Skip to content

fixes for parallel section and add to toc #17

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

Merged
fjansson merged 1 commit into from
Sep 11, 2026
Merged

fixes for parallel section and add to toc #17

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions book/_toc.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,7 @@ parts:
# - file: running/chemistry.ipynb
# - file: running/ibm.ipynb
# - file: running/openbcs.ipynb
- file: running/parallel.md
- file: running/gpu.md
- file: running/quirks.md
- caption: For DALES developers
Expand Down
8 changes: 4 additions & 4 deletions book/running/parallel.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,7 @@ The best configuration (choice of `itot`, `jtot`, `nprocx`, `nprocy`) for a give

* a reasonable number of grid rows and columns per task (`imax = itot/nprocx` and `jmax = jtot/nprocy`) is around 32-64.

* the Fourier-transform-based Poisson solver works best (fastest) when the total grid sizes (`itot`, `jtot`) factorize into powers of small numbers, e.g. 2**n * 3**m.
* the Fourier-transform-based Poisson solver works best (fastest) when the total grid sizes (`itot`, `jtot`) factorize into powers of small numbers, e.g. `2**n * 3**m`.

* it seems square tiles are efficient, or tiles that are
longer along x than y, perhaps up to two times longer.
Expand All @@ -37,9 +37,9 @@ the `total_number_of_cores = number_of_nodes * cores_per_node` should match the
i.e. an integer number of tile rows fit on one node. This means that the
communication in the x direction is always within a node.

* often the number of cores in a node is of the form 2**n * 3**m, where
* often the number of cores in a node is of the form `2**n * 3**m`, where
m is small, e.g. 48, 128, 192.
We have had good experiences with domain sizes (`itot`, `jtot`) that are 2**n or 3*2**n.
We have had good experiences with domain sizes (`itot`, `jtot`) that are `2**n` or `3*2**n`.

To increase speed and to save memory, also consider running in
single-precision mode, see [](sec:compilation:single)=
single-precision mode, see [](sec:compilation:single)
Loading