Intel® MPI Library
Get help with building, analyzing, optimizing, and scaling high-performance computing (HPC) applications.
2222 Discussions

Can't start mpitune after terminate it with ctrl-c

superlee
Beginner
766 Views
mpitune can't be started again after it was terminate with ctrl-c, and show the message below.
How should I do to fix it?

$ mpitune -f ~/machines.LINUX -o ./ --app mpiexec -n 4 ./test
20'Aug'11 06:25:24 INF | Starting. Please wait...

20'Aug'11 06:25:24 | MPITune started at 19 August'11 (Friday) 22:25:24
20'Aug'11 06:25:24 | MPITune has been started by: l
20'Aug'11 06:25:24 | Preparing tuner's components...
20'Aug'11 06:25:24 | Initialization of signals handlers...
20'Aug'11 06:25:24 | Start catching signal with code 15 (SIGTERM) ...
20'Aug'11 06:25:24 | Success.
20'Aug'11 06:25:24 | Start catching signal with code 2 (SIGINT) ...
20'Aug'11 06:25:24 | Success.
20'Aug'11 06:25:24 | Initialization of signals handlers completed.
20'Aug'11 06:25:24 | Existing application conf file with zero lenght has been passed. No copy created.
20'Aug'11 06:25:24 | Obtained following information about Intel MPI Library:
MPI Root : /opt/intel/impi/4.0.1.007
MPI Bin : /opt/intel/impi/4.0.1.007/bin64
MPI Version: 4.0
MPI Build : 20100910
mpiexec : /opt/intel/impi/4.0.1.007/bin64/mpiexec
20'Aug'11 06:25:24 | No batch system has been detected.
20'Aug'11 06:25:24 ERR | Invalid default value ('/home/l/test_intel_mpi/mpd.hosts') of argument ('host-file').
20'Aug'11 06:25:24 WRN | No hosts were obtained for work. Only current host will be used
20'Aug'11 06:25:24 CER | Saving hosts list to file failed.
20'Aug'11 06:25:24 CER | A critical error has occurred!
Details:----------------------------------------------------------------------
Type : exceptions.Exception
Value : Saving hosts list to file failed.
Raised:
{
'('mpitune.py', 342, 'EntryPoint', None)'
'('/tmp/vgusev.xtmpdir.svsmpi011.isv.intel.com.15693/mpi4.32e.svlmpibld05.20100910/dev/src/I_MPI/tuning/BaseModules/Configurator.py', 1124, 'Prepare', None)'
'('/tmp/vgusev.xtmpdir.svsmpi011.isv.intel.com.15693/mpi4.32e.svlmpibld05.20100910/dev/src/I_MPI/tuning/BaseModules/Messenger.py', 88, 'OutputMessage', None)'
}
------------------------------------------------------------------------------
20'Aug'11 06:25:24 | Tuning has been completed for 0h:0m:0s:104ms
20'Aug'11 06:25:24 CER | Error while terminating child processes. Description: 'NoneType' object has no attribute 'DestroyAllChildProcesses'
20'Aug'11 06:25:24 | Report saving...
20'Aug'11 06:25:24 | Complete.
20'Aug'11 06:25:24 WRN | Safe application's termination completed.
20'Aug'11 06:25:26 WRN | Filemanager object executed unlocking by timeout.
20'Aug'11 06:25:26 | Attention! No results have been obtained during current tuning process. It may be caused by:
- Intel MPI Library has optimal settings already
- Tuning process has not been completed at all due to one of follow reasons:
* Time limitations
* Critical errors in process
* Abort of the process by user
* Other
You can find more information in log-file/output.
20'Aug'11 06:25:26 | Tuning has been completed for 0h:0m:1s:416ms
[l@l-pc test_intel_mpi]$ mpitune -f ~/machines.LINUX -o ./ --app \\"mpiexec -n 4 ./test\\"
20'Aug'11 06:25:34 INF | Starting. Please wait...

20'Aug'11 06:25:34 | MPITune started at 19 August'11 (Friday) 22:25:33
20'Aug'11 06:25:34 | MPITune has been started by: l
20'Aug'11 06:25:34 | Preparing tuner's components...
20'Aug'11 06:25:34 | Initialization of signals handlers...
20'Aug'11 06:25:34 | Start catching signal with code 15 (SIGTERM) ...
20'Aug'11 06:25:34 | Success.
20'Aug'11 06:25:34 | Start catching signal with code 2 (SIGINT) ...
20'Aug'11 06:25:34 | Success.
20'Aug'11 06:25:34 | Initialization of signals handlers completed.
20'Aug'11 06:25:34 | Existing application conf file with zero lenght has been passed. No copy created.
20'Aug'11 06:25:34 | Obtained following information about Intel MPI Library:
MPI Root : /opt/intel/impi/4.0.1.007
MPI Bin : /opt/intel/impi/4.0.1.007/bin64
MPI Version: 4.0
MPI Build : 20100910
mpiexec : /opt/intel/impi/4.0.1.007/bin64/mpiexec
20'Aug'11 06:25:34 | No batch system has been detected.
20'Aug'11 06:25:34 ERR | Invalid default value ('/home/l/test_intel_mpi/mpd.hosts') of argument ('host-file').
20'Aug'11 06:25:34 WRN | No hosts were obtained for work. Only current host will be used
20'Aug'11 06:25:34 CER | Saving hosts list to file failed.
20'Aug'11 06:25:34 CER | A critical error has occurred!
Details:----------------------------------------------------------------------
Type : exceptions.Exception
Value : Saving hosts list to file failed.
Raised:
{
'('mpitune.py', 342, 'EntryPoint', None)'
'('/tmp/vgusev.xtmpdir.svsmpi011.isv.intel.com.15693/mpi4.32e.svlmpibld05.20100910/dev/src/I_MPI/tuning/BaseModules/Configurator.py', 1124, 'Prepare', None)'
'('/tmp/vgusev.xtmpdir.svsmpi011.isv.intel.com.15693/mpi4.32e.svlmpibld05.20100910/dev/src/I_MPI/tuning/BaseModules/Messenger.py', 88, 'OutputMessage', None)'
}
------------------------------------------------------------------------------
20'Aug'11 06:25:34 | Tuning has been completed for 0h:0m:0s:107ms
20'Aug'11 06:25:34 CER | Error while terminating child processes. Description: 'NoneType' object has no attribute 'DestroyAllChildProcesses'
20'Aug'11 06:25:34 | Report saving...
20'Aug'11 06:25:34 | Complete.
20'Aug'11 06:25:34 WRN | Safe application's termination completed.
20'Aug'11 06:25:35 WRN | Filemanager object executed unlocking by timeout.
20'Aug'11 06:25:35 | Attention! No results have been obtained during current tuning process. It may be caused by:
- Intel MPI Library has optimal settings already
- Tuning process has not been completed at all due to one of follow reasons:
* Time limitations
* Critical errors in process
* Abort of the process by user
* Other
You can find more information in log-file/output.
20'Aug'11 06:25:35 | Tuning has been completed for 0h:0m:1s:424ms

0 Kudos
1 Reply
Dmitry_K_Intel2
Employee
766 Views
Hi SuperLee,

It seems to me that the issue is related to the command line. Please use '-hf' instead of '-f' option.
Use '-od' instead of '-o' option.
And the most important thing - any command line for an application should be enclosed in " (backslash + double quotes):
--application "mpiexec -n 4 ./test"


Regards!
Dmitry
0 Kudos
Reply