MPI df hangs

[hungpham2017]

Hello,

Recently, I have updated the lasted PySCF 1.6 and mpi4pyscf. As a result, mpi4pyscf didn't work.

1. When I used mpi4pyscf df module interactively to compute JK, the mpi_df just hangs while the serial df works just fine

from pyscf.pbc import gto, df
from mpi4pyscf.pbc import df as mpidf
import numpy
cell = gto.M(atom='He 0 0 0; He 0 0 1', a=numpy.eye(3)*4, mesh=[11]*3)
kpts = cell.make_kpts([1,1,2], wrap_around=True)
nao = cell.nao_nr()
dm = numpy.ones((nao,nao))

# df
mydf = df.MDF(cell, kpts)
vj, vk = mydf.get_jk(dm, kpts=kpts[0])

# MPI df
myMPIdf = mpidf.MDF(cell, kpts)
vj, vk = myMPIdf.get_jk(dm, kpts=kpts[0])

2. If I run the script above using mpirun, I got this error message:

Traceback (most recent call last):
  File "/home/gagliard/phamx494/mpi4pyscf/__init__.py", line 32, in <module>
    mpi.pool.wait()
  File "/home/gagliard/phamx494/mpi4pyscf/tools/mpi_pool.py", line 111, in wait
    ans = self.function(*task)  # task = worker_args
  File "/home/gagliard/phamx494/mpi4pyscf/tools/mpi.py", line 565, in _distribute_call
    return fn(dev, *args, **kwargs)
  File "/home/gagliard/phamx494/mpi4pyscf/pbc/df/mdf_jk.py", line 78, in get_jk
    vj, vk = mdf_jk.get_jk(mydf, dm, hermi, kpt, kpt_band, with_j, with_k, exxdiv)
  File "/home/gagliard/phamx494/pyscf/pyscf/pbc/df/mdf_jk.py", line 90, in get_jk
    vj1, vk1 = df_jk.get_jk(mydf, dm, hermi, kpt, kpts_band, with_j, with_k, None)
  File "/home/gagliard/phamx494/pyscf/pyscf/pbc/df/df_jk.py", line 257, in get_jk
    mydf.build(kpts_band=kpts_band)
  File "/home/gagliard/phamx494/mpi4pyscf/pbc/df/df.py", line 43, in build
    log.debug('MPI info (rank, host, pid)  %s', mpi.platform_info())
  File "/home/gagliard/phamx494/mpi4pyscf/tools/mpi.py", line 737, in platform_info
    return pool.apply(info, (), ())
  File "/home/gagliard/phamx494/mpi4pyscf/tools/mpi_pool.py", line 123, in apply
    self.wait()
  File "/home/gagliard/phamx494/mpi4pyscf/tools/mpi_pool.py", line 111, in wait
    ans = self.function(*task)  # task = worker_args
TypeError: info() takes 0 positional arguments but 2 were given

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "He2.py", line 4, in <module>
    from mpi4pyscf.pbc import df as mpidf
  File "/home/gagliard/phamx494/mpi4pyscf/__init__.py", line 34, in <module>
    traceback.print_exc(err)
  File "/home/gagliard/phamx494/anaconda/lib/python3.6/traceback.py", line 159, in print_exc
    print_exception(*sys.exc_info(), limit=limit, file=file, chain=chain)
  File "/home/gagliard/phamx494/anaconda/lib/python3.6/traceback.py", line 100, in print_exception
    type(value), value, tb, limit=limit).format(chain=chain):
  File "/home/gagliard/phamx494/anaconda/lib/python3.6/traceback.py", line 497, in __init__
    capture_locals=capture_locals)
  File "/home/gagliard/phamx494/anaconda/lib/python3.6/traceback.py", line 332, in extract
    if limit >= 0:
TypeError: '>=' not supported between instances of 'TypeError' and 'int'

Not sure where it goes wrong with the new update.
Thank you