Skip to content

Performance baseline pointers: add linux_latest and repoint CI #13

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
iahncajigas merged 1 commit into from
Mar 4, 2026
Merged

Performance baseline pointers: add linux_latest and repoint CI #13

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
5 changes: 4 additions & 1 deletion .github/workflows/performance-parity.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@ jobs:
--python-report output/performance/python_performance_report.json \
--matlab-report tests/performance/fixtures/matlab/performance_baseline_470fde8.json \
--policy parity/performance_gate_policy.yml \
--previous-python-report tests/performance/fixtures/python/performance_baseline_linux_20260304.json \
--previous-python-report tests/performance/fixtures/python/performance_baseline_linux_latest.json \
--report-out output/performance/performance_parity_report.json \
--csv-out output/performance/performance_parity_report.csv \
--fail-on-regression \
Expand All @@ -81,5 +81,8 @@ jobs:
output/performance/*.json
output/performance/*.csv
tests/performance/fixtures/matlab/performance_baseline_470fde8.json
tests/performance/fixtures/python/performance_baseline_linux_latest.json
tests/performance/fixtures/python/performance_baseline_linux_latest.csv
tests/performance/fixtures/python/performance_baseline_linux_20260304.json
tests/performance/fixtures/python/performance_baseline_linux_20260304.csv
if-no-files-found: warn
2 changes: 1 addition & 1 deletion README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -187,7 +187,7 @@ python tools/performance/compare_matlab_python_performance.py \
--python-report output/performance/python_performance_report.json \
--matlab-report tests/performance/fixtures/matlab/performance_baseline_470fde8.json \
--policy parity/performance_gate_policy.yml \
--previous-python-report tests/performance/fixtures/python/performance_baseline_linux_20260304.json \
--previous-python-report tests/performance/fixtures/python/performance_baseline_linux_latest.json \
--report-out parity/performance_parity_report.json \
--csv-out parity/performance_parity_report.csv \
--fail-on-regression \
Expand Down
2 changes: 1 addition & 1 deletion parity/CYCLE_VALIDATION_CHECKLIST.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@ Commands used each cycle:
- `python tools/reports/build_image_parity_pdfs.py --report-json <latest-json> --python-out output/pdf/image_mode_parity/python_pages.pdf --matlab-out output/pdf/image_mode_parity/matlab_pages.pdf --pairs-json output/pdf/image_mode_parity/pairs.json`
- `python tools/reports/check_pdf_image_parity.py --python-pdf output/pdf/image_mode_parity/python_pages.pdf --matlab-pdf output/pdf/image_mode_parity/matlab_pages.pdf --out-dir output/pdf/image_mode_parity --dpi 150 --ssim-threshold 0.70 --max-failing-pages 0`
- `python tools/performance/run_python_benchmarks.py --tiers S --repeats 5 --warmup 1 --out-json output/performance/python_performance_report.json --out-csv output/performance/python_performance_report.csv`
- `python tools/performance/compare_matlab_python_performance.py --python-report output/performance/python_performance_report.json --matlab-report tests/performance/fixtures/matlab/performance_baseline_470fde8.json --policy parity/performance_gate_policy.yml --previous-python-report tests/performance/fixtures/python/performance_baseline_linux_20260304.json --report-out output/performance/performance_parity_report.json --csv-out output/performance/performance_parity_report.csv --fail-on-regression --require-regression-env-match`
- `python tools/performance/compare_matlab_python_performance.py --python-report output/performance/python_performance_report.json --matlab-report tests/performance/fixtures/matlab/performance_baseline_470fde8.json --policy parity/performance_gate_policy.yml --previous-python-report tests/performance/fixtures/python/performance_baseline_linux_latest.json --report-out output/performance/performance_parity_report.json --csv-out output/performance/performance_parity_report.csv --fail-on-regression --require-regression-env-match`
- Local macOS reruns use `tests/performance/fixtures/python/performance_baseline_20260303.json` with the same command to satisfy strict env matching.

