[OMPI devel] openmpi-1.8.2rc2 and f08 interface built with PGI-14.7 causes link error

Hi folks, I tried to build openmpi-1.8.2rc2 with PGI-14.7 and execute a sample program. Then, it causes linking error: [mishima@manage work]$ cat test.f program hello_world use mpi_f08 implicit none type(MPI_Comm) :: comm integer :: myid, npes, ierror

[OMPI devel] MPI_T SEGV on DSO

Hi, I encountered the same SEGV reported on the users list when running varList program. http://www.open-mpi.org/community/lists/users/2014/07/24792.php mpiexec -n 1 ./varList: ... snip ... event

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

Hi, On 05/10/2014 02:46 PM, Bert Wesarg wrote: Hi, Btw, I'm pretty confident, that this Open SHMEM implementation does not recognize global or static variables in shared libraries as symmetric objects. It is probably wise to note this somewhere to the users. I've never got an reply to this

Re: [OMPI devel] MPI_T SEGV on DSO

The problem is the code in question does not check the return code of MPI_T_cvar_handle_alloc . We are returning an error and they still try to use the handle (which is stale). Uncomment this section of the code: //if (MPI_T_ERR_INVALID_INDEX == err)// { NOTE TZI: This variable

[OMPI devel] Github exploration status

Folks -- Here's the slides I went through (including the URLs) showing the current status of github testing. -- Jeff Squyres jsquy...@cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/ ompi-github-testing-status.pptx Description:

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

> > On 05/10/2014 02:46 PM, Bert Wesarg wrote: >> Hi, >> >> Btw, I'm pretty confident, that this Open SHMEM implementation does not >> recognize global or static variables in shared libraries as symmetric >> objects. It is probably wise to note this somewhere to the users. > > I've never got an

Re: [OMPI devel] MPI_T SEGV on DSO

FYI: We have pinged the upstream/LLNL authors of varlist about this issue. On Jul 29, 2014, at 11:38 AM, Nathan Hjelm wrote: > > The problem is the code in question does not check the return code of > MPI_T_cvar_handle_alloc . We are returning an error and they still try > to

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

On 07/29/2014 06:23 PM, Shamis, Pavel wrote: On 05/10/2014 02:46 PM, Bert Wesarg wrote: Hi, Btw, I'm pretty confident, that this Open SHMEM implementation does not recognize global or static variables in shared libraries as symmetric objects. It is probably wise to note this somewhere to the

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

Are you claiming that in the following test, the static variable "val" will not be seen as a symmetric object? #include "shmem.h" int main( int argc, char **argv){ long my_pe, npes, master; start_pes(0); my_pe = shmem_my_pe(); npes = shmem_n_pes(); master = npes - 1; /*

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

On 07/29/2014 06:57 PM, Joshua Ladd wrote: Are you claiming that in the following test, the static variable "val" will not be seen as a symmetric object? No. #include "shmem.h" int main( int argc, char **argv){ long my_pe, npes, master; start_pes(0); my_pe = shmem_my_pe();

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

Hi Josh, No, I think what Bert is talking about is having something like foobar.c == #include “shmem.h” long my_dso_val = 0; int foobar(void) { printf(“Hello from a shared library\n”); } == cc –fPIC –c foobar.c cc –shared –o libfoobar.so foobar.o cc –o

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

Gotcha, missed the shared lib part... On Tue, Jul 29, 2014 at 1:30 PM, Pritchard Jr., Howard wrote: > Hi Josh, > > > > No, I think what Bert is talking about is having something like > > > > foobar.c > == > > #include “shmem.h” > > > > long my_dso_val

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

then in your main example below do a shmem_long_fadd on my_dso_val. It won’t work unless you’ve put smarts in the shmem library to go through the segments of loaded shared libraries and register them with the same mechanism used for the data segment of the executable. In this case the "smart"

Re: [OMPI devel] opal_config_bottom.h problem with trunk

On Jul 29, 2014, at 1:53 PM, Pritchard Jr., Howard wrote: > If I replace the opal_config.h.in with an older version the build works > again. It seems the opal_config.h wasn't getting generated in my build area, > even though the config.log indicates claims it was. Hmm; that's

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

Btw, I'm pretty confident, that this Open SHMEM implementation does not recognize global or static variables in shared libraries as symmetric objects. It is probably wise to note this somewhere to the users. >>> >>> I've never got an reply to this query. Any comments on it?

