“Voxel Modeling: Introduction to Applied GPU-based Parallel Computing” Course

Posted on by

Winter Quarter 2016 Special Course Offering for Graduate Students and Undergraduates with Instructor permission: Current Topics in Computer-Aided Engineering:

Voxel Modeling: Introduction to Applied GPU-based Parallel Computing:

  • Time Schedule: MW (and some Fridays) from 2:30 to 3:20 PM
  • Location: MEB Rm 234
  • Credits: 3 credits
  • Instructor: Professor Duane Storti (storti@uw.edu)
  • Registration (email for instructor permission):
    • ME 599, SLN #16714
    • For ME Masters: ME 599O, SLN#16721 or Registration #15265
    • ME498A, SLN#16660 (with instructor permission)

Course Overview:

  •  Introduction to GPU-based parallel computing using CUDA including:
    • Setting up a CUDA-capable system
    • C language basics
    • Essential CUDA language: APIs and C extensions
    • Introductory projects to get you up the CUDA learning curve
  • Applications to geometric modeling based on volumetric medical imaging
  • Need to know” CUDA memory tricks for:
    • Creating applications with live graphical interactions
    • Enhancing development or performance efficiency
  • A final project that you propose and implement:
    • Choosing a project topic that is tightly couple with your research is encouraged.
    • You will get to give presentations to introduce your topic and present your results.

Course Description:

The course is very hands-on, so plan on writing CUDA code on a regular basis. CUDA is available for a variety of operating systems (Linux, OS X, and Windows) and uses an extension of C/C++, so experience with C/C++ programming is advantageous (but NOT required). The class is open to graduate students and highly motivated undergraduates, and students from other departments/colleges are welcome. Grading is based on projects, consulting interactions, and class presentations. Expected background includes some math (basic vector calculus, linear algebra, and differential equations are all helpful) and some programming experience. Many students already have access to a CUDA-enabled system (with a recent NVIDIA graphics card), and we will do our best to make arrangements for those who do not.

If you are interested in exploring the parallel computing capabilities supported by modern GPUs, you are encouraged to register and contact the instructor for further information