public inbox for fortran@gcc.gnu.org
 help / color / mirror / Atom feed
* Fwd: This OpenMP code doesn't seem to compile correctly (5.4, 6.1).
       [not found] <ac471fb9-a142-2104-39b9-8a41dbd0bc5b@moene.org>
@ 2016-06-28 19:16 ` Toon Moene
  2016-06-28 19:59   ` Anton Shterenlikht
  0 siblings, 1 reply; 4+ messages in thread
From: Toon Moene @ 2016-06-28 19:16 UTC (permalink / raw)
  To: fortran

[-- Attachment #1: Type: text/plain, Size: 776 bytes --]




-------- Forwarded Message --------
Subject: This OpenMP code doesn't seem to compile correctly (5.4, 6.1).
Date: Tue, 28 Jun 2016 21:13:43 +0200
From: Toon Moene <toon@moene.org>
Organization: Moene Computational Physics, Maartensdijk, The Netherlands
To: fortran@gcc.gn.org

See the attached.

I can't find fault with the code, yet it produces:

     99999.0000       99999.0000       99999.0000       99999.0000

instead of zeros.

Is this indeed an error (in that case I will open a BugZilla report).

Kind regards,

-- 
Toon Moene - e-mail: toon@moene.org - phone: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
At home: http://moene.org/~toon/; weather: http://moene.org/~hirlam/
Progress of GNU Fortran: http://gcc.gnu.org/wiki/GFortran#news


[-- Attachment #2: elkhatib.f90 --]
[-- Type: text/x-fortran, Size: 1179 bytes --]

MODULE MYFIELDS_MOD

! The definition of a derived type containing an allocatable array

IMPLICIT NONE
SAVE

TYPE :: MYFIELDS
  REAL, ALLOCATABLE :: GMVT1S(:)
END TYPE MYFIELDS

END MODULE MYFIELDS_MOD

PROGRAM GP_MODEL

! A program to allocate then use a structure containing a allocatable array

USE MYFIELDS_MOD , ONLY : MYFIELDS

IMPLICIT NONE

TYPE(MYFIELDS) :: YRFIELDS

ALLOCATE(YRFIELDS%GMVT1S(4))

YRFIELDS%GMVT1S=99999.

CALL GP_TEST(YRFIELDS)

END PROGRAM GP_MODEL

SUBROUTINE GP_TEST(YDFIELDS)

! A subroutine to initialize the component of a structure, which is an
! allocatable array, via an association

! DOES NOT WORK SINCE GFORTRAN 4.9.2 (4.9.1 ?), UNLESS ONE OF THE FOLLOWING CONDITIONS IS RESPECTED :
! - Open-mp is disabled
! - the attribute is POINTER instead of ALLOCATABLE
! - the association is not used
! COULD BE AN ISSUE WITH Version 4.0 of the OpenMP specification

USE MYFIELDS_MOD , ONLY : MYFIELDS

IMPLICIT NONE

TYPE(MYFIELDS), INTENT(INOUT)      :: YDFIELDS

INTEGER :: J

ASSOCIATE(YDGMV=>YDFIELDS)

!$OMP PARALLEL DO
DO J=1,4
  YDGMV%GMVT1S(J)=0.
ENDDO
!$OMP END PARALLEL DO

print*,YDFIELDS%GMVT1S(:)

END ASSOCIATE

END SUBROUTINE GP_TEST


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: This OpenMP code doesn't seem to compile correctly (5.4, 6.1).
  2016-06-28 19:16 ` Fwd: This OpenMP code doesn't seem to compile correctly (5.4, 6.1) Toon Moene
@ 2016-06-28 19:59   ` Anton Shterenlikht
  2016-06-28 21:40     ` Toon Moene
  0 siblings, 1 reply; 4+ messages in thread
From: Anton Shterenlikht @ 2016-06-28 19:59 UTC (permalink / raw)
  To: Toon Moene; +Cc: fortran

mech-as222> uname -a
FreeBSD mech-as222.men.bris.ac.uk 10.3-RELEASE-p4 FreeBSD
10.3-RELEASE-p4 #0: Sat May 28 12:23:44 UTC 2016
root@amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC  amd64


mech-as222> gfortran6 -Wl,-rpath="/usr/local/lib/gcc6" z.f90
mech-as222> ./a.out
   0.00000000       0.00000000       0.00000000       0.00000000
mech-as222> gfortran6 --version
GNU Fortran (FreeBSD Ports Collection) 6.0.0 20150927 (experimental)

mech-as222> gfortran5 -Wl,-rpath="/usr/local/lib/gcc5" z.f90
mech-as222> ./a.out
   0.00000000       0.00000000       0.00000000       0.00000000
mech-as222> gfortran5 --version
GNU Fortran (FreeBSD Ports Collection) 5.3.0

Anton

On 28/06/2016, Toon Moene <toon@moene.org> wrote:
>
>
>
> -------- Forwarded Message --------
> Subject: This OpenMP code doesn't seem to compile correctly (5.4, 6.1).
> Date: Tue, 28 Jun 2016 21:13:43 +0200
> From: Toon Moene <toon@moene.org>
> Organization: Moene Computational Physics, Maartensdijk, The Netherlands
> To: fortran@gcc.gn.org
>
> See the attached.
>
> I can't find fault with the code, yet it produces:
>
>      99999.0000       99999.0000       99999.0000       99999.0000
>
> instead of zeros.
>
> Is this indeed an error (in that case I will open a BugZilla report).
>
> Kind regards,
>
> --
> Toon Moene - e-mail: toon@moene.org - phone: +31 346 214290
> Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
> At home: http://moene.org/~toon/; weather: http://moene.org/~hirlam/
> Progress of GNU Fortran: http://gcc.gnu.org/wiki/GFortran#news
>
>

