Intel® MPI Library
Get help with building, analyzing, optimizing, and scaling high-performance computing (HPC) applications.

How to manage disk access after splitting the communicator

Anders_S_1
New Contributor III
3,981 Views

Hi,

I have been using MPI  succesfully without splitting the group of threads into subgroups. Reading and writing of files have been handled by

a subroutine that allows serial access to the hard disk by the threads (see attached subroutine).

However, as I have an excess of threads for my application, I therefore want to use subgroups by use of  MPI_GROUP_SPLIT. in order

to get an additional speed increase. Can my treatment in the attached subroutine be generalized in some way in order to handle disk access?

Best regards

Anders S

 

0 Kudos
22 Replies
Anders_S_1
New Contributor III
589 Views

Hi Prasanth,

Thanks for your kind question!

Yes, i solved my problem. I use file storage at start and finish of each time step but I will change from file storage to using derived variables for 

saving data inside the time step. In this way I avoid the disk READ/WRITE problem and also get a speed up by working in the primary memory.

Thanks for the help!

Best regards

Anders S

0 Kudos
PrasanthD_intel
Moderator
589 Views

Hi Anders,

Glad to hear that the problem is resolved. We are closing this thread for now.

Please raise a new thread for further issues.

 

Thanks

Prasanth

0 Kudos
Reply