I'm not quite sure when you're invoking "cd" and it's not working -- is
it in main() of your MPI application? If so, that will only execute in
a sub shell and have no effect on the parent process. If you want a C
application to change directory, you can use the chdir(2) system call.
Hope that helps.
On Oct 6, 2004, at 4:43 PM, Sriram Rallabhandi wrote:
>
> I too thought about that. Thanks for replying.
>
> However, each node has to run some other applications based on the
> files created. It would be much simpler
> if I could separate the files into different directories.
>
> Sriram
>
>
> At 05:29 PM 10/6/2004 -0300, you wrote:
>
>> Hi Sriram,
>>
>> A simple solution could to let each rank output to a file
>> associated with itself, like output_1, output_2, ...and so on.
>> So that all of them could live in the same directory without being
>> overwritten.
>>
>> Regards,
>>
>> Roberto
>>
>>
>>
>> On Wed, 6 Oct 2004, Sriram Rallabhandi wrote:
>>
>> >
>> > Hi all,
>> >
>> > I have a shared space cluster and I'm trying to distribute my work
>> among
>> > various nodes. I create and use various files during each run on
>> each node.
>> > If nothing is done, the files created by one node will get
>> overwritten by
>> > the same files created by other nodes.
>> >
>> > So, I wish to create different directories for each node and build
>> all the
>> > files and use them in that directory. When all work is done, I
>> would like
>> > to get all the information back to the root and proceed to the final
>> > computations.
>> >
>> > I start my mpirun from /rootdirectory/
>> >
>> > The root node creates all files in this directory and all other
>> nodes
>> > should create the exact same files in directories /rootdirectory/1/,
>> > /rootdirectory/2/ ..etc.
>> >
>> > I have tried doing "cd itoa(rank)" from /rootdirectory/ for each
>> node
>> > before the actual computation begins. However, the pwd remains the
>> > /rootdirectory/
>> > and files are being overwritten.
>> >
>> > How can I do this? Is there any other way to do the above?
>> >
>> > Thanks
>> > Sriram
>> >
>> >
>> >
>> ----------------------------------------------------------------------
>> ---------
>> > Sriram K. Rallabhandi
>> > Graduate Research Assistant Work: 404 385 2789
>> > Aerospace Engineering Res: 404 603 9160
>> > Georgia Inst. of Technology
>> >
>> ----------------------------------------------------------------------
>> ---------
>> >
>> > _______________________________________________
>> > This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>> >
>>
>> _______________________________________________
>> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
> -----------------------------------------------------------------------
> --------
> Sriram K. Rallabhandi
> Graduate Research Assistant Work: 404 385 2789
> Aerospace Engineering Res: 404 603 9160
> Georgia Inst. of Technology
> -----------------------------------------------------------------------
> --------
>
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
