Tests known to fail with PETSc configured with gfortran -fcheck=bounds,do,mem,pointer,recursion:

================================================================================

Fortran runtime error: Actual string length is shorter than the declared one for dummy argument 'transa' (-873257856/1)
At line 188 of file dgemm.f

geh - this only occurs when there are 15 chemical componetns, > 1 grid cell and with ILU preconditioning. The error occurs from within PETSc.
--------------------------------
543_hanford_srfcplx_param-np8.in
surface_complexation_vf.in
surface_complexation_den.in

================================================================================

 Closing group: Materials
 Closing hdf5 file: ./543_river_setup.h5
At line 57 of file /home/gehammo/software/lib/petsc-git/src/sys/f90-src/fsrc/f90_fwrap.F
Fortran runtime error: Index '1' of dimension 1 of array 'ptr' above upper bound of 0

geh - this occurs within:

      call VecGetArrayF90(map_ids_1,vec_ptr,ierr);CHKERRQ(ierr)
      do ii=1,dataset_map_hdf5%map_dims_local(2)
        vec_ptr(ii)=dataset_map_hdf5%mapping(1,ii)
      enddo
->    call VecRestoreArrayF90(map_ids_1,vec_ptr,ierr);CHKERRQ(ierr)
    
      call VecScatterBegin(vec_scatter,map_ids_1,map_ids_2, &
                       INSERT_VALUES,SCATTER_FORWARD,ierr);CHKERRQ(ierr)

in PatchCreateFlowConditionDatasetMap() when there are no entries in map_ids_1 on one of the processes.  DDT confirms that the local size is zero.

--------------------------------
543_river_het_map_dirich-np4

