[mpich-discuss] Problems with MPI Romio code

[Rajeev Thakur]

MPI_File_set_atomicity(... true) enables stronger consistency semantics
similar to POSIX. The data written by one process can be read by another
process belonging to the same file open as long as the read occurs after
the write. The barrier ensures that the read happens after the write.

Rajeev

> -----Original Message-----
> From: mpich-discuss-bounces at mcs.anl.gov 
> [mailto:mpich-discuss-bounces at mcs.anl.gov] On Behalf Of shankha
> Sent: Friday, July 02, 2010 10:07 AM
> To: mpich-discuss at mcs.anl.gov
> Subject: [mpich-discuss] Problems with MPI Romio code
> 
> Hi,
> Your slides mention that MPI_Barrier should still be used along with
> MPI_Atomicity. Atomicty only ensures that other processes are not 
> performing the same operation but it does not ensure that 
> there cannot 
> be another operation on the same file.
> 
> e.g.
> MPI_open
> MPI_set_atomic
> MPI_File_write
> -------------------->   MPI_Barrier
> MPI_File_read
> 
> 
> Am I correct in my assumption and reasoning ?
> 
> 
> Thanks
> Shankha
> _______________________________________________
> mpich-discuss mailing list
> mpich-discuss at mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>