Re: [OMPI users] OBJ_CLASS_INSTANCE error: initializer element is not constant - Windows 7 64 bit mingw-w64 4.6.3

2011-12-28 Thread Roger Martin 

Hi Shiqing,

I was able to build the open-rte and the rest by applying g++.exe 
instead of gcc.exe.  I couldn't find any combination of --std=C9x to 
make the c compiling happy for gcc.exe.


For open-rte build I ran into issues with opal_ifisloopback impl 
unreferenced.  I'll regenerate my steps and post more discrete pieces.


Regards,
Roger


On 12/28/2011 06:15 PM, Shiqing Fan wrote:

Hi Roger,

To be honest, this MinGW support for Open MPI was integrated when the 
64 bit MinGW wasn't available, so I haven't tested it for the 64 bit 
build. I'll give a try, and come back to you soon.


Thanks for your report.

Regards,
Shiqing

On 2011-12-22 4:54 PM, Roger Martin wrote:

Hi,

Working on building on windows with MinGW.  Have built libopen-pal, 
opal-restart, opal-wrapper and am at the open-rte stage of the build.
The  OBJ_CLASS_INSTANCE hits errors about 'initializer element is not 
constant'. Open-pal also uses OBJ_CLASS_INSTANCE and doesn't throw 
this error.
Tried g++ instead of gcc and source such as rmcast_base_open.c 
compiled successfully but then I get into name mangling issues.


Being not familiar with the code I haven't yet determined why 
open-pal is successful and open-rte isn't.  Flag difference?  Is 
there code difference between them in how they apply the 
OBJ_CLASS_INSTANCE?


$ ../../mingw64/bin/gcc.exe --version
gcc.exe (GCC) 4.6.3 2005 (prerelease)

..line 325 of 
rmcast_base_open.c.

 OBJ_CLASS_INSTANCE(rmcast_base_send_t,
   opal_list_item_t,
   send_construct,
   send_destruct);


C:\Software\openmpi-build>gmake
[ 14%] Built target libopen-pal
[ 14%] Built target opal-restart
[ 14%] Built target opal-wrapper
[ 14%] Building C object 
orte/CMakeFiles/libopen-rte.dir/mca/rmcast/base/rmcast_

base_open.obj
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:325:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:325:1: 
erro

r: (near initialization for 'rmcast_base_send_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:363:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:363:1: 
erro

r: (near initialization for 'rmcast_base_recv_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:403:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:403:1: 
erro

r: (near initialization for 'rmcast_base_channel_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:413:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:413:1: 
erro

r: (near initialization for 'rmcast_seq_tracker_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:434:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:434:1: 
erro

r: (near initialization for 'rmcast_recv_log_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:451:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:451:1: 
erro

r: (near initialization for 'rmcast_send_log_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:466:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:466:1: 
erro

r: (near initialization for 'orte_rmcast_msg_t_class.cls_parent')
gmake[2]: *** 
[orte/CMakeFiles/libopen-rte.dir/mca/rmcast/base/rmcast_base_open.

obj] Error 1
gmake[1]: *** [orte/CMakeFiles/libopen-rte.dir/all] Error 2
gmake: *** [all] Error 2

___
users mailing list
us...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/users

Re: [OMPI users] OBJ_CLASS_INSTANCE error: initializer element is not constant - Windows 7 64 bit mingw-w64 4.6.3

2011-12-28 Thread Shiqing Fan 

Hi Roger,

To be honest, this MinGW support for Open MPI was integrated when the 64 
bit MinGW wasn't available, so I haven't tested it for the 64 bit build. 
I'll give a try, and come back to you soon.


Thanks for your report.

Regards,
Shiqing

On 2011-12-22 4:54 PM, Roger Martin wrote:

Hi,

Working on building on windows with MinGW.  Have built libopen-pal, 
opal-restart, opal-wrapper and am at the open-rte stage of the build.
The  OBJ_CLASS_INSTANCE hits errors about 'initializer element is not 
constant'. Open-pal also uses OBJ_CLASS_INSTANCE and doesn't throw 
this error.
Tried g++ instead of gcc and source such as rmcast_base_open.c 
compiled successfully but then I get into name mangling issues.


