Issue No. 08 - August (2012 vol. 45)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MC.2012.255
Stanley Tzeng , University of California, Davis
Brandon Lloyd , Microsoft
John D. Owens , University of California, Davis
A task-parallel approach to programming commodity graphics hardware is useful for implementing irregular parallel workloads with dependencies, particularly for applications such as video encoding and backtracking algorithms. The featured Web extra is a video that demonstrates how to use a GPU task-parallel model for H.264 intra prediction. The authors first describe the dependency structure and then how to map parallel work units to a problem. YouTube URL: http://youtu.be/K4VB1h4xres
Graphics processing unit, Programming, Hardware, Encoding, Prediction algorithms, Predictive models, Youtube, Multithreading, multithreading, massively threaded systems, graphics processors, parallel processing, video coding, GPU, CPU
S. Tzeng, B. Lloyd and J. D. Owens, "A GPU Task-Parallel Model with Dependency Resolution," in Computer, vol. 45, no. , pp. 34-41, 2012.