Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

PYTHONHASHSEED #5

Open
rainwoodman opened this issue Aug 10, 2017 · 0 comments
Open

PYTHONHASHSEED #5

rainwoodman opened this issue Aug 10, 2017 · 0 comments

Comments

@rainwoodman
Copy link
Collaborator

It is important that all MPI ranks carry the same PYTHONHASHSEED.
python destroys member variables in a random order determined by the hashseed. If the seed is different it is possible some MPI enabled objects goes out of scope earlier than others. This is actually a MPI related issue and shall be noted in mpi4py.

I see a strange stalling with SGI's MPT and anaconda 3.5 due to this on Plaedias

@nickhand -- I suggest you mention this in the nbodykit documentation.
rainwoodman added a commit to rainwoodman/python-mpi-bcast that referenced this issue Aug 11, 2017
@rainwoodman
set PYTHONHASHSEED.
914c139 
see MP-Gadget/pfft-python#5
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@rainwoodman