[MPICH] ROMIO error
Rajeev Thakur
thakur at mcs.anl.gov
Fri May 4 16:27:40 CDT 2007
Yong,
ROMIO is missing support for the MPI-2 function
MPI_Type_create_resized. Until we fix it, you can use MPI_type_struct with
an explicit MPI_LB and MPI_UB.
Rajeev
> -----Original Message-----
> From: owner-mpich-discuss at mcs.anl.gov
> [mailto:owner-mpich-discuss at mcs.anl.gov] On Behalf Of Yong Chen
> Sent: Friday, May 04, 2007 3:24 PM
> To: mpich-discuss at mcs.anl.gov
> Subject: [MPICH] ROMIO error
>
> Hi,
>
> I kept getting the following error when I run the sample
> MPI_File_set_view() code,
>
> Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks
> [cli_0]: aborting job:
> application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0
> rank 0 in job 125 x86_32838 caused collective abort of all ranks
> exit status of rank 0: killed by signal 9
>
> I did a debugging with gdb, the filetype caused the problem and
> MPI_File_set_view() threw the error. Indeed, if filetype is replaced
> with MPI_INT, it works well. Did a search on google, still
> cannot figure
> out why. Can anyone give me a clue? thanks very much! The
> following is
> the sample code.
>
> #include "mpi.h"
> int main( int argc, char *argv[] )
> {
> MPI_Aint lb, extent;
> MPI_Datatype etype, filetype, contig;
> MPI_Offset disp;
> MPI_File fh;
> int buf[1000];
>
> MPI_Init( &argc, &argv );
> /* ... other application code ... */
> MPI_File_open(MPI_COMM_WORLD, "datafile",
> MPI_MODE_CREATE | MPI_MODE_RDWR,
> MPI_INFO_NULL, &fh);
>
> MPI_Type_contiguous(2, MPI_INT, &contig);
> lb = 0;
> extent = 6 * sizeof(int);
> MPI_Type_create_resized(contig, lb, extent, &filetype);
> MPI_Type_commit(&filetype);
>
> disp = 5 * sizeof(int); /* assume displacement in
> this file view
> is of size equal to 5 integers */
> etype = MPI_INT;
> /* filetype = MPI_INT; */
> MPI_File_set_view(fh, disp, etype, filetype, "native",
> MPI_INFO_NULL);
> MPI_File_write(fh, buf, 1000, MPI_INT, MPI_STATUS_IGNORE);
> /* ... */
> MPI_Finalize();
> return 0;
> }
>
>
> Thanks,
> Yong
>
>
>
More information about the mpich-discuss
mailing list