Hello,
please, can somebody comment on the statement I found at several
places that coarray cannot be a part of a common block?
I would like to understand if there is a good reason for this...
It complicates easy parallelization of existing code a lot, because
instead of just changing things inside a common one has to create
modules, and then (since I can't change all commons into modules at
once) copy information between commons and modules (and have it
explicitly twice, in effect).
I believe that Cray complilers explicitly state that there *can* be
coarrays in commons... On the other and, gfortran seems to reject
that.
Thanks for any comments and explanations!
Best regards
Rudolf Sykora