Re: [OMPI users] Checkpoint/Restart in OpenMPI

2009-11-02 Thread Andreea m. (Costea)
Hi again. I have installed BLCR, passed all the tests, and now I tried to install OpenMPI. This is the configuration call: -- ./configure --prefix=/home/andreea/MPIrelated/OpenMPI/openmpi_1.3.3/

Re: [OMPI users] Runtime error while running mpirun

2009-11-02 Thread basant.lakhotiya
Hi Terry, Its not creating mca_paffinity_windows.dll but there is a file mca_paffinity_windowsd.dll. Can you have any idea why it is not creating mca_paffinity_windows's dll. Thanks, Basant -Original Message- From: users-boun...@open-mpi.org [mailto:users-boun...@open-mpi.org] On

Re: [OMPI users] Runtime error while running mpirun

2009-11-02 Thread basant.lakhotiya
Hi Terry, I complied OMPI again, Its creating mca_paffinity_windows and mca_paffinity_windowsd.dll. And also I have checked the dependency of both dll through "VS2005 Depends". Its not showing any missing dependency for these dlls. While running orterun.exe its giving the same error. Thanks,

Re: [OMPI users] Checkpoint/Restart in OpenMPI for Sergio

2009-11-02 Thread Sergio Díaz
Hi, The option --enable-ft-thread doesn't exist. By default is enable. The options --with-blcr and --with-blcr-libdir have to point to your BLCR installation dir. Before compiling openmpi, you have to load the blcr modules. When I installed BLCR, my installation dir was in a different

Re: [OMPI users] Checkpoint/Restart in OpenMPI

2009-11-02 Thread Andreea m. (Costea)
Ok... Maybe someone can clear me :) I have BLCR sources in this directory : /home/andreea/MPIrelated/BLCR/blcr-0.8.2 it's installation is in this directory: /home/andreea/MPIrelated/BLCR/blcr-0.8.2/builddir OpenMPI: sources: /home/andreea/MPIrelated/OpenMPI/SOURCES installation:

Re: [OMPI users] Checkpoint/Restart in OpenMPI

2009-11-02 Thread Sergio Díaz
Maybe the variable MPI_HOME?? Andreea m. (Costea) escribió: Ok... Maybe someone can clear me :) I have BLCR sources in this directory : /home/andreea/MPIrelated/BLCR/blcr-0.8.2 it's installation is in this directory: /home/andreea/MPIrelated/BLCR/blcr-0.8.2/builddir OpenMPI: sources:

Re: [OMPI users] Runtime error while running mpirun

2009-11-02 Thread Shiqing Fan
Hi Basant, The mca_paffinity_windowsd.dll is the debug version of mca_paffinity_windows.dll, but orterun.exe should know which one it can use when you build it. Because you were building Open MPI with libtool support, probably the problem could be that libtool is not loaded correctly.

Re: [OMPI users] Checkpoint/Restart in OpenMPI

2009-11-02 Thread Andreea m. (Costea)
MPI_HOME is set to be /home/andreea/MPIrelated/OpenMPI/openmpi_1.3.3 But as far as I see it needs files that are only in SOURCES... so maybe I did something wrong in the installation of OpenMPI? If I run an already compiled file it works fine, of course. But it does not compile --- On Mon,

Re: [OMPI users] Checkpoint/Restart in OpenMPI

2009-11-02 Thread Sergio Díaz
The mpi.h file is also in the installation dir. It is in the include directory of your installation. I think that it is a path problem... Andreea m. (Costea) escribió: MPI_HOME is set to be /home/andreea/MPIrelated/OpenMPI/openmpi_1.3.3 But as far as I see it needs files that are only in

Re: [OMPI users] Checkpoint/Restart in OpenMPI

2009-11-02 Thread Andreea m. (Costea)
After a reboot and a new setting of the variables I could compile. Is there a way to set the variables for all the sessions? The only one that is set after reboot is PATH, because I modified the file /etc/profile I added new files to

Re: [OMPI users] Runtime error while running mpirun

2009-11-02 Thread basant.lakhotiya
Hi Shiqing, Sorry for misunderstanding. I checked its not creating mca_paffinity_windows.dll and mca_paffinity_windowsd.dll instead that files are libmca_paffinity_windows.dll and libmca_paffinity_windowsd.dll. Only .lib files are there for mca_paffinity_windows and mca_paffinity_windowsd.

[OMPI users] using specific algorithm for collective communication, and knowing the root cpu?

2009-11-02 Thread George Markomanolis
Dear all, I would like to ask about collective communication. With debug mode enabled, I can see many info during the execution which algorithm is used etc. But my question is that I would like to use a specific algorithm (the simplest I suppose). I am profiling some applications and I want

Re: [OMPI users] Runtime error while running mpirun

2009-11-02 Thread Shiqing Fan
Hi Basant, It sounds very strange to me, because all the mca libraries on Windows doesn't have the "lib" prefix, they should all have the name like mca__xxx.dll, so I'm wondering how the libmca_pafinity_windows stuff comes, might be a bug too. Could you provide more information about

Re: [OMPI users] Runtime error while running mpirun

2009-11-02 Thread Shiqing Fan
Hi Basant, Could you please also check in your Open MPI solutions, that do you have the mca_paffinity_windows project? and in the property of this project, go to linker->General, in "Output File" field, you will see the file name of this DLL, is the name "libmca_paffinity_windows.dll" or

Re: [OMPI users] Runtime error while running mpirun

2009-11-02 Thread Shiqing Fan
Hi Basant, Yes, rename the output file name should solve the problem. I'll try to fix it ASAP. Could you send me again the log files, but maybe off list, I can't open it, it got corrupted somehow. And please also send me the output of ompi_info, and

Re: [OMPI users] Runtime error while running mpirun

2009-11-02 Thread Jeff Squyres
On Nov 2, 2009, at 7:43 AM, Shiqing Fan wrote: Because you were building Open MPI with libtool support, probably the problem could be that libtool is not loaded correctly. Could you check that libtool bin directory is in the PATH environment variable? If Open MPI can't find correct libtool

Re: [OMPI users] checkpoint opempi-1.3.3+sge62

2009-11-02 Thread Sergio Díaz
Hi again, I found a C program to test ompi-checkpoint/restart an it works fine. The program was written by Alan Woodland and shared in the following distribution list: debian-bugs-d...@lists.debian.org This program starts a countdown from 10 to 0 and when the countdown is 6, do a checkpoint,

Re: [OMPI users] checkpoint opempi-1.3.3+sge62

2009-11-02 Thread Andreea m. (Costea)
I am having the same problem when I want to checkpoint manually: "HNP with PID Not found!", though I am sure I put the right PID --- On Mon, 11/2/09, Sergio Díaz wrote: From: Sergio Díaz Subject: Re: [OMPI users] checkpoint opempi-1.3.3+sge62 To: "Open

Re: [OMPI users] segmentation fault: Address not mapped

2009-11-02 Thread Jeff Squyres
Many thanks for all this information. Unfortunately, it's not enough to know what's going on. :-( Do you know for sure that the application is correct? E.g., is it possible that a bad buffer is being passed to MPI_Isend? I note that it is fairly odd to fail in MPI_Isend itself because