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

Pasha, Is v1.1 posted somewhere? I don't see it up on the LBNL site. Josh On Tue, Jul 29, 2014 at 2:05 PM, Shamis, Pavel wrote: > > Btw, I'm pretty confident, that this Open SHMEM implementation does > not > recognize global or static variables in shared

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

Hi Pasha, It can be done. Cray Fortran 2008 compiler had to do this to be fully compliant with the Fortran 2008 co-array feature. Howard -Original Message- From: devel [mailto:devel-boun...@open-mpi.org] On Behalf Of Shamis, Pavel Sent: Tuesday, July 29, 2014 12:02 PM To: Open MPI

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

Is v1.1 posted somewhere? I don't see it up on the LBNL site. www.openshmem.org "Get documentation" - > "Specification" (For some reason I can not get direct link) Pasha Josh On Tue, Jul 29, 2014 at 2:05 PM, Shamis, Pavel

Re: [OMPI devel] SHMEM symmetric objects in shared libraries

On 07/29/2014 08:23 PM, Shamis, Pavel wrote: Is v1.1 posted somewhere? I don't see it up on the LBNL site. www.openshmem.org "Get documentation" - > "Specification" (For some reason I can not get direct link)

[OMPI devel] OMPI_XXX defines in opal_config.h.in question

Hi Folks, So I'm trying to get my pmix project back in order after making the big mistake of pulling in the BTL move stuff. I get bizillions of warning messages from the compiler now about undefined macros. A number of the culprits are in opal_config.h.in, and begin with OMPI, like

Re: [OMPI devel] OMPI_XXX defines in opal_config.h.in question

Interesting - I'm not seeing those myself On Jul 29, 2014, at 1:01 PM, George Bosilca wrote: > Good catch. As you already have the warnings, please go ahead and fix them. > > Thanks, > George. > > > > On Tue, Jul 29, 2014 at 3:58 PM, Pritchard Jr., Howard

Re: [OMPI devel] OMPI_XXX defines in opal_config.h.in question

Hi Ralph, Oh sorry for the confusion. forgot I'd been trying to fix a different problem that caused this one. Sorry. Howard From: devel [mailto:devel-boun...@open-mpi.org] On Behalf Of Ralph Castain Sent: Tuesday, July 29, 2014 2:15 PM To: Open MPI Developers Subject: Re: [OMPI devel]

[OMPI devel] RFC: add atomic compare-and-swap that returns old value

What: Add new versions of opal_atomic_cmpset_* that return the old value not whether they succeeded. Why: I plan on adding support for network atomics to the BTL interface. The compare-and-swap function will return the old value from the target memory location. In order to implement a similar

Re: [OMPI devel] RFC: add atomic compare-and-swap that returns old value

On Tue, Jul 29, 2014 at 2:10 PM, Nathan Hjelm wrote: > Is there a reason why the > current implementations of opal atomics (add, cmpset) do not return the > old value? > Because some CPUs don't implement such an atomic instruction? On any CPU one *can* certainly synthesize the

Re: [OMPI devel] btl_openib_receive_queues mca param not always taken into account

Hi, Nadia I CMRed your patch to 1.8.2 and applied the fix on the trunk in: https://svn.open-mpi.org/trac/ompi/changeset/32346 Thanks for reporting! Josh On Fri, Jul 11, 2014 at 6:04 AM, Nadia Derbey wrote: > Hi, > > I noticed that specifying the receive_queues

Re: [OMPI devel] [OMPI svn] svn:open-mpi r32346 - trunk/opal/mca/btl/openib

Josh, you can not free from a memory location that has been registered with the MCA variable system. It will likely SEGV when the component is unloaded. You should call mca_base_var_get_value to get the source of the value. The following should do it: vari = mca_base_var_find ("opal", "btl",

Re: [OMPI devel] [OMPI svn] svn:open-mpi r32346 - trunk/opal/mca/btl/openib

Thanks, Nathan This is a user's patch. So, if I understand correctly, the best thing would be to: vari = mca_base_var_find ("opal", "btl", "openib", "receive_queues"); mca_base_var_get_value (vari, NULL, , NULL); mca_btl_openib_component. receive_queues_source = source Instead of doing this

