GPU programming is now a much richer environment that it used to be a few years ago. On top of the two major programming languages, CUDA and OpenCL, libraries (e.g., cufft) and high level interfaces (e.g., thrust) have been developed that allow a fast access to the computing power of GPUs without detailed knowledge or programming of GPU hardware.
Annotation-based programming models (e.g., OpenACC), GPU plug-ins for existing mathematical software (e.g., Jacket in Matlab), GPU script languages (e.g., PyOpenCL), and new data parallel languages (e.g., Copperhead) bring GPU programming to a new level.
A major criticism of programming abstractions is that they look great on small examples but fail on practical problems. Therefore, this symposium invites, in particular, submissions that deal with practical applications that have successfully employed GPU libraries or high level programming tools. The focus may lie both on the development of the libraries or utilization of existing tools. Workshop topics include, but are not limited to:
- GPU applications coded with high level programming tools
- GPU library development and application
- Comparison of different programming abstractions on the same/similar applications
- Comparison of the same/similar programming abstractions on different applications
- Performance and coding effort of high level tools against hand-coded approaches on the GPU
- Performance and coding effort on multi-core CPUs against GPUs utilizing programming abstractions
- Classification of different programming abstractions with respect to their best application area
The highest quality papers of the minisymposium will receive an invitation to a special issue of the journal “Concurrency and Computation: Practice and Experience”.
Full CFP: Minisymposium on GPU Computing at the 10th International Conference on Parallel Processing and Applied Mathematics (PPAM). Note that PPAM will also host a full-day tutorial on Advanced GPU Programming.