## Cycle 1
Expand Down
36 changes: 21 additions & 15 deletions tests/performance/fixtures/python/performance_baseline_linux_20260304.csv
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,16 +1,22 @@
case,tier,repeats,median_runtime_ms,mean_runtime_ms,std_runtime_ms,median_peak_memory_mb,summary
unit_impulse_basis,S,7,1.7549330000008467,1.767814000002334,0.03712841168356124,0.39142704010009766,"{""cols"": 50.0, ""rows"": 501.0, ""total_mass"": 500.0}"
unit_impulse_basis,M,7,5.230034999996747,5.244222714285992,0.04359156024491504,3.076793670654297,"{""cols"": 100.0, ""rows"": 2001.0, ""total_mass"": 2000.0}"
unit_impulse_basis,L,7,10.598025999996707,10.61451614285959,0.03639945260812847,18.373775482177734,"{""cols"": 200.0, ""rows"": 6001.0, ""total_mass"": 6000.0}"
covariate_resample,S,7,0.3187809999900537,0.31151314285742565,0.027315777992366848,0.06198883056640625,"{""cols"": 1.0, ""rows"": 1001.0, ""signal_energy"": 0.5195204795204795}"
covariate_resample,M,7,0.3733120000077861,0.3833688571432958,0.027679721024068783,0.1355915069580078,"{""cols"": 1.0, ""rows"": 3001.0, ""signal_energy"": 0.5198042747802833}"
covariate_resample,L,7,0.4388640000030364,0.43310742857321266,0.016053381349413916,0.2376575469970703,"{""cols"": 1.0, ""rows"": 6001.0, ""signal_energy"": 0.5199200133311115}"
history_design_matrix,S,7,0.34009099999821046,0.3473752857193598,0.02510644529088935,0.0890655517578125,"{""cols"": 4.0, ""rows"": 1000.0, ""total_count"": 9737.0}"
history_design_matrix,M,7,0.7935309999993478,0.8029685714266829,0.02089668436143026,0.43688201904296875,"{""cols"": 4.0, ""rows"": 5000.0, ""total_count"": 243740.0}"
history_design_matrix,L,7,1.6487759999961327,1.6487702857140059,0.03152991976246771,0.8869400024414062,"{""cols"": 4.0, ""rows"": 10000.0, ""total_count"": 1462420.0}"
simulate_cif_thinning,S,7,11.90902599999788,11.841307428570401,0.2313883554103292,0.07111740112304688,"{""mean_spikes_per_unit"": 13.8, ""num_units"": 5.0, ""total_spikes"": 69.0}"
simulate_cif_thinning,M,7,46.7842029999872,46.623392285716086,0.34392725136599767,0.13444900512695312,"{""mean_spikes_per_unit"": 23.5, ""num_units"": 10.0, ""total_spikes"": 235.0}"
simulate_cif_thinning,L,7,138.86579399999732,138.25536671428398,2.017449271316567,0.20075607299804688,"{""mean_spikes_per_unit"": 36.65, ""num_units"": 20.0, ""total_spikes"": 733.0}"
decoding_spike_rate_cis,S,7,29.966428000008705,29.846530285713666,0.3328735707401364,0.2328357696533203,"{""num_trials"": 6.0, ""prob_mean"": 0.1509259259259259, ""rate_mean"": 50.4457886636761, ""sig_count"": 0.0}"
decoding_spike_rate_cis,M,7,62.59277799999552,63.25780771428567,1.3782261025910456,0.7640476226806641,"{""num_trials"": 8.0, ""prob_mean"": 0.18562499999999998, ""rate_mean"": 50.12398439148756, ""sig_count"": 0.0}"
decoding_spike_rate_cis,L,7,138.46276700000715,138.051728285717,1.750384804578685,2.7336788177490234,"{""num_trials"": 12.0, ""prob_mean"": 0.21328124999999998, ""rate_mean"": 50.073736692667104, ""sig_count"": 0.0}"
unit_impulse_basis,S,7,1.7951030000062929,1.800190142854977,0.030044572115161792,0.39142704010009766,"{""cols"": 50.0, ""rows"": 501.0, ""total_mass"": 500.0}"
unit_impulse_basis,M,7,6.6628950000051645,6.653057285713463,1.4829550166016998,3.076824188232422,"{""cols"": 100.0, ""rows"": 2001.0, ""total_mass"": 2000.0}"
unit_impulse_basis,L,7,16.024357000020473,16.10430300001196,0.23143857554091385,18.37380599975586,"{""cols"": 200.0, ""rows"": 6001.0, ""total_mass"": 6000.0}"
