[MOAB-dev] build question

[Rajeev Jain]

Many be it's the classes with same names in CGM/MOAB/Lasso/MK causing problems in shared builds.


CGM's GeometryQueryTool.cpp:1356 create an object CpuTimer timer; and the debugger goes to a CpuTimer class in WriteHDF5.cpp in MOAB. 

At this point since_birth() function is called in CpuTimer class and assygen fails with an error: Attempting to use an MPI routine before initializing MPICH



#0  CpuTimer (this=0x7fffffffc7f0) at WriteHDF5.cpp:66
#1  0x00007ffff5954e9d in GeometryQueryTool::make_Body (this=0x7d16e0, bodysm_ptr=0x7d9110) at GeometryQueryTool.cpp:1356
#2  0x00007ffff59060dc in GeometryModifyTool::cylinder (this=0x7d1cc0, hi=40, r1=1.6000000000000001, r2=1.6000000000000001, r3=1.6000000000000001) at GeometryModifyTool.cpp:637
#3  0x00007ffff5e4160e in igeom_createcylinder_ (instance=0x7ffff605ce20, height=40, major_rad=1.6000000000000001, minor_rad=1.6000000000000001, geom_entity=0x7fffffffcc98, err=0x7fffffffde4c) at iGeom_CGMA.cc:5695
#4  0x000000000041a347 in CNrgen::CreatePinCell (this=0x7fffffffd6d0, i=1, dX=0, dY=-0, dZ=0) at io.cpp:2480
#5  0x00000000004160e0 in CNrgen::Create_CartAssm (this=0x7fffffffd6d0, szInputString=...) at io.cpp:1927
#6  0x000000000040c037 in CNrgen::ReadAndCreate (this=0x7fffffffd6d0) at io.cpp:674
#7  0x0000000000404c24 in main (argc=1, argv=0x7fffffffdfe8) at assygen.cpp:38

 
Rajeev



>________________________________
>From: Tim Tautges <tautges at mcs.anl.gov>
>To: Iulian Grindeanu <iulian at mcs.anl.gov>
>Cc: moab-dev at mcs.anl.gov
>Sent: Tuesday, May 24, 2011 11:14 AM
>Subject: Re: [MOAB-dev] build question
>
>
>
>On 05/24/2011 11:08 AM, Iulian Grindeanu wrote:
>> It is now a static build; I will set the PATH and LD_LIBRARY_PATH to point to the same mpich, and I will let you know
>> On my laptop (32 bit) works fine too, I am using default mpich from synaptic
>>
>
>On one of your earlier messages you were specifying --with-mpi=/home/fathom/..., I think for MOAB.
>
>> I am wondering why the error appears only for cubit-based cgm
>>
>
>For cubit-based cgm, you'll have at least some shared libs, since that's the only way cubit libs are distributed.  Maybe 
>in your non-cubit cgm you have true static libs.
>
>- tim
>
>> ----- Original Message -----
>>> If this is a shared build, then the test file is actually a shell
>>> script, the actual executable is in the .libs
>>> directory under that.
>>>
>>> - tim
>>>
>>> On 05/24/2011 10:50 AM, Jason Kraftcheck wrote:
>>>> On 05/24/2011 10:44 AM, Iulian Grindeanu wrote:
>>>>>
>>>>>
>>>>> FAIL: partest (exit: 1)
>>>>> =======================
>>>>>
>>>>> Attempting to use an MPI routine before initializing MPICH
>>>>> ...
>>>>>
>>>>>
>>>>> The debugger will not start, and I don't know which one is called
>>>>> before MPI_Init
>>>>>
>>>>
>>>> The debugger will not start? Assuming you figure out what is going
>>>> on with
>>>> that, try setting break points in MPI_Comm_rank and MPI_Comm_size.
>>>> Most
>>>> codes using MPI call at least one of those before anything else.
>>>>
>>>> - jason
>>>>
>>>>
>>>
>>> --
>>> ================================================================
>>> "You will keep in perfect peace him whose mind is
>>> steadfast, because he trusts in you." Isaiah 26:3
>>>
>>> Tim Tautges Argonne National Laboratory
>>> (tautges at mcs.anl.gov) (telecommuting from UW-Madison)
>>> phone: (608) 263-8485 1500 Engineering Dr.
>>> fax: (608) 263-4499 Madison, WI 53706
>>
>
>-- 
>================================================================
>"You will keep in perfect peace him whose mind is
>   steadfast, because he trusts in you."               Isaiah 26:3
>
>              Tim Tautges            Argonne National Laboratory
>          (tautges at mcs.anl.gov)      (telecommuting from UW-Madison)
>          phone: (608) 263-8485      1500 Engineering Dr.
>            fax: (608) 263-4499      Madison, WI 53706
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/moab-dev/attachments/20110524/1f80c421/attachment.htm>