If I have a multi-process application, one rank of which generates a region reference into a dataset for a file open in that rank, but then sends that information to a rank managing a different file containing an external link to that dataset, is it possible to save that region reference in a dataset in the second file such that it "just works" via the external link in that second file? If not, is there something I could / should be doing instead? Parallel I/O to a single file is not really looking like an option in this case due to the possible numbers of files involved and the difficulty of managing collective dataset extension operations.
Thanks for any help
Chris Green <email@example.com>, FNAL CS/SCD/ADSS/SSI/TAC;
'phone (630) 840-2167; Skype: chris.h.green;
IM: firstname.lastname@example.org, chissgreen (AIM),
chris.h.green (Google Talk).