-
Notifications
You must be signed in to change notification settings - Fork 28
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
Refactoring Basis
/Transform
/Grid
stuff
#1405
base: master
Are you sure you want to change the base?
Conversation
| benchmark_name | dt(%) | dt(s) | t_new(s) | t_old(s) |
| -------------------------------------- | ---------------------- | ---------------------- | ---------------------- | ---------------------- |
-test_build_transform_fft_lowres | +14.92 +/- 4.41 | +7.90e-02 +/- 2.34e-02 | 6.09e-01 +/- 2.1e-02 | 5.29e-01 +/- 1.1e-02 |
test_equilibrium_init_medres | -3.73 +/- 1.91 | -1.56e-01 +/- 7.94e-02 | 4.01e+00 +/- 6.9e-02 | 4.16e+00 +/- 4.0e-02 |
+test_equilibrium_init_highres | -8.21 +/- 1.69 | -4.47e-01 +/- 9.17e-02 | 4.99e+00 +/- 7.9e-02 | 5.44e+00 +/- 4.6e-02 |
test_objective_compile_dshape_current | +2.81 +/- 4.64 | +1.14e-01 +/- 1.88e-01 | 4.17e+00 +/- 1.4e-01 | 4.05e+00 +/- 1.2e-01 |
-test_objective_compute_dshape_current | +6.46 +/- 1.62 | +3.32e-04 +/- 8.35e-05 | 5.48e-03 +/- 6.6e-05 | 5.15e-03 +/- 5.1e-05 |
test_objective_jac_dshape_current | +1.05 +/- 6.45 | +4.47e-04 +/- 2.76e-03 | 4.32e-02 +/- 2.0e-03 | 4.28e-02 +/- 1.9e-03 |
test_perturb_2 | +3.74 +/- 1.68 | +7.34e-01 +/- 3.30e-01 | 2.04e+01 +/- 2.7e-01 | 1.96e+01 +/- 1.9e-01 |
test_proximal_freeb_jac | -0.20 +/- 1.30 | -1.48e-02 +/- 9.62e-02 | 7.36e+00 +/- 4.9e-02 | 7.38e+00 +/- 8.3e-02 |
test_solve_fixed_iter | +1.32 +/- 1.36 | +4.20e-01 +/- 4.34e-01 | 3.22e+01 +/- 2.9e-01 | 3.18e+01 +/- 3.2e-01 |
test_LinearConstraintProjection_build | +2.73 +/- 1.59 | +2.80e-01 +/- 1.62e-01 | 1.05e+01 +/- 9.6e-02 | 1.02e+01 +/- 1.3e-01 |
test_build_transform_fft_midres | +5.38 +/- 5.63 | +3.25e-02 +/- 3.40e-02 | 6.37e-01 +/- 3.4e-02 | 6.04e-01 +/- 3.4e-03 |
test_build_transform_fft_highres | -5.36 +/- 7.69 | -5.20e-02 +/- 7.46e-02 | 9.18e-01 +/- 6.8e-02 | 9.70e-01 +/- 3.0e-02 |
+test_equilibrium_init_lowres | -10.74 +/- 2.66 | -4.42e-01 +/- 1.09e-01 | 3.67e+00 +/- 7.7e-02 | 4.11e+00 +/- 7.8e-02 |
test_objective_compile_atf | -3.20 +/- 2.10 | -2.75e-01 +/- 1.80e-01 | 8.31e+00 +/- 1.4e-01 | 8.58e+00 +/- 1.1e-01 |
test_objective_compute_atf | +0.19 +/- 3.47 | +3.03e-05 +/- 5.58e-04 | 1.61e-02 +/- 1.9e-04 | 1.61e-02 +/- 5.3e-04 |
test_objective_jac_atf | +0.45 +/- 2.01 | +8.78e-03 +/- 3.96e-02 | 1.98e+00 +/- 1.5e-02 | 1.97e+00 +/- 3.6e-02 |
test_perturb_1 | +4.59 +/- 3.12 | +6.74e-01 +/- 4.58e-01 | 1.54e+01 +/- 3.3e-01 | 1.47e+01 +/- 3.2e-01 |
test_proximal_jac_atf | -1.84 +/- 1.72 | -1.53e-01 +/- 1.43e-01 | 8.16e+00 +/- 9.4e-02 | 8.31e+00 +/- 1.1e-01 |
test_proximal_freeb_compute | -0.87 +/- 1.76 | -1.75e-03 +/- 3.54e-03 | 2.00e-01 +/- 2.3e-03 | 2.02e-01 +/- 2.7e-03 |
test_solve_fixed_iter_compiled | +0.59 +/- 1.35 | +1.22e-01 +/- 2.78e-01 | 2.07e+01 +/- 2.5e-01 | 2.06e+01 +/- 1.3e-01 | |
replace at axis with 0.0 for quantities we know must be the case |
@@ -257,29 +257,26 @@ def test_3d_tz(self): | |||
assert "Z" in data.keys() | |||
|
|||
assert "|F|" in data.keys() | |||
return fig |
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.
newer versions of pytest complain if tests return something that isn't used
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #1405 +/- ##
==========================================
- Coverage 95.69% 95.63% -0.07%
==========================================
Files 101 101
Lines 25633 25775 +142
==========================================
+ Hits 24529 24649 +120
- Misses 1104 1126 +22
|
Is this slowdown only on CPU? It was faster at some point but some recent change made it slower. |
Transform(method="fft")
now works with grids that have an even number of toroidal nodesBasis
classes for unique mode numbers, inverse indices etc, similar to gridsunique
optionBasis.evaluate
in favor of always using precomputed unique modes/nodes (Take advantage of separable basis functions #1530)