-
Notifications
You must be signed in to change notification settings - Fork 357
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.
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
Add documentation about deadlock in mpi processes #2758
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@jessica-mitchell Thanks! I have some minor suggestions for the example code, but overall this looks good and is a very useful addition to the documentation.
Co-authored-by: Nicolai Haug <39106781+nicolossus@users.noreply.github.com>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Great! This is exactly what was needed. Perhaps could you add a sentence like:
The PyNEST API was designed so that all parallel operations are handled through the nestkernel, hence the users don't need to handle MPI communication from the python script.
Ideally the script should be written in a serial manner, however if specific operations unrelated to the simulation execution are needed, the NEST API offers the Rank() function to retrieve the MPI rank of the process and handle process specific tasks.
Note that one should never call any ....
So that it is clear that parallelization is fully handled from the kernel and not the python script.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@jessica-mitchell Thanks for this PR! I wonder if it would not work even better if the "Important" box came at the end of the "MPI related commands" section that is now just below the "Important" box. Also, in the "Run distributed simulations" section, an "r" is missing in "distibuted simulation scripts". Could you fix that as part of this PR?
Concerning the suggestion by @JoseJVS, I think this is pretty much covered if one looks at the "Using distributed computing" section as a whole, so adding another sentence would make it a bit repetitive.
…mulator into docs-mpi-update
This PR add a note in the documentation in the distributed computing section of the parallel computing page, regarding the issue resolved in #2749.
See output here