@inbook {875, title = {A New Approach to MPI Collective Communication Implementations}, booktitle = {Distributed and Parallel Systems}, year = {2007}, pages = {45-54}, publisher = {Springer US}, organization = {Springer US}, abstract = {Recent research into the optimization of collective MPI operations has resulted in a wide variety of algorithms and corresponding implementations, each typically only applicable in a relatively narrow scope: on a specific architecture, on a specific network, with a specific number of processes, with a specific data size and/or data-type {\textendash} or any combination of these (or other) factors. This situation presents an enormous challenge to portable MPI implementations which are expected to provide optimized collective operation performance on all platforms. Many portable implementations have attempted to provide a token number of algorithms that are intended to realize good performance on most systems. However, many platform configurations are still left without well-tuned collective operations. This paper presents a proposal for a framework that will allow a wide variety of collective algorithm implementations and a flexible, multi-tiered selection process for choosing which implementation to use when an application invokes an MPI collective function.}, keywords = {Automatic Selection, Collective Operation, Framework, Message Passing (MPI), Open MPI}, isbn = {978-0-387-69857-1}, doi = {10.1007/978-0-387-69858-8_5}, author = {Torsten Hoefler and Jeffrey M. Squyres and Graham Fagg and George Bosilca and Wolfgang Rehm and Andrew Lumsdaine} }