The documents distributed by this server have been provided by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a noncommercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.
Publications of SPCL
|T. Hoefler, M. Mosch, T. Mehlan, W. Rehm:|
|CollGM - A Myrinet/GM optimized collective component for Open MPI|
(In Proceedings of 3rd KiCC Workshop 2007, presented in Aachen, Germany, RWTH Aachen, Dec. 2007)
AbstractThe Open MPI collective framework offers a way to implement hardware-specific collective operations for Open MPI. We used this framework to develop a Myrinet/GM collective component by combining common knowledge of the implementation of collective algorithms with GM protocol optimized techniques to achieve highest performance. Our results show that the good performance of the existing point-to-point based tuned collective implementation in Open MPI can be improved with the use of these techniques.