Re: [OMPI users] Help: MPIBLAST Crash

2008-09-02 Thread Brock Palen 

Check out the first error:

ADIOI_GEN_DELETE (line 22): **io No such file or directoryFatal Error:
mpiblast.cppError reading database mbf file


Check to make sure all your files are in place.
Also many times you need to set you .ncbirc


Brock Palen
www.umich.edu/~brockp
Center for Advanced Computing
bro...@umich.edu
(734)936-1985



On Aug 30, 2008, at 1:27 AM, Lee Amy wrote:

Hello,

I use Open MPI 1.2.5 to compile MPIBLAST 1.5, it runs well. When I  
run job by this following command:


mpirun -np 21 -mca pls_rsh_agent rsh mpiblast -p blastn -d test.fa - 
i /usr/local/aaa.fa -o output


It shows these error messages:

ADIOI_GEN_DELETE (line 22): **io No such file or directoryFatal Error:
mpiblast.cppError reading database mbf file
[gnode1:26957] MPI_ABORT invoked on rank 0 in communicator  
MPI_COMM_WORLD with errorcode -1

40.010139Bailing out with signal 15
[gnode2:11975] MPI_ABORT invoked on rank 4 in communicator  
MPI_COMM_WORLD with errorcode 0

200.01599Bailing out with signal 15
[gnode1:26961] MPI_ABORT invoked on rank 20 in communicator  
MPI_COMM_WORLD with errorcode 0

160.00986695Bailing out with signal 15
[gnode5:10829] MPI_ABORT invoked on rank 16 in communicator  
MPI_COMM_WORLD with errorcode 0

50.0122449Bailing out with signal 15
[gnode2:11976] MPI_ABORT invoked on rank 5 in communicator  
MPI_COMM_WORLD with errorcode 0

10.0102971Bailing out with signal 15
[gnode1:26958] MPI_ABORT invoked on rank 1 in communicator  
MPI_COMM_WORLD with errorcode 0

170.012876Bailing out with signal 15
[gnode5:10830] MPI_ABORT invoked on rank 17 in communicator  
MPI_COMM_WORLD with errorcode 0

60.014256Bailing out with signal 15
[gnode2:11977] MPI_ABORT invoked on rank 6 in communicator  
MPI_COMM_WORLD with errorcode 0

20.0124009Bailing out with signal 15
[gnode1:26959] MPI_ABORT invoked on rank 2 in communicator  
MPI_COMM_WORLD with errorcode 0

180.0144899Bailing out with signal 15
[gnode5:10831] MPI_ABORT invoked on rank 18 in communicator  
MPI_COMM_WORLD with errorcode 0

70.016083Bailing out with signal 15
[gnode2:11978] MPI_ABORT invoked on rank 7 in communicator  
MPI_COMM_WORLD with errorcode 0

30.011174Bailing out with signal 15
[gnode1:26960] MPI_ABORT invoked on rank 3 in communicator  
MPI_COMM_WORLD with errorcode 0

190.0175061Bailing out with signal 15
[gnode5:10832] MPI_ABORT invoked on rank 19 in communicator  
MPI_COMM_WORLD with errorcode 0

120.00988579Bailing out with signal 15
[gnode4:13462] MPI_ABORT invoked on rank 12 in communicator  
MPI_COMM_WORLD with errorcode 0

130.012902Bailing out with signal 15
[gnode4:13463] MPI_ABORT invoked on rank 13 in communicator  
MPI_COMM_WORLD with errorcode 0

140.0149741Bailing out with signal 15
[gnode4:13464] MPI_ABORT invoked on rank 14 in communicator  
MPI_COMM_WORLD with errorcode 0

150.016958Bailing out with signal 15
[gnode4:13465] MPI_ABORT invoked on rank 15 in communicator  
MPI_COMM_WORLD with errorcode 0

80.0100341Bailing out with signal 15
[gnode3:13459] MPI_ABORT invoked on rank 8 in communicator  
MPI_COMM_WORLD with errorcode 0

90.0302022Bailing out with signal 15
[gnode3:13460] MPI_ABORT invoked on rank 9 in communicator  
MPI_COMM_WORLD with errorcode 0

100.032104Bailing out with signal 15
[gnode3:13461] MPI_ABORT invoked on rank 10 in communicator  
MPI_COMM_WORLD with errorcode 0

110.0338449Bailing out with signal 15
[gnode3:13462] MPI_ABORT invoked on rank 11 in communicator  
MPI_COMM_WORLD with errorcode 0


And the job stop. However, I don't know why this problem happened.

Could you tell me how to fix that?

Huge thanks to you.

Best Regards,

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

Re: [OMPI users] Problems with compilig of OpenMPI 1.2.7

2008-09-02 Thread Rolf Vandevaart 

On 08/29/08 19:27, Jeff Squyres wrote:

On Aug 29, 2008, at 10:48 AM, Rolf Vandevaart wrote:

In the file mpicxx.cc there is a declaration near the bottom that 
looks like this.


const int LOCK_SHARED = MPI_LOCK_SHARED;

The preprocessor is going through that file and replacing LOCK_SHARED 
with 0x01.  Then when it tries to compile it you are trying to compile 
a line that looks like this.


const int 0x01 = 2;

That is why you see the error.


Hmm.  This hasn't changed in mpicxx.cc for a long time.  What made it 
get activated now?




I think I touched upon this in my earlier post.  There was a change in 
/usr/include/sys/synch.h Solaris header file. And one of the changes was 
adding the following line.


#define LOCK_SHARED 0x01/* same as USYNC_PROCESS */

Therefore, we are seeing it on later versions of Solaris.

Rolf

--

=
rolf.vandeva...@sun.com
781-442-3043
=

[OMPI users] CPU burning in Wait state

2008-09-02 Thread Vincent Rotival 

Dear all

I think I already read some comments on this issue, but I'd like to know 
of latest versions of OpenMPI have managed to solve it. I am now running 
1.2.5


If I run a MPI program with synchronization routines (e.g. MPI_barrier, 
MPI_bcast...), all threads waiting for data are still burning CPU. On 
the other hand when using non-blocking receives all threads waiting for 
data are not consuming any CPU.


Would there be a possibility to use MPI_Bcast without  burning CPU power ?

Thanks in advance

Vincent

--
**
* CEA/DAM Ile-De-France  * 15 Boulevard BLANQUI  *
* DAM/DIF/DPTA/PN/LSNT   * 75013 Paris   *
* 91680 Bruyeres-Le-Chatel BP12  * FRANCE*
* FRANCE *   *
* office : (+33)169264972* home : (+33)678625323 *
* fax: (+33)169267063*   *
**