MPI Programming Guide

Features of MPI-IO

The primary features of MPI-IO are:

1. Portability: As part of MPI-2, programs written to use MPI-IO must be portable across MPI-2 implementations and across hardware/software platforms. The PE MPI-IO implementation guarantees portability of object code on RS/6000 SP computers and clusters of IBM pSeries and RS/6000 workstations. The MPI-IO API ensures portability at the source code level.
2. Versatility: The PE MPI-IO implementation provides support for:
   • basic file manipulations (open, close, delete, sync)
   • get/set file attributes (view, size, group, mode, info)
   • blocking data access operations with explicit offsets (both independent and collective)
   • nonblocking data access operations with explicit offsets (independent only)
   • blocking and nonblocking data access operations with file pointers (individual and shared)
   • split collective data access operations
   • any derived datatype for memory and file mapping
   • file interoperability through data representations (internal, external, user-defined)
   • atomic mode for data accesses.
3. Robustness: PE MPI-IO performs as robustly as possible in the event of error occurrences. Because the default behavior, as dictated by the MPI-2 standard, is for I/O errors to return, PE MPI-IO tries to prevent any deadlock that might result from an I/O error returning. The intent of the "errors return" default is that the sort of errors considered almost routine in doing I/O should not be fatal in MPI ("file not found", for example). However, deadlocks resulting from erroneous user codes cannot be entirely avoided.