(j3.2006) an alternative to Aleks' asynchronous proposal
Fri Jun 20 12:54:37 EDT 2008
On Friday 20 June 2008 08:05, Lionel, Steve wrote:
> unless we
> are to add a block construct with syntax listing the variables that
> should be considered asynchronous in that block. ?
This is already there in Fortran 2008 via the BLOCK construct:
ASYNCHRONOUS :: a, b
The problem is that it cannot be split across programming units, so you can
start the transfer at one point and end elsewhere (isn't this common?).
> Should a statement be added saying that there is an implicit memory
> synchronization point at the "return" point for the current procedure?
My proposed SYNC attribute for the procedure was something like this, though
maybe there are other better ways.
More information about the J3