Here is how I call my data_out object to write HDF5,

DataOutBase::DataOutFilter
data_filter(DataOutBase::DataOutFilterFlags(true,true));

data_out.write_filtered_data(data_filter);

data_out.write_hdf5_parallel(data_filter,
                             write_mesh_file,
                             mesh_filename,
                             solution_filename,
                             mpi_comm);


Here are the lines pulled from data_out_base.cc (9.7.0 documentation)
833 node_dataset_id = H5Pcreate(H5P_DATASET_CREATE);
    #  ifdef DEAL_II_WITH_ZLIB
    H5Pset_deflate(node_dataset_id,
                   get_zlib_compression_level(flags.compression_level));
    H5Pset_chunk(node_dataset_id, 2, cell_ds_dim); //<------- Here it
sets some chunk size?
    #  endif
    cell_dataset = H5Dcreate(h5_mesh_file_id,
                             "cells",
                             H5T_NATIVE_UINT,
                             cell_dataspace,
                             H5P_DEFAULT,
                             node_dataset_id,
                             H5P_DEFAULT);
    H5Pclose(node_dataset_id);
    #  endif
 840   AssertThrow(cell_dataset >= 0, ExcIO());   // <--- line 8330 in my error


On Tue, Sep 16, 2025 at 11:29 PM 'Wolfgang Bangerth' via deal.II User Group
<[email protected]> wrote:

>
> On 9/16/25 05:04, Jalil Khan wrote:
> >
> >  From inspecting the source in data_out_base.cc, HDF5 uses chunked
> > layout, with chunk dimensions set internally in deal.II. At 512³, the
> > computed chunk size exceeds the HDF5 hard limit of 4 GB per chunk, which
> > leads to the failure.
> >
> > Is there a recommended way for users to avoid this error when writing
> > large output files?
>
> Jalil:
> that likely falls in the category of "the person who wrote these HDF5
> calls didn't think of this scenario, and nobody has run into it since
> then". Can you identify which line in data_out_base.cc specifically
> triggers the issue, and can you show how you call deal.II functions to
> run into this issue?
>
> In the end, this ought to be fixed of course. Any help is obviously
> always welcome!
>
> Best
>   Wolfgang
>
>
> --
> The deal.II project is located at http://www.dealii.org/
> For mailing list/forum options, see
> https://groups.google.com/d/forum/dealii?hl=en
> ---
> You received this message because you are subscribed to the Google Groups
> "deal.II User Group" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To view this discussion visit
> https://groups.google.com/d/msgid/dealii/f1bd81c7-9629-460b-977f-9b2c627bdd59%40colostate.edu
> .
>


-- 
Regards
J.R.Khan
8803985667

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/dealii/CAALjy%3DiFENpOMRjwek75AMh_F0SvTUnPoBo%3DtQ56upmX3s%2B%2Bug%40mail.gmail.com.

Reply via email to