Fixing issues related to linking parallel hdf5 on cmake

[dutta-alankar]

On our Freya cluster at MPCDF, I tried compiling Idefix using the following after setting IDEFIX_DIR:

module purge
module load gcc/15 cuda/13.0 cmake/4.0
OMPI=/freya/ptmp/mpa/adutt/openmpi_gpu-5.0
HDF5=/freya/ptmp/mpa/adutt/hdf5-2.0.0-openmpi_gpu-5.0/parallel
export PATH=$OMPI/bin:$HDF5/bin:$PATH
export LD_LIBRARY_PATH=$OMPI/lib:$HDF5/lib:${LD_LIBRARY_PATH:-}
CC=mpicc CXX=mpic++ cmake \
-DIdefix_MPI=ON -DIdefix_DEBUG=OFF \
-DIdefix_HDF5=ON -DKokkos_ENABLE_CUDA=ON \
-DKokkos_ARCH_AMPERE80=ON \
-DHDF5_ROOT=/freya/ptmp/mpa/adutt/hdf5-2.0.0-openmpi_gpu-5.0/parallel \
-DIdefix_PROBLEM_DIR=$PWD \
-GNinja -B ./build -S $IDEFIX_DIR cmake --build ./build/

This resulted in an error in linking hdf5 library. Following is the error message:

-- Found HDF5: hdf5-shared (found version "2.0.0")
-- Found HDF5 include directories: /freya/ptmp/mpa/adutt/hdf5-2.0.0-openmpi_gpu-5.0/parallel/include
CMake Error at CMakeLists.txt:122 (message):
Parallel HDF5 required for Idefix_MPI but the found HDF5 library does not
support it

This comes up despite having parallel support enabled hdf5 library. This PR addresses the issue and introduces a robust hdf5 linking section in CMakeLists.txt.
Apart from this, there are two minor edits to suppress unused variable warnings.

[Copilot]

Pull request overview

This PR adjusts Idefix’s CMake HDF5 discovery/linking to better support parallel HDF5 builds (especially when HDF5 is found via a CMake config package), and it also suppresses a couple of unused-variable warnings in C++ sources.

Changes:

• Update CMakeLists.txt to prefer HDF5 imported targets when available and add a robust parallel-HDF5 detection probe for MPI builds.
• Suppress unused-variable warnings in src/output/xdmf.cpp and src/fluid/boundary/axis.cpp using [[maybe_unused]].

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
CMakeLists.txt	Improves HDF5 linking target selection and adds a more reliable parallel-HDF5 detection mechanism when Idefix_MPI=ON.
src/output/xdmf.cpp	Adds [[maybe_unused]] for tot_dim in the DIMENSIONS==2 branch to silence warnings.
src/fluid/boundary/axis.cpp	Marks local variables as [[maybe_unused]] in Axis::ExchangeMPI to suppress unused-variable warnings.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[dutta-alankar]

This is first of few PRs (as discussed in #349) that I am going to put in the next few weeks, final being the introduction of the Tabulated radiative cooling module once all these other changes get merged.

[dutta-alankar]

@svalat I plan to put in the Cooling module on top of these commits. So, it would be nice if you could review this and this PR gets merged.

[svalat]

A minor change request and it looks ok to me, thanks for the proposal.

[dutta-alankar]

@svalat Let me know if you're happy with the changes introduced.

[svalat]

Hi, the patch is OK to me, thanks for it.
Waiting the status of the test suite to merge.
We just merged a fix about the test suite which might collide with your fixes about "maybe unused", I solved the conflict, if you can accept it.

[dutta-alankar]

Hi, the patch is OK to me, thanks for it. Waiting the status of the test suite to merge. We just merged a fix about the test suite which might collide with your fixes about "maybe unused", I solved the conflict, if you can accept it.

@svalat I do not seem to find the option to accept your change.