covariate_resample,S,7,0.29961199999206656,0.29913757142756886,0.009541632013909111,0.06201934814453125,"{""cols"": 1.0, ""rows"": 1001.0, ""signal_energy"": 0.5195204795204795}"
covariate_resample,M,7,0.3440330000046288,0.34703042856725524,0.01738929698449399,0.1356220245361328,"{""cols"": 1.0, ""rows"": 3001.0, ""signal_energy"": 0.5198042747802833}"
covariate_resample,L,7,0.4108699999960663,0.4196135714299284,0.017089334564883416,0.2376880645751953,"{""cols"": 1.0, ""rows"": 6001.0, ""signal_energy"": 0.5199200133311115}"
history_design_matrix,S,7,0.3993699999966793,0.4031285714266427,0.013794363505286652,0.0890960693359375,"{""cols"": 4.0, ""rows"": 1000.0, ""total_count"": 9737.0}"
history_design_matrix,M,7,0.9118490000048496,0.9101645714265162,0.05885667560476089,0.43691253662109375,"{""cols"": 4.0, ""rows"": 5000.0, ""total_count"": 243740.0}"
history_design_matrix,L,7,1.8411630000230161,1.8406257142860991,0.014739662895922209,0.8869705200195312,"{""cols"": 4.0, ""rows"": 10000.0, ""total_count"": 1462420.0}"
simulate_cif_thinning,S,7,12.83120299999041,12.864690857132441,0.17008704628575258,0.07111740112304688,"{""mean_spikes_per_unit"": 12.6, ""num_units"": 5.0, ""total_spikes"": 63.0}"
simulate_cif_thinning,M,7,49.6265170000072,49.62809900000106,0.2452487903460643,0.13447189331054688,"{""mean_spikes_per_unit"": 23.9, ""num_units"": 10.0, ""total_spikes"": 239.0}"
simulate_cif_thinning,L,7,148.1780189999995,147.96911385714193,0.8086479240372416,0.20068740844726562,"{""mean_spikes_per_unit"": 35.2, ""num_units"": 20.0, ""total_spikes"": 704.0}"
decoding_spike_rate_cis,S,7,9.922279000022627,9.8990294285792,0.09977266978679186,0.08637428283691406,"{""num_trials"": 6.0, ""prob_mean"": 0.1509259259259259, ""rate_mean"": 50.4457886636761, ""sig_count"": 0.0}"
decoding_spike_rate_cis,M,7,15.846046000007163,15.825482000000843,0.11100674963412104,0.19997024536132812,"{""num_trials"": 8.0, ""prob_mean"": 0.18562499999999998, ""rate_mean"": 50.12398439148756, ""sig_count"": 0.0}"
decoding_spike_rate_cis,L,7,29.582787000009603,29.530547285718608,0.2645464377729889,0.5168647766113281,"{""num_trials"": 12.0, ""prob_mean"": 0.21328124999999998, ""rate_mean"": 50.073736692667104, ""sig_count"": 0.0}"
nspiketrain_get_sigrep,S,7,0.45410599997808276,0.45695271428160467,0.02263874211807069,0.05826568603515625,"{""binary_sum"": 746.0, ""count_sum"": 800.0, ""n_bins"": 1000.0}"
nspiketrain_get_sigrep,M,7,0.5966520000129094,0.6046927142863622,0.017903597884198864,0.18491363525390625,"{""binary_sum"": 2577.0, ""count_sum"": 3000.0, ""n_bins"": 3000.0}"
nspiketrain_get_sigrep,L,7,0.9368639999820516,0.9338544285648303,0.016064032275827016,0.39853668212890625,"{""binary_sum"": 5000.0, ""count_sum"": 9000.0, ""n_bins"": 5000.0}"
analysis_fit_glm_pipeline,S,7,5.966330000006792,6.041203285718666,0.12616461532934142,0.10829353332519531,"{""coeff_norm"": 8.000957080893618e-16, ""intercept"": -12.01100351405273, ""log_likelihood"": -5.469277525370689e-06, ""pred_mean"": 6.076975028189657e-06}"
analysis_fit_glm_pipeline,M,7,7.278939000002538,7.281217000001204,0.1861142338621149,0.22568511962890625,"{""coeff_norm"": 9.782140347964175e-16, ""intercept"": -12.704150666835071, ""log_likelihood"": -5.46927767729412e-06, ""pred_mean"": 3.038487598496732e-06}"
analysis_fit_glm_pipeline,L,7,9.020953999993253,9.130242857143653,0.1717714863551795,0.4345836639404297,"{""coeff_norm"": 8.223875012950201e-16, ""intercept"": -12.704150666835066, ""log_likelihood"": -9.723160315189595e-06, ""pred_mean"": 3.0384875984967483e-06}"
Loading