Main Page

From Pheet
Jump to: navigation, search

Pheet is a template-based C++ library in the tradition of Intel Threading Building Blocks and C++11 tasks. It provides an easy to use programming interface, while maintaining a good performance and a low overhead of scheduling tasks. The main focus of Pheet lies on configurability. Almost every performance relevant aspect in the system can be replaced in a modular manner at compile time, which allows to quantify the contribution of different scheduler components and data-structures to application performance. The system also allows to compile multiple configurations of a component into a single program.

Contents

Getting started

Documentation

Pheet Publications

  • Martin Wimmer, Jesper Larsson Träff. Work-stealing for mixed-mode parallelism by deterministic team-building. In 23rd ACM Symposium on Parallelism in Algorithms and Architectures (SPAA 2011), pages 105-115. ACM Press, 2011.
  • Martin Wimmer, Jesper Larsson Träff. An extended work-stealing framework for mixed-mode parallel applications. In Workshop on Multi-threaded Architectures and Applications (MTAAP) at 25th IEEE International Parallel and Distributed Processing Symposium (IPDPS), page 48. IEEE Press, 2011.
  • Martin Wimmer, Daniel Cedermann, Jesper Larsson Träff, Philippas Tsigas. Work-stealing with configurable strategies. In 18th ACM Symposium on Principles & Practice of Parallel Programming (PPoPP), 2013.
  • Martin Wimmer. Wait-free Hyperobjects for Task-parallel Programming Systems. In 27th IEEE International Parallel and Distributed Processing Symposium (IPDPS), 2013.
  • Martin Wimmer, Manuel Pöter, Jesper Larsson Träff. The Pheet Task-Scheduling Framework on the Intel® Xeon Phi™ Coprocessor and other Multicore Architectures. In Workshop on Multi-threaded Architectures and Applications (MTAAP) at 27th IEEE International Parallel and Distributed Processing Symposium (IPDPS), 2013.
  • Martin Wimmer, Francesco Versaci, Jesper Larsson Träff, Daniel Cederman, Philippas Tsigas. Data structures for task-based priority scheduling. In 19th ACM Symposium on Principles and Practice of Parallel Programming (PPoPP), 2014.

Topics related to the Advanced Multiprocessor Programming lecture

Personal tools
Namespaces

Variants
Actions
Navigation
Tools