3
0
mirror of https://github.com/triqs/dft_tools synced 2024-12-26 22:33:48 +01:00

Fixed freezing issue in CSC run; better exception handling

In some MPI implementations the self-consistent run of `sc_dmft.py`
resulted in random freezings. This issue was fixed by polling the
VASP lock file only on the master node.

Also, better exception handling is implemented, allowing all nodes
to exit properly in case of error.
This commit is contained in:
Oleg E. Peil 2016-12-31 14:42:43 +01:00
parent 61f11bcff4
commit ac46611e51

View File

@ -33,6 +33,13 @@ import sys
import pytriqs.utility.mpi as mpi import pytriqs.utility.mpi as mpi
import converter import converter
xch = sys.excepthook
def excepthook(typ, value, traceback):
xch(typ, value, traceback)
if mpi.MPI.COMM_WORLD.size > 1:
mpi.MPI.COMM_WORLD.Abort(1)
sys.excepthook = excepthook
debug = True debug = True
# #
# Helper functions # Helper functions
@ -41,7 +48,11 @@ def sigint_handler(signal, frame):
raise SystemExit(1) raise SystemExit(1)
def is_vasp_lock_present(): def is_vasp_lock_present():
return os.path.isfile('./vasp.lock') res_bool = False
if mpi.is_master_node():
res_bool = os.path.isfile('./vasp.lock')
res_bool = mpi.bcast(res_bool)
return res_bool
def is_vasp_running(vasp_pid): def is_vasp_running(vasp_pid):
""" """
@ -122,34 +133,15 @@ def run_all(vasp_pid, dmft_cycle, cfg_file, n_iter):
if debug: print bcolors.MAGENTA + "rank %s"%(mpi.rank) + bcolors.ENDC if debug: print bcolors.MAGENTA + "rank %s"%(mpi.rank) + bcolors.ENDC
err = 0 err = 0
exc = None exc = None
try: if debug: print bcolors.BLUE + "plovasp: rank %s"%(mpi.rank) + bcolors.ENDC
if debug: print bcolors.BLUE + "plovasp: rank %s"%(mpi.rank) + bcolors.ENDC if mpi.is_master_node():
if mpi.is_master_node(): converter.generate_and_output_as_text(cfg_file, vasp_dir='./')
converter.generate_and_output_as_text(cfg_file, vasp_dir='./') # Read energy from OSZICAR
# Read energy from OSZICAR dft_energy = get_dft_energy()
dft_energy = get_dft_energy()
except Exception, exc:
err = 1
err = mpi.bcast(err)
if err:
if mpi.is_master_node():
raise exc
else:
raise SystemExit(1)
mpi.barrier() mpi.barrier()
try: if debug: print bcolors.GREEN + "rank %s"%(mpi.rank) + bcolors.ENDC
if debug: print bcolors.GREEN + "rank %s"%(mpi.rank) + bcolors.ENDC corr_energy, dft_dc = dmft_cycle()
corr_energy, dft_dc = dmft_cycle()
except:
if mpi.is_master_node():
print " master forwarding the exception..."
raise
else:
print " rank %i exiting..."%(mpi.rank)
raise SystemExit(1)
mpi.barrier() mpi.barrier()
if mpi.is_master_node(): if mpi.is_master_node():