[gpaw-users] Installing GPAW with parallel and scalapack support

[Gaël Donval]

> >> Still not working scalapack even though OpenBlas, PBLAS, BLACS and
> >> Scalapack itself passed all their respective built-in tests perfectly.
> > Got it working by disabling no_numpy_depreciated_api precompiler flag.
> > However, TypeError: Not a proper NumPy array for MPI communication.
> > remains as well as the RunetimeError.
> 
> What are the stack traces?
> 
> Regards
> Ask

Here we are:


parallel/ut_hsblacs.py                  1.345  FAILED! (rank 0,1,2,3)
#############################################################################
RANK 0,1,2,3:
Traceback (most recent call last):
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python2.7/site-packages/gpaw/test/__init__.py", line 489, in run_one
    execfile(filename, loc)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python2.7/site-packages/gpaw/test/parallel/ut_hsblacs.py", line 389, in <module>
    raise SystemExit('Test failed. Check ut_hsblacs.log for details.')
SystemExit: Test failed. Check ut_hsblacs.log for details.
#############################################################################





fileio/parallel.py                 Traceback (most recent call last):
  File "/u/gdonval/opt_gcc/gpaw_python/bin/gpaw-test", line 150, in <module>
    nfailed = run()
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python2.7/site-packages/gpaw/test/test.py", line 125, in run
    failed = TestRunner(tests, jobs=opt.jobs, show_output=opt.show_output).run()
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python2.7/site-packages/gpaw/test/__init__.py", line 417, in run
    self.run_single()
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python2.7/site-packages/gpaw/test/__init__.py", line 439, in run_single
    self.run_one(test)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python2.7/site-packages/gpaw/test/__init__.py", line 505, in run_one
    mpi.ibarrier(timeout=60.0)  # guard against parallel hangs
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python2.7/site-packages/gpaw/mpi/__init__.py", line 750, in ibarrier
    raise RuntimeError('MPI barrier timeout.')
RuntimeError: MPI barrier timeout.
GPAW CLEANUP (node 0): <type 'exceptions.RuntimeError'> occurred.  Calling MPI_Abort!
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD 
with errorcode 42.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------





parallel/overlap.py              0.363  FAILED! (rank 0,1)
#############################################################################
RANK 0,1:
Traceback (most recent call last):
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/test/__init__.py", line 489, in run_one
    execfile(filename, loc)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/test/parallel/overlap.py", line 124, in <module>
    psit_mG = run(psit_mG)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/test/parallel/overlap.py", line 87, in run
    S_nn = overlap.calculate_matrix_elements(psit_mG, P_ani, S, dS)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/hs_operators.py", line 365, in calculate_matrix_elements
    sbuf_In, rbuf_In, cycle_P_ani)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/hs_operators.py", line 207, in _initialize_cycle
    self.req2.append(band_comm.send(sbuf_In, rankm, 31, False))
TypeError: Not a proper NumPy array for MPI communication.
#############################################################################





pw/slab.py                       1.499  FAILED! (rank 0,1)
#############################################################################
RANK 0,1:
Traceback (most recent call last):
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/test/__init__.py", line 489, in run_one
    execfile(filename, loc)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/test/pw/slab.py", line 20, in <module>
    BFGS(slab).run(fmax=0.01)
  File "/u/gdonval/opt_gcc/ase/ase/optimize/optimize.py", line 114, in run
    f = self.atoms.get_forces()
  File "/u/gdonval/opt_gcc/ase/ase/atoms.py", line 669, in get_forces
    forces = self._calc.get_forces(self)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/aseinterface.py", line 69, in get_forces
    force_call_to_set_positions=force_call_to_set_positions)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/paw.py", line 269, in calculate
    self.occupations):
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/scf.py", line 46, in run
    wfs.eigensolver.iterate(hamiltonian, wfs)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/eigensolvers/eigensolver.py", line 63, in iterate
    wfs.overlap.orthonormalize(wfs, kpt)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/overlap.py", line 76, in orthonormalize
    S_nn = operator.calculate_matrix_elements(psit_nG, P_ani, S, dS)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/hs_operators.py", line 365, in calculate_matrix_elements
    sbuf_In, rbuf_In, cycle_P_ani)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/hs_operators.py", line 207, in _initialize_cycle
    self.req2.append(band_comm.send(sbuf_In, rankm, 31, False))
TypeError: Not a proper NumPy array for MPI communication.
#############################################################################





exx_acdf.py                      5.825  FAILED! (rank 0,1)
#############################################################################
RANK 0,1:
Traceback (most recent call last):
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/test/__init__.py", line 489, in run_one
    execfile(filename, loc)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/test/exx_acdf.py", line 25, in <module>
    E_k = E + calc.get_xc_difference(exx)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/aseinterface.py", line 423, in get_xc_difference
    return self.hamiltonian.get_xc_difference(xc, self.density) * Hartree
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/hamiltonian.py", line 405, in get_xc_difference
    xc.calculate_exx()
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/xc/hybridk.py", line 273, in calculate_exx
    kpt2_q[0].start_sending(srank)
  File "/u/gdonval/opt_gcc/gpaw_python/lib/python/gpaw/xc/hybridk.py", line 77, in start_sending
    self.kd.comm.send(P_In, rank, block=False, tag=5)]
TypeError: Not a proper NumPy array for MPI communication.
#############################################################################