Being not familiar with the code I haven't yet determined why open-pal 
is successful and open-rte isn't.  Flag difference?  Is there code 
difference between them in how they apply the OBJ_CLASS_INSTANCE?


$ ../../mingw64/bin/gcc.exe --version
gcc.exe (GCC) 4.6.3 2005 (prerelease)

..line 325 of 
rmcast_base_open.c.

 OBJ_CLASS_INSTANCE(rmcast_base_send_t,
   opal_list_item_t,
   send_construct,
   send_destruct);


C:\Software\openmpi-build>gmake
[ 14%] Built target libopen-pal
[ 14%] Built target opal-restart
[ 14%] Built target opal-wrapper
[ 14%] Building C object 
orte/CMakeFiles/libopen-rte.dir/mca/rmcast/base/rmcast_

base_open.obj
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:325:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:325:1: 
erro

r: (near initialization for 'rmcast_base_send_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:363:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:363:1: 
erro

r: (near initialization for 'rmcast_base_recv_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:403:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:403:1: 
erro

r: (near initialization for 'rmcast_base_channel_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:413:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:413:1: 
erro

r: (near initialization for 'rmcast_seq_tracker_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:434:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:434:1: 
erro

r: (near initialization for 'rmcast_recv_log_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:451:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:451:1: 
erro

r: (near initialization for 'rmcast_send_log_t_class.cls_parent')
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:466:1: 
erro

r: initializer element is not constant
C:\Software\openmpi-1.5.5rc1\orte\mca\rmcast\base\rmcast_base_open.c:466:1: 
erro

r: (near initialization for 'orte_rmcast_msg_t_class.cls_parent')
gmake[2]: *** 
[orte/CMakeFiles/libopen-rte.dir/mca/rmcast/base/rmcast_base_open.

obj] Error 1
gmake[1]: *** [orte/CMakeFiles/libopen-rte.dir/all] Error 2
gmake: *** [all] Error 2

___
users mailing list
us...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/users




--
---
Shiqing Fan
High Performance Computing Center Stuttgart (HLRS)
Tel: ++49(0)711-685-87234  Nobelstrasse 19
Fax: ++49(0)711-685-65832  70569 Stuttgart
http://www.hlrs.de/organization/people/shiqing-fan/
email: f...@hlrs.de

Re: [OMPI users] MPI 2 support in sm btl

2011-12-28 Thread Sabela Ramos Garea 
Thanks for your answers.
Is there a similar issue with mx BTL? I mean, using the mx BTL, is it
possible to reach a parent job from a spawned process?

Thanks again! :-)

El 16 de diciembre de 2011 15:08, Jeff Squyres escribió:

> On Dec 14, 2011, at 4:01 PM, Ralph Castain wrote:
>
> >> As far as I know, there is no support for some MPI-2 features in the
> shared memory BTL as dynamic process creation or port connection. Are you
> planning to include this support?
> >
> > It depends on what exactly you mean. Dynamically spawned processes do
> use the shared memory BTL between themselves, but not with the processes in
> their parent job. We plan to support that as well at some future time.
>
> ...when enough people ask for it.  :-)
>
> We've talked about supporting this for a long time, but to be blunt, there
> hasn't been enough demand for it to justify spending the time to figure out
> all the issues and implement it properly.  :-(
>
> Patches would be welcome, though!  :-)
>
> --
> Jeff Squyres
> jsquy...@cisco.com
> For corporate legal information go to:
> http://www.cisco.com/web/about/doing_business/legal/cri/
>
>
> ___
> users mailing list
> us...@open-mpi.org
> http://www.open-mpi.org/mailman/listinfo.cgi/users
>