[FLASH-USERS] compiling issue--error: too few arguments to function [WINDOWS-1252?] $B![ (B5Dopen2? size=

[Chris Daley]

I think the reason FLASH is failing is because you have not promoted 
reals to 8 bytes.

All floating point MPI communication in FLASH describes the data as type:
FLASH_REAL = MPI_DOUBLE_PRECISION, so there is a mismatch in
communication when you are using 4 byte reals.

Add the following flags to FFLAGS_OPT, FFLAGS_DEBUG and FFLAGS_TEST:
-fdefault-real-8 -fdefault-double-8

Chris


Anshu Dubey wrote:
> Ken,
>
> Your code is crashing at the first instance of MPI use after MPI initialization.
> The routine nameValueLL  creates an MPI datatype, and then has several
> MPI_bcast calls. Could you try a simpler mpi program to verify that it is not
> your installation that is the problem ?
>
> Anshu
>
> On Tue, Feb 10, 2009 at 2:06 PM, Ken Chen <ken at asiaa.sinica.edu.tw> wrote:
>   
>> Hi Nathan and all:
>> I tried the both sedov and sod problems. They worked fine for single cpu, but didn't
>> work multi cpus. Is there a problem in open-mp or mpi stuff? I can compile the
>> problems with these setups(mpi, open-mp) in Makefile.h. May something go wrong ?
>> Thanks for any comment.
>>
>> cheers;
>>
>> Ken
>>
>>
>> ---------- Original Message -----------
>> From: Nathan Hearn <nhearn at ucar.edu>
>> To: Ken Chen <ken at asiaa.sinica.edu.tw>
>> Cc: flash-users at flash.uchicago.edu
>> Sent: Tue, 10 Feb 2009 12:22:49 -0700
>> Subject: Re: [FLASH-USERS] compiling issue--error: too few arguments to function
>> [WINDOWS-1252?] $B![ (B5Dopen2? size=
>>
>>     
>>> Hi Ken,
>>>
>>>     I'm sorry about that ... I read through your message too quickly.
>>> It is difficult to ascertain the problem from the error messages, as a
>>> segmentation fault can occur in many places.  Which setup are you
>>> trying to run?  (And have you tried any of the other setups, such as
>>> Sod or Sedov)?
>>>
>>> - Nathan
>>>
>>> P.S., Please be sure to reply to the flash-users list so that others
>>> can offer advice.
>>>
>>> 2009/2/10 Ken Chen <ken at asiaa.sinica.edu.tw>:
>>>       
>>>> Hi Nathan:
>>>> Thanks for prompt reply. Maybe you misunderstood my question, I am curious that I
>>>>         
>> can
>>     
>>>> run single cpu for flash on our machine (32 cpus work station). However, if I want
>>>>         
>> to run
>>     
>>>> more than 1 cpu such a  2,4,even 32. However I got the error message from below. I
>>>> don't understand where the problem came from. Could you take a look the error
>>>>         
>> message
>>     
>>>> and give me some comments? Thanks a lot~~
>>>>
>>>> cheers;
>>>>
>>>> Ken
>>>>
>>>> *******************************************
>>>> [ken at c object]$ mpirun -np 2 flash3
>>>>  RuntimeParameters_read:  ignoring unknown parameter "iProcs"...
>>>>  RuntimeParameters_read:  ignoring unknown parameter "jProcs"...
>>>>  RuntimeParameters_read:  ignoring unknown parameter "kProcs"...
>>>> [c:31677] *** Process received signal ***
>>>> [c:31677] Signal: Segmentation fault (11)
>>>> [c:31677] Signal code: Address not mapped (1)
>>>> [c:31677] Failing at address: 0x3200000011
>>>> [c:31677] [ 0] /lib64/libpthread.so.0 [0x3317c0f0f0]
>>>> [c:31677] [ 1] /usr/lib64/openmpi/1.2.4-gcc/libopen-pal.so.0(_int_malloc+0x19c)
>>>> [0x327882d2dc]
>>>> [c:31677] [ 2] /usr/lib64/openmpi/1.2.4-gcc/libopen-pal.so.0(malloc+0x8f)
>>>>         
>> [0x327882ebcf]
>>     
>>>> [c:31677] [ 3] flash3(__namevaluell_data_MOD_namevaluell_addreal+0x92) [0x4fbb92]
>>>> [c:31677] [ 4] flash3(namevaluell_bcast_+0x6b1) [0x4fb481]
>>>> [c:31677] [ 5] flash3(runtimeparameters_init_+0x83) [0x44e573]
>>>> [c:31677] [ 6] flash3(driver_initflash_+0x4a) [0x42f9ea]
>>>> [c:31677] [ 7] flash3(MAIN__+0x1a) [0x43282a]
>>>> [c:31677] [ 8] flash3(main+0x2c) [0x5d97ec]
>>>> [c:31677] [ 9] /lib64/libc.so.6(__libc_start_main+0xe6) [0x331701e576]
>>>> [c:31677] [10] flash3 [0x42b529]
>>>> [c:31677] *** End of error message ***
>>>> mpirun noticed that job rank 0 with PID 31676 on node c.spa.umn.edu exited on
>>>>         
>> signal 15
>>     
>>>> (Terminated).
>>>> 1 additional process aborted (not shown)
>>>> ******************************************************************
>>>>         
>> ------- End of Original Message -------
>>
>>
>>