I spent a great summer in Boston, working at the Boston Design Center atAdvanced Micro Devices (AMD), under the direction of Norm Rubin. Perhaad Mistry and I worked on producing an OpenCL computer vision application of the “Speeded Up Robust Features” algorithm. We worked on analysis of the OpenCL code and on a few demonstrations of the interest-point detection algorithm, most notably a real-time video stabilization routine.
Author Archives: Chris
Contention-Aware Scheduling of Parallel Code for Heterogeneous Systems (update)
I presented the poster for our paper, Contention-Aware Scheduling of Parallel Code for Heterogeneous Systems at HotPar 2010.
Contention-Aware Scheduling of Parallel Code for Heterogeneous Systems
Another computer engineering student, Jeff Brantley, and I (along with Kim Hazelwood) wrote a paper on contention-aware dynamic scheduling for heterogeneous systems. We utilized OpenCL for the code, and ran our experiments on an AMD GPU and an Intel Core 2 Duo CPU.
Enabling Task Parallelism in the CUDA Scheduler
Marisabel Guevara and I, along with Kevin Skadron and Kim Hazelwood, wrote a paper entitled Enabling Task Parallelism in the CUDA Scheduler. It was accepted into the PMEA Workshop at the PACT 2009 Conference in Raleigh, NC. Marisabel will be presenting the paper.
Genetic Algorithms in Autonomous Embedded Systems
I took an Embedded Computing class during the Spring 2009 semester, and I did a project investigating how to use genetic algorithms to program my Roomba vacuum cleaner to go between two points on a floor with barriers in between: Genetic Algorithms in Autonomous Embedded Systems.
Measuring Unix
Marisabel Guevara and I wrote a paper on Measuring Unix for our Operating Systems Class.