Use --parallel when running tests on CI and in precommit

[karyon]

On my machine, wall clock time for "./manage.py run test" drops from 43 to 20s with 6 threads. More threads increased the overhead for creating databases more than it reduced the time for the actual tests. The github runners only have 4 cores.

[karyon]

I put a hardcoded value there because on my machine with a i7-13800H (6 P-cores with hyperthreadig, 8 E-cores) it would start 20 processes and actually run quite a bit slower while burning much more CPU cycles than with 6 processes. When I opened the PR I thought 6 was an "organic" sweet spot based on my measurements, but it's suspiciously equal to the number of P-cores :) I would still think that in the absence of a percentage option a hardcoded 4-8 would be better for most, but I don't mind much. Feel free to edit the PR and merge it :)

…

On March 11, 2026 12:15:24 PM GMT+01:00, Richard Ebeling ***@***.***> wrote: @richardebeling commented on this pull request. > @@ -43,7 +45,7 @@ jobs: start-db: true - name: Run tests - run: python manage.py test --shuffle + run: python manage.py test --shuffle --parallel 4 Django can only do "auto", anything else would require code from us. I think "auto" will be fine, even for the weaker laptops. Note that this is only `precommit`. For `test`, we don't change the default. -- Reply to this email directly or view it on GitHub: #2666 (comment) You are receiving this because you authored the thread. Message ID: ***@***.***>

[richardebeling]

Updated everything to use auto detection. For environments where this is not appropriate, users can set the DJANGO_TESTS_PROCESSES environment variable to override the auto choice. I think this is a reasonable compromise.