High-Performance Computing w/ Joe Hummel Inc.
Joe Hummel, PhD

Downloads: PPT, code, and more

 

 

 

  • 8 April 2014: delivered 2 talks at Freescale's annual FTF conference in Dallas (FTF == Freescale Technology Forum). The talks were focused on Multicore software development, in particular understanding the critical role that caching plays in not only performance, but correctness.
    • Part1: Understanding Memory in Multicore World
    • Part2: Software Tools for Improving Cache Perf
    • Materials: PPT and demos

 

 

 

  • 12 October 2013: gave a 1-hour talk on "Going Parallel for Performance" in .NET, as part of Leetspeak 2013 in Stockholm. If you're interested in more details, the PPT and demo code can be found here. Note that the datafile driving the Netflix demo has to be downloaded separately.

 

  • 21 September 2013: 1.5 hr workshop on "Going Parallel with C++11" at the U. of Findlay, as part of CCSC-MW 2013; here's the talk, and here are the source materials.zip (PPT and demo code). Note that the datafile driving the Netflix demo has to be downloaded separately.

 

  • 2 September 2013: Freescale's "Make It Multicore" 1-day workshop running on Power-based QorIQ CPU, held in Porto Allegre, Brazil; materials here.

 

  • 01 August 2013: Freescale's "Make It Multicore" 1-day workshop running on ARM-based CPU, held in San Diego; materials here.

 

  • 28-29 May 2013: 2-day Workshop on "Going Parallel: Asynchronous and Parallel Programming in C++, F#, and Hadoop", held in Denmark. Here are the workshop materials.zip : PPT, demos, and lab exercises.

 

  • 22 May 2013: Freescale's "Make It Multicore" 1-day workshop, held in Cedar Rapids; materials here.

 

  • 30 April 2013: Freescale's "Make It Multicore" 1-day workshop, held in Chicago; materials here.

 

 

  • 23 April 2013: Freescale's "Make It Multicore" 1-day workshop, held in Minneapolis; materials here.

 

  • 7 March 2013:  at SIGCSE 2013, gave a short "lightning" talk as part of Intel's Lightning Round on how educators integrate parallel programming into their curriculums.  I spoke on how I introduce parallel programming in CS1 via matrix multiply, which is not just about multicore, but also data locality.  Here's the talk in PPT, and the MM source code I use in class (this is Visual Studio 2012 C++ code).

 

Website Builder