Well, this is very odd. That was the set of instructions I used, and I
followed them to set up the build last night. Last night it failed three
times, I just checked Bmake.inc again, everything looked right, I tried
the build again and this time it worked. <Insert sheepish expression
here>. Um, I'll go away now and stop being a Pebkac....
Damien
> Hi,
>
> Please have a look at this page -
> http://www.lam-mpi.org/3rd-party/blacs.php3
>
> I was able to build the most recent version of BLACS for MPI, by
> changing MPIdir macro in Bmake.inc to the directory where I installed
> LAM-7.0.3.
>
> Hope this helps.
> --
> Shashwat Srivastav
> LAM / MPI Developer (http://www.lam-mpi.org)
> Indiana University
> http://www.cs.indiana.edu/~ssrivast
>
> On Tuesday, Nov 25, 2003, at 00:51 America/Chicago, Damien Hocking
> wrote:
>
>> Hi folks. I'm a bit of a newbie to Unix builds, so this might be a
>> dumb sort of question that I should be able to work out myself.
>>
>> I just built LAM 7.0.3, had no problems. I'm in the process of
>> building BLACS, LAPACK and SCALAPACK for the MUMPS linear solver. I
>> followed the instructions for building a LAM version of BLACS using
>> the mpicc and mpif77 from LAM and started the build. Everything is
>> fine for quite a while, until it tries to build BI_TransUserComm.c and
>> it produces the error "No rule to make
>> target /mpi/share/include/mpif.h needed by mpif.h". A circular
>> reference from a Fortran header? I've never seen that before.
>>
>> It's a Slackware 9.1 OS on a PC. Can anyone help?
>>
>> Damien
>> _______________________________________________
>> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>>
>
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
|
