Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

Subject: Re: [OMPI users] MPI-IO puzzlement
From: Ricardo Reis (rreis_at_[hidden])
Date: 2012-05-15 16:13:16


On Tue, 15 May 2012, Jeff Squyres wrote:

> On May 15, 2012, at 2:19 PM, Ricardo Reis wrote:
>
>>>> INTEGER(kind=MPI_OFFSET_KIND) :: offset
>>>>
>>>> MPI_OFFSET_KIND is insuficient to represent my offset...
>>>
>>> Is it not a 64 bit integer for your compiler?
>
> I'm still interested in the answer to this question.

printing the result of

bit_size(offset)

does give the value of 64

I reckon I had an error in my debug code, I was truncating the output
format, that explains why I'm chasing a gambuzino with this point.

>>> There *is* a bug in OMPI at the moment that we're casting the result back down to 32 bits. I can give you a patch for that, if you'd like to try it.
>>
>> I'll get it from svn and try it out. On another note, me thinks with a derived_data type I nail two with one stroke. This problem and the limitation of 2GB per core when writing files...
>
> You should be able to get just a patch that you can apply to your v1.6 tarball here (the trunk patch was different because some filenames have changed between the v1.6 branch and the trunk):
>
> https://svn.open-mpi.org/trac/ompi/raw-attachment/ticket/3095/fortran-file-int-cast-fix.diff

hmmm... I downloaded the tarball of 1.6, cd to the openmpi-1.6 directory
and tried to apply the patch with

patch < /home/rreis/fortran-file-int-cast-fix.diff

but the result was

rreis_at_galego: /opt/build/openmpi-1.6
5044.0 $ patch -p1 < /home/rreis/fortran-file-int-cast-fix.diff
patching file mpi/f77/file_get_view_f.c
Hunk #1 FAILED at 9.
Hunk #2 FAILED at 71.
2 out of 2 hunks FAILED -- saving rejects to file
mpi/f77/file_get_view_f.c.rej
patching file mpi/f77/file_get_position_f.c
Hunk #1 FAILED at 9.
Hunk #2 FAILED at 65.
2 out of 2 hunks FAILED -- saving rejects to file
mpi/f77/file_get_position_f.c.rej
patching file mpi/f77/file_get_size_f.c
Hunk #1 FAILED at 9.
Hunk #2 FAILED at 65.
2 out of 2 hunks FAILED -- saving rejects to file
mpi/f77/file_get_size_f.c.rej

so I'm probably doing something wrong here?

  Ricardo Reis

  'Non Serviam'

  PhD/MSc Mechanical Engineering | Lic. Aerospace Engineering

  Computational Fluid Dynamics, High Performance Computing, Turbulence
  http://www.lasef.ist.utl.pt

  Cultural Instigator @ Rádio Zero
  http://www.radiozero.pt

  http://www.flickr.com/photos/rreis/

  contacts: gtalk: kyriusan_at_[hidden] skype: kyriusan

  Institutional Address:

  Ricardo J.N. dos Reis
  IDMEC, Instituto Superior Técnico, Technical University of Lisbon
  Av. Rovisco Pais
  1049-001 Lisboa
  Portugal

                       - email sent with alpine 2.00 -