Kernel Tuner
Kernel Tuner greatly simplifies the development of highly-optimized and auto-tuned CUDA, OpenCL, and C code, supporting many advanced use-cases and optimization strategies that speed up the auto-tuning process.
Verified construction of correct and optimised parallel software
GPUs have an increasingly big impact on industry and academia, due to their great computational capabilities. However, in practice, one usually needs to have expert knowledge on GPU architectures to optimally gain advantage of those capabilities. At the Eindhoven University of Technology, Wijs will work on modelling GPU applications using a Domain Specific Language, formally verifying the correctness of the models, and automatically generating GPU code. At the University of Twente, Huisman will work on the structured optimisation of GPU code, while ensuring that functional correctness is preserved. Existing formal verification techniques, model checking and code verification, will be combined to create, for the first time, a complete end-to-end development workflow for GPU applications.
To ensure the practical effectiveness of the resulting workflow, a users committee, consisting of SURFsara, the Netherlands eScience Center, Stream HPC, and CodePlay (UK), will provide real-life cases and provide feedback throughout the project.
Consolidating and Future-proofing Kernel Tuner by developing Software Engineering Best Practices
Supporting researchers to easily set-up benchmarks
Smart, secure container networks for trusted big data sharing
Providing computing solutions for exascale challenges
Automated multi-scale graph manipulation with topological and flow-based methods
From the Things to the Cloud and back
Boosting the performance of current and future programs
Storytelling as a means of visual data communication
Using point clouds to their full potential
Making breakthroughs in data-driven research
Developing an eScience technology platform
Kernel Tuner greatly simplifies the development of highly-optimized and auto-tuned CUDA, OpenCL, and C code, supporting many advanced use-cases and optimization strategies that speed up the auto-tuning process.