[petsc-users] Questions about HDF5 parallel I/O

[Barry Smith]

> On Sep 3, 2020, at 12:19 PM, Sajid Ali <sajidsyed2021 at u.northwestern.edu> wrote:
> 
> Hi PETSc-developers, 
> 
> Currently, VecView loads an entire dataset from an hdf5 file into a PETSc vector (with any number of mpi ranks). Given that there is no routine to load a subset of an HDF5 dataset into a PETSc vector, the next best thing is to load the entire data-set into memory and select a smaller region as a sub-vector. Is there an example that demonstrates this ? (Mainly to get an idea on how to select a 2d array from a 3d array using a PETSc IS. Given that it's a regular 3D vector is it best to use a DMDA 3Dvec which gives ownership ranges that may aid with creating the IS ?)

  Sajid,

   DMDACreatePatchIS() might be what you need.

> 
> I've seen on earlier threads that XDMF can be used to create a map of where data is present in hdf5 files, is there an example for doing this with regular vectors to select subvectors as described above ?
> 
> Also, is it possible to have different sub-comms read different hdf5 groups ? 

   I think this should be possible if you have each sub-comm separately open the file. 

  Barry

> 
> Thank You,
> Sajid Ali | PhD Candidate
> Applied Physics
> Northwestern University
> s-sajid-ali.github.io <http://s-sajid-ali.github.io/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20200903/697c155b/attachment.html>