Re: [gomp3.1] Allow pointers and cray pointers in firstprivate/lastprivate, handle not allocated allocatable in firstprivate

[Tobias Burnus <burnus@net-b.de>]

Thomas Koenig wrote:
>> This patch includes assorted OpenMP 3.1 changes for Fortran.
>> Haven't changed COPYIN with not allocated allocatables yet, waiting
>> for explanation on OpenMP forum there.
>
> I'm not an OpenMP expert, but I'd say this is OK for trunk (unless 
> somebody else speaks up, quickly :-)

I suggest to wait until OpenMP 3.1 is released - and then merge over all 
of the gomp-3_1-branch. I think that also what Jakub planed to do. 
Nevertheless, proof reading patches is good.

OpenMP 3.1 timeline:
- Public review announced Frebruary 5
- End of public comment period: "Monday, May 1, 2011" (SIC!)
- Scheduled release: At IWOMP 2011 (June 13-15)
Cf. http://openmp.org/wp/2011/02/31-draft-specs-ready-for-public-comment/

Tobias

PS: Jakub was already rather active; the following patches have been 
committed to the branch. Missing is the support for the OMP_PROC_BIND 
environment variable and the actual implementation of the final and 
mergeable clauses, taskyield construct and in_final function - plus some 
"minor" stuff. The commits:

http://gcc.gnu.org/ml/gcc-cvs/2011-02/msg01012.html
   Branch for OpenMP 3.1 implementation.
http://gcc.gnu.org/ml/gcc-cvs/2011-02/msg01021.html
   omp_in_final: Prototype
http://gcc.gnu.org/ml/gcc-cvs/2011-03/msg00081.html
   min/max reductions
http://gcc.gnu.org/ml/gcc-cvs/2011-03/msg00158.html
   OMP_NUM_THREADS
http://gcc.gnu.org/ml/gcc-cvs/2011-03/msg00163.html
   Docu update
http://gcc.gnu.org/ml/gcc-cvs/2011-03/msg00270.html
   C/C++ checking
http://gcc.gnu.org/ml/gcc-cvs/2011-04/msg00915.html
   Fortran checking and extensions:
   Handle non-allocated allocatable, Allow POINTERs
   Cray pointers in clauses other than REDUCTION
http://gcc.gnu.org/ml/gcc-cvs/2011-04/msg01019.html
   Handle atomics in C
http://gcc.gnu.org/ml/gcc-cvs/2011-04/msg01209.html
   Handle atomics in C++
http://gcc.gnu.org/ml/fortran/2011-04/msg00280.html
   Handle atomics in Fortran
http://gcc.gnu.org/ml/gcc-patches/2011-04/msg02122.html
   Fix atomic write for C/C++
http://gcc.gnu.org/ml/gcc-patches/2011-04/msg02128.html
   Update libgomp.texi
http://gcc.gnu.org/ml/gcc-patches/2011-04/msg02230.html
   Add support for parsing final and mergeable task, taskyield