-
Notifications
You must be signed in to change notification settings - Fork 26
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
Bump python versions #923
Bump python versions #923
Conversation
| benchmark_name | dt(%) | dt(s) | t_new(s) | t_old(s) |
| -------------------------------------- | ---------------------- | ---------------------- | ---------------------- | ---------------------- |
test_build_transform_fft_lowres | +2.77 +/- 1.75 | +2.50e-04 +/- 1.58e-04 | 9.28e-03 +/- 9.7e-05 | 9.03e-03 +/- 1.2e-04 |
test_build_transform_fft_midres | +3.07 +/- 1.54 | +2.72e-03 +/- 1.36e-03 | 9.12e-02 +/- 8.8e-04 | 8.85e-02 +/- 1.0e-03 |
test_build_transform_fft_highres | +2.57 +/- 0.98 | +1.18e-02 +/- 4.49e-03 | 4.69e-01 +/- 2.7e-03 | 4.57e-01 +/- 3.6e-03 |
test_equilibrium_init_lowres | +0.04 +/- 3.26 | +1.11e-04 +/- 8.86e-03 | 2.71e-01 +/- 5.3e-03 | 2.71e-01 +/- 7.1e-03 |
test_equilibrium_init_medres | +3.34 +/- 2.03 | +2.01e-02 +/- 1.22e-02 | 6.22e-01 +/- 9.4e-03 | 6.02e-01 +/- 7.8e-03 |
test_equilibrium_init_highres | +0.93 +/- 1.80 | +2.05e-02 +/- 3.96e-02 | 2.22e+00 +/- 2.4e-02 | 2.20e+00 +/- 3.2e-02 |
test_objective_compile_dshape_current | +0.21 +/- 8.86 | +8.54e-03 +/- 3.57e-01 | 4.04e+00 +/- 2.4e-01 | 4.03e+00 +/- 2.7e-01 |
test_objective_compile_atf | -1.62 +/- 5.01 | -1.29e-01 +/- 4.00e-01 | 7.85e+00 +/- 2.6e-01 | 7.98e+00 +/- 3.0e-01 |
-test_objective_compute_dshape_current | +19.41 +/- 3.07 | +1.01e-03 +/- 1.60e-04 | 6.23e-03 +/- 1.4e-04 | 5.21e-03 +/- 8.2e-05 |
test_objective_compute_atf | +8.26 +/- 4.91 | +9.28e-04 +/- 5.52e-04 | 1.22e-02 +/- 2.4e-04 | 1.12e-02 +/- 5.0e-04 |
test_objective_jac_dshape_current | -2.37 +/- 6.93 | -1.11e-03 +/- 3.25e-03 | 4.57e-02 +/- 2.2e-03 | 4.69e-02 +/- 2.4e-03 |
test_objective_jac_atf | -0.97 +/- 3.12 | -2.11e-02 +/- 6.82e-02 | 2.17e+00 +/- 5.7e-02 | 2.19e+00 +/- 3.8e-02 |
test_perturb_1 | +1.44 +/- 4.28 | +2.35e-01 +/- 7.00e-01 | 1.66e+01 +/- 4.2e-01 | 1.63e+01 +/- 5.6e-01 |
test_perturb_2 | +0.12 +/- 4.75 | +2.65e-02 +/- 1.02e+00 | 2.16e+01 +/- 9.2e-01 | 2.15e+01 +/- 4.5e-01 |
test_proximal_jac_atf | +0.39 +/- 1.61 | +2.84e-02 +/- 1.18e-01 | 7.32e+00 +/- 9.4e-02 | 7.29e+00 +/- 7.0e-02 |
test_proximal_freeb_compute | +1.60 +/- 1.26 | +1.97e-03 +/- 1.55e-03 | 1.25e-01 +/- 7.5e-04 | 1.23e-01 +/- 1.4e-03 |
test_proximal_freeb_jac | +0.15 +/- 1.27 | +1.09e-02 +/- 8.96e-02 | 7.08e+00 +/- 7.9e-02 | 7.07e+00 +/- 4.3e-02 | |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #923 +/- ##
=======================================
Coverage 94.96% 94.96%
=======================================
Files 83 83
Lines 21012 21012
=======================================
+ Hits 19953 19954 +1
+ Misses 1059 1058 -1 |
@@ -204,7 +204,7 @@ Commit `a2fe711ffa3f` (an older version of the `master` branch) was tested to wo | |||
|
|||
module load anaconda3/2020.11 cudatoolkit/11.1 cudnn/cuda-11.1/8.0.4 | |||
|
|||
conda create --name desc-env python=3.8 |
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.
was this tested to work? I guess it should not really matter but the power architectures were tricky to work with install-wise
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.
i had 3.9 installed when i was running stuff on traverse
Here is the description for this PR by @YigitElma - I created a new workflow that runs every 2 days and updates the requirements and stores them in a venv that is cached - Black, unit, regression, notebook and benchmark workflows use the cached venv (which takes around 4-8seconds whereas installing dependencies usually more than a minute) - Increase the spitting of `unit`, `notebook`, and `benchmark` workflows. - For some reason I couldn't make the linting test faster using venv (so set it back to old version which is fine), actually for some reason venv made it way slower around 8 minutes?!!!? ### Other Fixes - We stopped using Python 3.8 in our tests with #923 but we forgot to remove them from weekly tests. Since then (I guess the dates check) the weekly tests fail (for the last 3 months). I removed Python 3.8 from weekly tests.
Python 3.8 will reach end of life later this year, and some packages (such as diffrax) already require Python 3.9
This PR moves us in that direction, removing testing against Python 3.8 and adds it for Python 3.12 (the most recent).
Also updates docs accordingly.