Somehow Brian's email didn't come through right in my mailing list
digest and his email is hidden. Brian, would you mind re-sending me
your "right-on" email?
Thank you!
Aaron Thompson
Vanderbilt University
aaron.p.thompson_at_[hidden]
On Jul 10, 2007, at 11:37 AM, lam-request_at_[hidden] wrote:
> Send lam mailing list submissions to
> lam_at_[hidden]
>
> To subscribe or unsubscribe via the World Wide Web, visit
> http://www.lam-mpi.org/mailman/listinfo.cgi/lam
> or, via email, send a message with subject or body 'help' to
> lam-request_at_[hidden]
>
> You can reach the person managing the list at
> lam-owner_at_[hidden]
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of lam digest..."
> Today's Topics:
>
> 1. Re: Undefined MPI symbols linker error with XLF (Brian Barrett)
> 2. Re: Running MPI program in a cluster (Brian Barrett)
> 3. Re: Undefined MPI symbols linker error with XLF (Joe Griffin)
> 4. Undefined reference to MPI_INIT (Viding Markus)
>
>
> Brian
>
> On Jul 9, 2007, at 3:50 PM, Aaron Thompson wrote:
>
>> Joe,
>> I'm not too familiar with Fortran, I'm only compiling a program
>> from source (DLPOLY). Could you tell me what compiler flags I
>> could try or where I could look? I didn't see anything too likely
>> in the xlf manual.
>>
>> Thank you,
>>
>> Aaron Thompson
>> Vanderbilt University
>> aaron.p.thompson_at_[hidden]
>>
>>
>>
>> On Jul 9, 2007, at 3:07 PM, lam-request_at_[hidden] wrote:
>>
>>> Aaron,
>>>
>>> Instead of trying to correct it with LDFLAGS at load time, Can you
>>> use compiler
>>> options for your code to match the MPI entry points?
>>>
>>> Joe
>>>
>>>
>>> From: lam-bounces_at_[hidden] [mailto:lam-bounces_at_[hidden]] On
>>> Behalf Of Aaron Thompson
>>> Sent: Monday, July 09, 2007 11:43 AM
>>> To: General LAM/MPI mailing list
>>> Subject: Re: LAM: Undefined MPI symbols linker error with XLF
>>>
>>> Underscores could be the issue. My program is calling MPI
>>> functions in all caps with no underscoring, e.g. MPI_SEND . I
>>> tried adding -fno_underscoring to my LDFLAGS, but that didn't
>>> help. Are there any other flags can I try using?
>>>
>>> Thank you,
>>>
>>> Aaron Thompson
>>> Vanderbilt University
>>> aaron.p.thompson_at_[hidden]
>>>
>>>
>>>> Could your problem be Case or leading or trailing underscores?
>>>>
>>>> Some code has:
>>>>
>>>> MPI_Send
>>>> mpi_send
>>>> mpi_send_
>>>> _mpi_send_
>>>>
>>>> The code and MPI libs need to have the same Case and underscore
>>>> policy.
>>>>
>>>> Regards,
>>>> joe Griffin
>>
>> _______________________________________________
>> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
> --
> Brian Barrett
> LAM/MPI Developer
> Make today a LAM/MPI day!
>
>
>
>
>
>
>
>> I use the lamboot comand to set up the LAM enviroment in a single
>> computer,
>> i.e., "lamboot -v". I am running a simple master slave program and
>> it runs
>> alright when I use the "mpirun -np 2 main" command. I then try to
>> set up a
>> simple cluster with only 2 computers making the hostfile as you
>> guys told in
>> the user guide pdf. It seems to boot fine. The strange thing is
>> that when I
>> run my program in a single computer with 2 processes it runs way
>> faster then
>> when I run it with process 0 in machine 0 and process 1 in machine
>> 1. Any
>> idea why this happens? Any help would be appreciated.
>
> This isn't totally unexpected, especially if one of the processes
> doesn't require much CPU time or you are on an SMP machine. When
> processes are on the same node, LAM/MPI will use shared memory for
> communication. If your cluster only has ethernet, bandwidth is
> halved and latency can easily be two or three orders of magnitude
> higher when compared to shared memory. If your application uses lots
> of short messages, the disparity between shared memory and TCP
> performance can be really easy to see.
>
>
> Hope this helps,
>
> Brian
>
> --
> Brian Barrett
> LAM/MPI Developer
> Make today a LAM/MPI day!
>
>
>
>
>
>
> From: "Joe Griffin" <joe.griffin_at_[hidden]>
> Date: July 10, 2007 1:21:26 AM EDT
> To: "General LAM/MPI mailing list" <lam_at_[hidden]>
> Subject: Re: LAM: Undefined MPI symbols linker error with XLF
> Reply-To: General LAM/MPI mailing list <lam_at_[hidden]>
>
>
> Aaron,
>
> Brian's advice about using XLF is right on.
