Re: [netcdf-hdf] bug in MPI cleanup


On Sep 13, 2007, at 2:00 PM, Robert Latham wrote:


On Thu, Sep 13, 2007 at 01:52:31PM -0500, Quincey Koziol wrote:
I would advise not hooking MPI-IO cleanup into atexit(3): you already 
correctly make the user call MPI_Init and MPI_Finalize.  Can hdf5
cleanup instead occur as part of nc_close?


        The way we normally suggest taking care of this sort of issue in MPI
programs is for the application program to call H5close() immediately
before calling MPI_Finalize().  Will that work for you?


Really?  Call H5Close inside a netcdf4 code?  Well, I can do that,
sure.  That seems to lack a certain symmetry, no?
I agree with you, but I don't think there's a corresponding "shut the netCDF-4 library down" API routine. :-) 

        Perhaps Ed or Russ can speak to that issue/possibility...

        Quincey

Attachment: smime.p7s
Description: S/MIME cryptographic signature