diff options
author | Fabio Zeiser <fzeiser@users.noreply.github.com> | 2020-01-24 18:33:02 +0100 |
---|---|---|
committer | Sebastian Berg <sebastian@sipsolutions.net> | 2020-01-24 09:33:02 -0800 |
commit | d5a444e9f263a1594643e64c37cb709074c59172 (patch) | |
tree | f04b2e1c237d409b6b62d77d8a3b8bf9dfff8e36 /numpy/array_api/_array_object.py | |
parent | 31e53bfddef063eba9c023999f8e5923a608ba71 (diff) | |
download | numpy-d5a444e9f263a1594643e64c37cb709074c59172.tar.gz |
DOC: Fix Multithreaded Generation example docs (gh-15367)
The old code examples did not actually run correctly and had a bug. They were
also never updated to use SeedSequence correctly. This fixes both.
* DOC, BUG: Fixes Multithreaded Generation docs
Working on #15365
The code runs now, but I get quite different runtimes than
in the example. May need some more work
* DOC: changes after PR review
* DOC: Changing input/output lines to sphinx standard
For ipython example: Adding `Out[i]` where i is the input cell number
and correcting vissual multiline inputs
* MAINT: Replace `jumped` by `SeedSequence` in multithreading snippet
After PR review, decided to replace `jumped` by `SeedSequence` as it is a more general method. Adopted the introduction to reflect this.
Also added a section on the reproducibility for different machines and number of threads + an example.
* MAINT: Changes after PR review
- Switch back to default_rng() instead of the Generator(BitGenerator)) syntax pursuant @rkern 's explanation in #15391
- small fixes
* MAINT: Delete unneccesary reference to PCG64
* MAINT: Added newline finxing codeblock error in docs
Trying to see whether this fixes circleci error
* MAINT: Clean up in imports of code snippet
Diffstat (limited to 'numpy/array_api/_array_object.py')
0 files changed, 0 insertions, 0 deletions