Skip to content

TimBashford/SharedMemoryMPI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Shared Memory MPI

A boost::mpi implementation for ensuring a specified physical machine is acting as the control node. Utilising rank 0 is naive because MPI itself assigns ranks to hostnames, but how this occurs is something of a black art. This implementation allows the programmer to specify a hostname and compare MPI_COMM_RANK against hnID to ensure the head node runs serial processes.

Designed for dual-mode parallelism (MPI to the enclosure, internal threading); deliberately halts single-mode parallelism implementations, but may easily be modified to alter this behaviour if desired; simply removing the check will likely yield success, with rank 0 ceding control to a process on the appropriate enclosure based on rank order.

By Tim Bashford - 16/04/11. Use as you see fit.

About

A boost::mpi implementation to reassign the serial post-processing node to a specified hostname.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages