc in Mathematical and Computing Sciences, 2014

    Tokyo Institute of Technology

  • BSc in Computing and Information Technology, 2008

    University of Technology, Jamaica

Experience

 
 
 
 
 

Argonne Scholar - Walter Massey Fellow

Argonne National Laboratory

Oct 2021 – Present Illinois, USA
  • Principal Investigator (PI) for the project “Improving Data Movement Performance for Emerging AI and Climate Science Workloads on Future Supercomputers”.
  • Using machine learning to configure network quality-of-service.
 
 
 
 
 

Postdoctoral Appointee - Exascale Interconnect Performance Evaluation

Argonne National Laboratory

Oct 2019 – Present Illinois, USA
  • Member of the Performance Engineering Group in the Argonne Leadership Computing Facility (ALCF).
  • Optimising network design and communication strategies for future generation supercomputers.
  • Using the CODES network simulation toolkit to explore novel network architectures, such as HPE’s Slingshot.
 
 
 
 
 

Research Assistant

AIST-Tokyo Tech Real World Big-Data Computing Open Innovation Laboratory (RWBC-OIL)

Oct 2017 – Sep 2018 Tokyo, Japan
  • Conducted I/O-MPI interference study at a joint lab between Tokyo Institute of Technology and the National Institute of Advanced Information Science and Technology (AIST).
  • Illustrated the interference trends between various I/O and MPI traffic patterns on fat-tree networks.
 
 
 
 
 

Summer Intern

Lawrence Livermore National Laboratory

Aug 2016 – Nov 2016 California, USA
  • Used the CODES simulation toolkit to investigate the interference between I/O and MPI traffic on fat-tree networks.
 
 
 
 
 

Summer Intern

Lawrence Livermore National Laboratory

Jun 2015 – Sep 2015 California, USA
  • Designed and implemented web-based, occlusion-free, topology-accurate visualization of fat-tree networks using d3.js.
 
 
 
 
 

Visiting Researcher

Barcelona Supercomputing Center (BSC-NCS)

Mar 2015 – Apr 2015 Barcelona, Spain
  • Extended the ibprof profiler to integrate traffic from multiple applications in order to compare traffic from multiple applications.
 
 
 
 
 

Unix Systems Administrator

Digicel Jamaica, Limited

Jun 2008 – Mar 2012 Kingston, Jamaica
  • Installed, configured, and monitored all production servers running Linux and Solaris.
  • Managed the organisation’s backup and recovery infrastructure.
  • Oversaw the upgrade of the organisations Netbackup solution and Quantum tape library.
  • Migrated production systems to virtual environments using VMWare ESZXi Servers.

Accomplish­ments

Japanese Government (Monbukagakusho: MEXT) Scholarships

Nominee - Prime Minister’s Youth Award for Excellence

UTech, Ja. Foundation Scholarship

Recent Publications

(2018). Interference Between I/O and MPI Traffic on Fat-tree Networks. Proceedings of the 47th International Conference on Parallel Processing.

PDF DOI

(2017). Co-locating Graph Analytics and HPC Applications. 2017 IEEE International Conference on Cluster Computing (CLUSTER).

DOI

Recent Posts

Keynote Speech at the 2021 ILSAMP Symposium

On February 2021, I presented the keynote the 2021 ILSAMP Annual Spring Symposium & Student Research Conference titled “Finding Opportunities in Everyday Encounters: An Agile Approach to Dealing with Life’s Unpredictability”.

Avoiding Type 3 Fonts in matplotlib Plots

So you’ve generated your figure/PDF with matplotlib and your conference/journal publishers tell you that you have unsupported font types. Chances are, based on my experience with ACM, you have some Type 3 fonts.

Contact