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: This OpenMP code doesn't seem to compile correctly (5.4, 6.1).
  2016-06-28 19:59   ` Anton Shterenlikht
@ 2016-06-28 21:40     ` Toon Moene
  2016-06-28 22:15       ` Anton Shterenlikht
  0 siblings, 1 reply; 4+ messages in thread
From: Toon Moene @ 2016-06-28 21:40 UTC (permalink / raw)
  To: fortran

On 06/28/2016 09:59 PM, Anton Shterenlikht wrote:

> mech-as222> uname -a
> FreeBSD mech-as222.men.bris.ac.uk 10.3-RELEASE-p4 FreeBSD
> 10.3-RELEASE-p4 #0: Sat May 28 12:23:44 UTC 2016
> root@amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC  amd64

Fascinating - a colleague of mine thought it might be related to a 
kernel problem. I am using:

toon@moene:~/src$ uname -a
Linux moene.org 4.6.0-1-amd64 #1 SMP Debian 4.6.1-1 (2016-06-06) x86_64 
GNU/Linux

> mech-as222> gfortran6 -Wl,-rpath="/usr/local/lib/gcc6" z.f90
> mech-as222> ./a.out
>    0.00000000       0.00000000       0.00000000       0.00000000
> mech-as222> gfortran6 --version
> GNU Fortran (FreeBSD Ports Collection) 6.0.0 20150927 (experimental)
>
> mech-as222> gfortran5 -Wl,-rpath="/usr/local/lib/gcc5" z.f90
> mech-as222> ./a.out
>    0.00000000       0.00000000       0.00000000       0.00000000
> mech-as222> gfortran5 --version
> GNU Fortran (FreeBSD Ports Collection) 5.3.0
>
> Anton
>
> On 28/06/2016, Toon Moene <toon@moene.org> wrote:
>>
>>
>>
>> -------- Forwarded Message --------
>> Subject: This OpenMP code doesn't seem to compile correctly (5.4, 6.1).
>> Date: Tue, 28 Jun 2016 21:13:43 +0200
>> From: Toon Moene <toon@moene.org>
>> Organization: Moene Computational Physics, Maartensdijk, The Netherlands
>> To: fortran@gcc.gn.org
>>
>> See the attached.
>>
>> I can't find fault with the code, yet it produces:
>>
>>      99999.0000       99999.0000       99999.0000       99999.0000
>>
>> instead of zeros.
>>
>> Is this indeed an error (in that case I will open a BugZilla report).
>>
>> Kind regards,
>>
>> --
>> Toon Moene - e-mail: toon@moene.org - phone: +31 346 214290
>> Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
>> At home: http://moene.org/~toon/; weather: http://moene.org/~hirlam/
>> Progress of GNU Fortran: http://gcc.gnu.org/wiki/GFortran#news
>>
>>
>


-- 
Toon Moene - e-mail: toon@moene.org - phone: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
At home: http://moene.org/~toon/; weather: http://moene.org/~hirlam/
Progress of GNU Fortran: http://gcc.gnu.org/wiki/GFortran#news

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: This OpenMP code doesn't seem to compile correctly (5.4, 6.1).
  2016-06-28 21:40     ` Toon Moene
@ 2016-06-28 22:15       ` Anton Shterenlikht
  0 siblings, 0 replies; 4+ messages in thread
From: Anton Shterenlikht @ 2016-06-28 22:15 UTC (permalink / raw)
  To: fortran, toon

>On 06/28/2016 09:59 PM, Anton Shterenlikht wrote:
>
>> mech-as222> uname -a
>> FreeBSD mech-as222.men.bris.ac.uk 10.3-RELEASE-p4 FreeBSD
>> 10.3-RELEASE-p4 #0: Sat May 28 12:23:44 UTC 2016
>> root@amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC  amd64
>
>Fascinating - a colleague of mine thought it might be related to a 
>kernel problem. I am using:
>
>toon@moene:~/src$ uname -a
>Linux moene.org 4.6.0-1-amd64 #1 SMP Debian 4.6.1-1 (2016-06-06) x86_64 
>GNU/Linux

And with 4.8:

mech-as222> gfortran48 -Wl,-rpath="/usr/local/lib/gcc48" z.f90
mech-as222> ./a.out 
   0.00000000       0.00000000       0.00000000       0.00000000    

Note that your versions of 6 and 5 branches are higher than mine.
Perhaps a regression?

Anton

>> mech-as222> gfortran6 -Wl,-rpath="/usr/local/lib/gcc6" z.f90
>> mech-as222> ./a.out
>>    0.00000000       0.00000000       0.00000000       0.00000000
>> mech-as222> gfortran6 --version
>> GNU Fortran (FreeBSD Ports Collection) 6.0.0 20150927 (experimental)
>>
>> mech-as222> gfortran5 -Wl,-rpath="/usr/local/lib/gcc5" z.f90
>> mech-as222> ./a.out
>>    0.00000000       0.00000000       0.00000000       0.00000000
>> mech-as222> gfortran5 --version
>> GNU Fortran (FreeBSD Ports Collection) 5.3.0

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2016-06-28 22:15 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <ac471fb9-a142-2104-39b9-8a41dbd0bc5b@moene.org>
2016-06-28 19:16 ` Fwd: This OpenMP code doesn't seem to compile correctly (5.4, 6.1) Toon Moene
2016-06-28 19:59   ` Anton Shterenlikht
2016-06-28 21:40     ` Toon Moene
2016-06-28 22:15       ` Anton Shterenlikht

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).