Help Requested

Hi All

Does anyone have any idea about the error massages I am getting which is attached at bellow?

Thanks

Orhan

The output (if any) follows:

  Initial static refinement
   iteration, no. not moved = 0 0
   iteration, no. not moved = 0 24
   iteration, no. not moved = 1 0
   iteration, no. not moved = 0 0
  after amr_morton_order : pe lnb 0 85
  after amr_morton_order : pe lnb 1 85
  after amr_morton_order : pe lnb 2 85
  after amr_morton_order : pe lnb 6 85
  after amr_morton_order : pe lnb 4 85
  after amr_morton_order : pe lnb 5 85
  after amr_morton_order : pe lnb 3 85
  after amr_morton_order : pe lnb 7 85
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
   #000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
     major(14): Dataspace interface
     minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
   #000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
     major(14): Dataspace interface
     minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
   #000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
     major(14): Dataspace interface
     minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
   #000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
     major(14): Dataspace interface
     minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
   #000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
     major(14): Dataspace interface
     minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
   #000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent

···

--
---------------------------
Dr. Orhan Donmez
Max-Planck-Institute for Gravitational Physics
(Albert-Einstein-Institute)
Division for Astrophysical Relativity
Am Muhlenberg 1
D-14476 Potsdam-Golm, Germany
e-mail: odonmez@aei.mpg.de
tel: 49-331-5677186

----------------------------------------------------------------------
This mailing list is for HDF software users discussion.
To subscribe to this list, send a message to hdf-forum-subscribe@hdfgroup.org.
To unsubscribe, send a message to hdf-forum-unsubscribe@hdfgroup.org.

I have seen recently several emails reporting errors when reading/writing HDF5 files. My 2 cents, maybe useful to someone.

I do not think putting HDF5 stuff directly in your complex codes make sense. It is way to easy to make mistakes, as you obviously have found. HDF5 is too low level for such usage. The way you do in in your codes is very error prone. You can of course sort it out at some point but then you go to another code, add lots of HDF5 stuff again and make similar errors again. I have gone through this myself too often.

What I think makes sense is to write a helper C++ class / FORTRAN module to do common tasks and a small test program. E.g. for C++ I use

class HDF5Writer
{
  public:
    HDF5Writer()
    {
      compression = 0;
      loud = 0;
      file = -1;
    };
    int open(const std::string&, const std::string& = "overwrite");
    int close();
    int createGroup(const std::string&);
    int write(const Math::Matrix<double>&, const std::string&);
    int write(const Math::Matrix<int>&, const std::string&);
    int write(const Math::Vector<double>&, const std::string&);
    int write(const Math::Vector<int>&, const std::string&);
    
    int compression;
    bool loud;
    std::vector<int> chunkSize;
  private:
    int writeData(const void*, const std::string&, const std::vector<int>&, const std::string&)const;
    hid_t file;
};

HDF5Reader would follow the same fashion as well as would a FORTRAN module. The idea is to get your data in/out of HDF5 file in 1-2 function calls (like in MATLAB), and not in 20-50. Of course I have spent considerable amount of time debugging the implementation of such an interface, but once done, the errors come much less frequently in the future.

regards,
Dominik

···

--
Dominik Szczerba, Ph.D.
Biomedical Simulation Group
Computer Vision Lab CH-8092 Zurich
http://www.vision.ee.ethz.ch/~domi

----------------------------------------------------------------------
This mailing list is for HDF software users discussion.
To subscribe to this list, send a message to hdf-forum-subscribe@hdfgroup.org.
To unsubscribe, send a message to hdf-forum-unsubscribe@hdfgroup.org.

Orhan,

Could you please check that you are providing the correct start, stride, count and block sizes to the H5Sselect_hyperslab function when you select a hyperslab for each MPI process?

If you are writing an unlimited dataset, please make sure that you extend it with the H5Dextend call (always collective) first.

Elena

···

On Jun 13, 2008, at 3:16 AM, Orhan Donmez wrote:

Hi All

Does anyone have any idea about the error massages I am getting which is attached at bellow?

Thanks

Orhan

The output (if any) follows:

Initial static refinement
iteration, no. not moved = 0 0
iteration, no. not moved = 0 24
iteration, no. not moved = 1 0
iteration, no. not moved = 0 0
after amr_morton_order : pe lnb 0 85
after amr_morton_order : pe lnb 1 85
after amr_morton_order : pe lnb 2 85
after amr_morton_order : pe lnb 6 85
after amr_morton_order : pe lnb 4 85
after amr_morton_order : pe lnb 5 85
after amr_morton_order : pe lnb 3 85
after amr_morton_order : pe lnb 7 85
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
#000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
   major(14): Dataspace interface
   minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
#000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
   major(14): Dataspace interface
   minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
#000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
   major(14): Dataspace interface
   minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
#000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
   major(14): Dataspace interface
   minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
#000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent
   major(14): Dataspace interface
   minor(04): Out of range
HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 MPI-process 7. Back trace follows.
#000: H5Dio.c line 676 in H5Dwrite(): file selection+offset not within extent

--
---------------------------
Dr. Orhan Donmez
Max-Planck-Institute for Gravitational Physics
(Albert-Einstein-Institute)
Division for Astrophysical Relativity
Am Muhlenberg 1
D-14476 Potsdam-Golm, Germany
e-mail: odonmez@aei.mpg.de
tel: 49-331-5677186

----------------------------------------------------------------------
This mailing list is for HDF software users discussion.
To subscribe to this list, send a message to hdf-forum-subscribe@hdfgroup.org.
To unsubscribe, send a message to hdf-forum-unsubscribe@hdfgroup.org.

----------------------------------------------------------------------
This mailing list is for HDF software users discussion.
To subscribe to this list, send a message to hdf-forum-subscribe@hdfgroup.org.
To unsubscribe, send a message to hdf-forum-unsubscribe@hdfgroup.org.