On Dec 23, 2014, at 5:49 AM, Anton Shterenlikht <[log in to unmask]> wrote:
>> From: John Reid <[log in to unmask]>
>>>
>>> Are there any plans to extend these intrinsics
>>> to accept complex args?
>>>
>>
>> No. The list of extensions beyond those in the two TSs (on further
>> interoperability and further coarray features) is in WG5 document N2023.
>
> The first item on that list is
>
> 1 UK-12 J3/14-174 Allow a file to be connected to more than one unit
> at one time 8-2-5
>
> This sounds like parallel I/O, which I thought was
> decisively ruled out of the next revision. Am I wrong?
The change was to make it processor dependent whether this is allowed, rather than strictly prohibited as was the case previously. There are no rules specified for how this works. However, it is certainly reasonable and straightforward to have a read-only file connected on multiple images, and this has been done in programs for years. The change was to allow this practice to be not in violation of the standard. It remains to be seen whether implementations will permit multiple units that allow writing to the same file.
Cheers,
Bill
>
> Thanks
>
> Anton
Bill Long [log in to unmask]
Fortran Technical Suport & voice: 651-605-9024
Bioinformatics Software Development fax: 651-605-9142
Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101
|