>
> Joe
>
> ________________________________
>
> From: lam-bounces_at_[hidden] on behalf of Aaron Thompson
> Sent: Mon 7/9/2007 2:50 PM
> To: General LAM/MPI mailing list
> Subject: Re: LAM: Undefined MPI symbols linker error with XLF
>
>
> Joe,
> I'm not too familiar with Fortran, I'm only compiling a program
> from source (DLPOLY). Could you tell me what compiler flags I
> could try or where I could look? I didn't see anything too likely
> in the xlf manual.
>
> Thank you,
>
> Aaron Thompson
> Vanderbilt University
> aaron.p.thompson_at_[hidden]
>
>
>
> On Jul 9, 2007, at 3:07 PM, lam-request_at_[hidden] wrote:
>
>
>
> Aaron,
>
> Instead of trying to correct it with LDFLAGS at load time, Can you
> use compiler
> options for your code to match the MPI entry points?
>
> Joe
>
>
>
> ________________________________
>
> From: lam-bounces_at_[hidden] [mailto:lam-bounces_at_[hidden]] On
> Behalf Of Aaron Thompson
> Sent: Monday, July 09, 2007 11:43 AM
> To: General LAM/MPI mailing list
> Subject: Re: LAM: Undefined MPI symbols linker error with XLF
>
> Underscores could be the issue. My program is calling MPI
> functions in all caps with no underscoring, e.g. MPI_SEND . I
> tried adding -fno_underscoring to my LDFLAGS, but that didn't
> help. Are there any other flags can I try using?
>
> Thank you,
>
> Aaron Thompson
> Vanderbilt University
> aaron.p.thompson_at_[hidden]
>
>
>
> Could your problem be Case or leading or trailing underscores?
>
> Some code has:
>
> MPI_Send
> mpi_send
> mpi_send_
> _mpi_send_
>
> The code and MPI libs need to have the same Case and underscore
> policy.
>
> Regards,
> joe Griffin
>
>
> <winmail.dat>
> From: Viding Markus <markus.viding_at_[hidden]>
> Date: July 10, 2007 11:37:21 AM EDT
> To: "'lam_at_[hidden]'" <lam_at_[hidden]>
> Cc: "'support_at_[hidden]'" <support_at_[hidden]>
> Subject: LAM: Undefined reference to MPI_INIT
> Reply-To: General LAM/MPI mailing list <lam_at_[hidden]>
>
>
> Hi,
>
>
> I am running LAM/MPI 7.1.3 with a Linux 2.4.21-27 kernel. The
> compiler I am using is an Absoft Fortran 95 10.0. Trying to run a
> simple Hello world:
>
>
> PROGRAM HELLO
>
> INCLUDE 'mpif.h'
>
>
> INTEGER :: RANK, SIZE, IERR
>
>
> CALL MPI_INIT(IERR)
>
> CALL MPI_COMM_RANK(MPI_COMM_WORLD,RANK,IERR)
>
> CALL MPI_COMM_SIZE(MPI_COMM_WORLD,SIZE,IERR)
>
>
> PRINT *, 'Hello! I am ', RANK, 'of', SIZE
>
>
> CALL MPI_FINALIZE(IERR)
>
> STOP
>
> END PROGRAM
>
>
> ------------------------------------
>
>
> When running mpif77 -showme, I get:
>
>
> F95 -I/opt/lam/lam-7_1_3/include -L/opt/lam/lam-7_1_3/lib -llammpio
> -llamf77mpi -lmpi -llam -laio -laio -lutil -lpthread -pthread -ldl
>
>
>
> When trying to compile the application this is what happens.
>
>
> [mavi4915_at_tor10 Multi]$ mpif77 Hello.f90
>
> /tmp/a9IZW7/Hello.o(.text+0x2f): In function `MAIN__':
>
> : undefined reference to `MPI_INIT'
>
> /tmp/a9IZW7/Hello.o(.text+0x4d): In function `MAIN__':
>
> : undefined reference to `MPI_COMM_RANK'
>
> /tmp/a9IZW7/Hello.o(.text+0x6b): In function `MAIN__':
>
> : undefined reference to `MPI_COMM_SIZE'
>
> /tmp/a9IZW7/Hello.o(.text+0x1bf): In function `MAIN__':
>
> : undefined reference to `MPI_FINALIZE'
>
> collect2: ld returned 1 exit status
>
> link failed.
>
> mpif77: No such file or directory
>
>
> I read that it could have to do with the compiler used when
> installing LAM and the compiler used when compiling the
> application. So, I installed LAM forcing it to use f95 when
> configuring:
>
>
> ./configure FC=f95 -with-fc=f95 -prefix=/opt/lam/lam-7_1_3/
>
>
> Any suggestions what to try next??
>
|