Re: [OMPI devel] [OMPI svn] svn:open-mpi r32346 - trunk/opal/mca/btl/openib

Yeah. Though it would be best to just check the source when you need to see if it should come from the ini. Then if we need to set the value from the ini either use mca_base_var_set_value or be sure to strdup when changing the receive_queues value. The thing to remember is the use may do this:

Re: [OMPI devel] [OMPI svn] svn:open-mpi r32346 - trunk/opal/mca/btl/openib

On Tue, Jul 29, 2014 at 04:12:18PM -0600, Nathan Hjelm wrote: > > Yeah. Though it would be best to just check the source when you need to > see if it should come from the ini. Then if we need to set the value > from the ini either use mca_base_var_set_value or be sure to strdup when > changing

Re: [OMPI devel] MPI_T SEGV on DSO

Nathan, Thanks for your response. Yes. My previous mail was the result of uncommented code. Now I also pulled latest varList source code which uncommented the section you mentioned, but the result was same. If MPI_T_cvar_get_info should return MPI_T_ERR_INVALID_INDEX for variables for unloaded

Re: [OMPI devel] openmpi-1.8.2rc2 and f08 interface built with PGI-14.7 causes link error

Tetsuya -- I am unable to test with the PGI compiler -- I don't have a license. I was hoping that LANL would be able to test today, but I don't think they got to it. Can you send more details? E.g., can you send the all the stuff listed on http://www.open-mpi.org/community/help/ for 1.8 and

Re: [OMPI devel] openmpi-1.8.2rc2 and f08 interface built with PGI-14.7 causes link error

I have license for PGI and installations of 14.1 and 14.4 I will see what I can do today in terms of testing. -Paul On Tue, Jul 29, 2014 at 4:23 PM, Jeff Squyres (jsquyres) wrote: > Tetsuya -- > > I am unable to test with the PGI compiler -- I don't have a license. I >

Re: [OMPI devel] openmpi-1.8.2rc2 and f08 interface built with PGI-14.7 causes link error

PGI 14.7 is VERY new -- I just received the announcement on Sunday. Larry Baker US Geological Survey 650-329-5608 ba...@usgs.gov On 29 Jul 2014, at 4:25 PM, Paul Hargrove wrote: > I have license for PGI and installations of 14.1 and 14.4 > I will see what I can do today in terms of testing. >

Re: [OMPI devel] openmpi-1.8.2rc2 and f08 interface built with PGI-14.7 causes link error

If I read the original email correctly then 1.8.2rc2 also failed on pgi-13.10. At the moment I am just hoping to reproduce at all. -Paul On Tue, Jul 29, 2014 at 4:46 PM, Larry Baker wrote: > PGI 14.7 is VERY new -- I just received the announcement on Sunday. > > Larry Baker >

Re: [OMPI devel] openmpi-1.8.2rc2 and f08 interface built with PGI-14.7 causes link error

On Tue, Jul 29, 2014 at 4:23 PM, Jeff Squyres (jsquyres) wrote: > Tetsuya -- > > I am unable to test with the PGI compiler -- I don't have a license. I > was hoping that LANL would be able to test today, but I don't think they > got to it. > > Can you send more details? > >

Re: [OMPI devel] openmpi-1.8.2rc2 and f08 interface built with PGI-14.7 causes link error

On Tue, Jul 29, 2014 at 6:33 PM, Paul Hargrove wrote: > I am trying again with an explicit --enable-mpi-fortran=usempi at > configure time to see what happens. > Of course that should have said --enable-mpi-fortran=usempif08 -- Paul H. Hargrove

Re: [OMPI devel] [OMPI svn] svn:open-mpi r32346 - trunk/opal/mca/btl/openib

Nathan, can you take a look at https://svn.open-mpi.org/trac/ompi/changeset/32350 when you get a chance. Thanks, Josh On Tue, Jul 29, 2014 at 6:14 PM, Nathan Hjelm wrote: > On Tue, Jul 29, 2014 at 04:12:18PM -0600, Nathan Hjelm wrote: > > > > Yeah. Though it would be best to

[hwloc-devel] Create success (hwloc git dev-169-g65abf30)

Creating nightly hwloc snapshot git tarball was a success. Snapshot: hwloc dev-169-g65abf30 Start time: Tue Jul 29 21:01:02 EDT 2014 End time: Tue Jul 29 21:02:29 EDT 2014 Your friendly daemon, Cyrador