Skip to content

fix: fix gt4py metrics extractor in the StencilTest benchmarking#1111

Merged
egparedes merged 8 commits intoC2SM:mainfrom
egparedes:fix/fix-bench-infra-for-new-gt4py
Mar 16, 2026
Merged

fix: fix gt4py metrics extractor in the StencilTest benchmarking#1111
egparedes merged 8 commits intoC2SM:mainfrom
egparedes:fix/fix-bench-infra-for-new-gt4py

Conversation

@egparedes
Copy link
Contributor

Get the metrics key using new gt4py hooks.

Copilot AI review requested due to automatic review settings March 13, 2026 15:14
@egparedes egparedes requested a review from iomaganaris March 13, 2026 15:16
Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Updates the StencilTest benchmarking path to extract the GT4Py metrics key via the newer gt4py.next.instrumentation.hooks mechanism, rather than deriving it from compiled-program internals.

Changes:

  • Switch metrics-enabled check to gtx_metrics.is_any_level_enabled() and import gtx_hooks.
  • Add a program-call hook to capture the current metrics source key during execution.
  • Add a pre-run of the configured program intended to “warm up” and set metrics_key for later metrics extraction.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

@egparedes
Copy link
Contributor Author

cscs-ci run benchmark-bencher

@gridtoolsjenkins
Copy link
Collaborator

@gridtoolsjenkins
Copy link
Collaborator

Copy link
Collaborator

@iomaganaris iomaganaris left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. Just one small suggestion. We can also run the bencher CI like before to verify that everything works as expected

@github-actions
Copy link

Mandatory Tests

Please make sure you run these tests via comment before you merge!

  • cscs-ci run default
  • cscs-ci run distributed

Optional Tests

To run benchmarks you can use:

  • cscs-ci run benchmark-bencher

To run tests and benchmarks with the DaCe backend you can use:

  • cscs-ci run dace

To run test levels ignored by the default test suite (mostly simple datatest for static fields computations) you can use:

  • cscs-ci run extra

For more detailed information please look at CI in the EXCLAIM universe.

@egparedes
Copy link
Contributor Author

cscs-ci run benchmark-bencher

@gridtoolsjenkins
Copy link
Collaborator

🐰 Bencher Report

Branchfix/fix-bench-infra-for-new-gt4py
Testbedci-runner:santis-gh200:gtfn_cpu:icon_benchmark_regional

🚨 9 Alerts

BenchmarkMeasure
Units
ViewBenchmark Result
(Result Δ%)
Upper Boundary
(Limit %)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
18.03 ms
(+29.21%)Baseline: 13.96 ms
15.35 ms
(117.46%)

TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
8.88 ms
(+28.43%)Baseline: 6.92 ms
7.61 ms
(116.76%)

TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
1.73 ms
(+39.90%)Baseline: 1.24 ms
1.36 ms
(127.19%)

TestCalculateNabla4[compile_time_domain]Latency
microseconds (µs)
📈 plot
🚷 threshold
🚨 alert (🔔)
1,024.68 µs
(+13.43%)Baseline: 903.38 µs
993.72 µs
(103.12%)

TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
2.89 ms
(+17.17%)Baseline: 2.47 ms
2.72 ms
(106.52%)

TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
12.43 ms
(+29.51%)Baseline: 9.60 ms
10.56 ms
(117.73%)

TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
41.29 ms
(+26.45%)Baseline: 32.65 ms
35.92 ms
(114.95%)

TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
3.17 ms
(+10.01%)Baseline: 2.89 ms
3.17 ms
(100.01%)

TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]Latency
microseconds (µs)
📈 plot
🚷 threshold
🚨 alert (🔔)
828.21 µs
(+11.26%)Baseline: 744.40 µs
818.84 µs
(101.14%)

Click to view all benchmark results
BenchmarkLatencyBenchmark Result
microseconds (µs)
(Result Δ%)
Upper Boundary
microseconds (µs)
(Limit %)
TestApplyDiffusionToVn[compile_time_domain]📈 view plot
🚷 view threshold
1,105.48 µs
(-1.79%)Baseline: 1,125.59 µs
1,238.15 µs
(89.29%)
TestApplyDiffusionToWAndComputeHorizontalGradientsForTurbulence[compile_time_domain]📈 view plot
🚷 view threshold
4,267.40 µs
(-23.08%)Baseline: 5,547.52 µs
6,102.27 µs
(69.93%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
18,032.47 µs
(+29.21%)Baseline: 13,955.79 µs
15,351.37 µs
(117.46%)

TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
8,883.29 µs
(+28.43%)Baseline: 6,916.68 µs
7,608.35 µs
(116.76%)

TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
1,731.74 µs
(+39.90%)Baseline: 1,237.81 µs
1,361.59 µs
(127.19%)

TestCalculateNabla4[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
1,024.68 µs
(+13.43%)Baseline: 903.38 µs
993.72 µs
(103.12%)

TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]📈 view plot
🚷 view threshold
3,000.35 µs
(-46.45%)Baseline: 5,602.43 µs
6,162.67 µs
(48.69%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
2,893.49 µs
(+17.17%)Baseline: 2,469.45 µs
2,716.39 µs
(106.52%)

TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[False]]📈 view plot
🚷 view threshold
45,900.81 µs
(-12.05%)Baseline: 52,190.26 µs
57,409.29 µs
(79.95%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[True]]📈 view plot
🚷 view threshold
46,493.87 µs
(-24.54%)Baseline: 61,614.06 µs
67,775.46 µs
(68.60%)
TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
12,428.50 µs
(+29.51%)Baseline: 9,596.92 µs
10,556.62 µs
(117.73%)

TestComputeHydrostaticCorrectionTerm[compile_time_domain]📈 view plot
🚷 view threshold
826.75 µs
(+1.93%)Baseline: 811.11 µs
892.22 µs
(92.66%)
TestComputePerturbedQuantitiesAndInterpolation[compile_time_domain]📈 view plot
🚷 view threshold
10,578.74 µs
(-11.25%)Baseline: 11,919.48 µs
13,111.42 µs
(80.68%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
41,287.00 µs
(+26.45%)Baseline: 32,651.14 µs
35,916.26 µs
(114.95%)

TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[True]-compile_time_domain]📈 view plot
🚷 view threshold
48,309.86 µs
(-10.38%)Baseline: 53,906.46 µs
59,297.11 µs
(81.47%)
TestFusedVelocityAdvectionStencilVMomentum[compile_time_domain]📈 view plot
🚷 view threshold
26,290.43 µs
(+9.29%)Baseline: 24,054.73 µs
26,460.20 µs
(99.36%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[False]]📈 view plot
🚷 view threshold
7,568.14 µs
(+8.50%)Baseline: 6,975.22 µs
7,672.75 µs
(98.64%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[True]]📈 view plot
🚷 view threshold
8,140.17 µs
(-31.91%)Baseline: 11,955.15 µs
13,150.66 µs
(61.90%)
TestInitCellKdimFieldWithZeroWp[compile_time_domain]📈 view plot
🚷 view threshold
788.67 µs
(-0.73%)Baseline: 794.47 µs
873.91 µs
(90.25%)
TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
3,174.24 µs
(+10.01%)Baseline: 2,885.31 µs
3,173.84 µs
(100.01%)

TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
828.21 µs
(+11.26%)Baseline: 744.40 µs
818.84 µs
(101.14%)

TestUpdateMassFluxWeighted[compile_time_domain]📈 view plot
🚷 view threshold
860.63 µs
(+1.93%)Baseline: 844.36 µs
928.80 µs
(92.66%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
31,342.84 µs
(-10.67%)Baseline: 35,086.65 µs
38,595.31 µs
(81.21%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
30,095.57 µs
(-31.39%)Baseline: 43,865.41 µs
48,251.95 µs
(62.37%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
30,596.38 µs
(-30.13%)Baseline: 43,791.42 µs
48,170.56 µs
(63.52%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[True]__lprep_adv[True]__is_iau_active[True]]📈 view plot
🚷 view threshold
30,451.91 µs
(-34.92%)Baseline: 46,788.80 µs
51,467.68 µs
(59.17%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[False]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
30,015.84 µs
(-14.78%)Baseline: 35,223.41 µs
38,745.75 µs
(77.47%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
29,707.98 µs
(-30.54%)Baseline: 42,771.04 µs
47,048.14 µs
(63.14%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]📈 view plot
🚷 view threshold
31,114.59 µs
(-24.15%)Baseline: 41,022.69 µs
45,124.96 µs
(68.95%)
🐰 View full continuous benchmarking report in Bencher

@gridtoolsjenkins
Copy link
Collaborator

🐰 Bencher Report

Branchfix/fix-bench-infra-for-new-gt4py
Testbedci-runner:santis-gh200:dace_cpu:icon_benchmark_regional

🚨 4 Alerts

BenchmarkMeasure
Units
ViewBenchmark Result
(Result Δ%)
Upper Boundary
(Limit %)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
3.32 ms
(+24.24%)Baseline: 2.67 ms
2.94 ms
(112.94%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
135.24 ms
(+258.27%)Baseline: 37.75 ms
41.52 ms
(325.70%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
136.87 ms
(+255.16%)Baseline: 38.54 ms
42.39 ms
(322.87%)

TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
144.86 ms
(+216.73%)Baseline: 45.74 ms
50.31 ms
(287.93%)

Click to view all benchmark results
BenchmarkLatencyBenchmark Result
microseconds (µs)
(Result Δ%)
Upper Boundary
microseconds (µs)
(Limit %)
TestApplyDiffusionToVn[compile_time_domain]📈 view plot
🚷 view threshold
2,090.30 µs
(-11.71%)Baseline: 2,367.67 µs
2,604.44 µs
(80.26%)
TestApplyDiffusionToWAndComputeHorizontalGradientsForTurbulence[compile_time_domain]📈 view plot
🚷 view threshold
10,493.22 µs
(-48.75%)Baseline: 20,475.28 µs
22,522.81 µs
(46.59%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]📈 view plot
🚷 view threshold
8,574.03 µs
(-90.61%)Baseline: 91,345.88 µs
100,480.46 µs
(8.53%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]📈 view plot
🚷 view threshold
9,423.63 µs
(-97.59%)Baseline: 391,089.25 µs
430,198.18 µs
(2.19%)
TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]📈 view plot
🚷 view threshold
2,339.98 µs
(-6.23%)Baseline: 2,495.45 µs
2,744.99 µs
(85.25%)
TestCalculateNabla4[compile_time_domain]📈 view plot
🚷 view threshold
1,116.95 µs
(-5.95%)Baseline: 1,187.60 µs
1,306.36 µs
(85.50%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
3,319.73 µs
(+24.24%)Baseline: 2,672.04 µs
2,939.25 µs
(112.94%)

TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]📈 view plot
🚷 view threshold
2,971.06 µs
(+9.41%)Baseline: 2,715.54 µs
2,987.10 µs
(99.46%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[False]]📈 view plot
🚷 view threshold
22,759.63 µs
(-48.06%)Baseline: 43,819.68 µs
48,201.65 µs
(47.22%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[True]]📈 view plot
🚷 view threshold
21,064.82 µs
(-50.15%)Baseline: 42,257.17 µs
46,482.89 µs
(45.32%)
TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]📈 view plot
🚷 view threshold
7,866.49 µs
(-5.76%)Baseline: 8,346.87 µs
9,181.56 µs
(85.68%)
TestComputeHydrostaticCorrectionTerm[compile_time_domain]📈 view plot
🚷 view threshold
2,422.33 µs
(-31.56%)Baseline: 3,539.56 µs
3,893.52 µs
(62.21%)
TestComputePerturbedQuantitiesAndInterpolation[compile_time_domain]📈 view plot
🚷 view threshold
10,922.66 µs
(+1.87%)Baseline: 10,722.65 µs
11,794.91 µs
(92.60%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]📈 view plot
🚷 view threshold
26,858.87 µs
(-31.17%)Baseline: 39,021.37 µs
42,923.51 µs
(62.57%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[True]-compile_time_domain]📈 view plot
🚷 view threshold
27,319.13 µs
(-31.56%)Baseline: 39,917.80 µs
43,909.58 µs
(62.22%)
TestFusedVelocityAdvectionStencilVMomentum[compile_time_domain]📈 view plot
🚷 view threshold
12,892.00 µs
(-54.41%)Baseline: 28,275.03 µs
31,102.54 µs
(41.45%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[False]]📈 view plot
🚷 view threshold
4,469.42 µs
(-50.55%)Baseline: 9,037.60 µs
9,941.36 µs
(44.96%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[True]]📈 view plot
🚷 view threshold
8,391.57 µs
(-43.17%)Baseline: 14,766.29 µs
16,242.92 µs
(51.66%)
TestInitCellKdimFieldWithZeroWp[compile_time_domain]📈 view plot
🚷 view threshold
794.88 µs
(-3.33%)Baseline: 822.25 µs
904.47 µs
(87.88%)
TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]📈 view plot
🚷 view threshold
2,960.17 µs
(-4.69%)Baseline: 3,105.92 µs
3,416.52 µs
(86.64%)
TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]📈 view plot
🚷 view threshold
1,195.54 µs
(-3.23%)Baseline: 1,235.47 µs
1,359.02 µs
(87.97%)
TestUpdateMassFluxWeighted[compile_time_domain]📈 view plot
🚷 view threshold
912.04 µs
(+1.69%)Baseline: 896.84 µs
986.53 µs
(92.45%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
135,235.88 µs
(+258.27%)Baseline: 37,747.35 µs
41,522.09 µs
(325.70%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
136,866.42 µs
(+255.16%)Baseline: 38,536.50 µs
42,390.15 µs
(322.87%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
16,580.86 µs
(-60.63%)Baseline: 42,110.58 µs
46,321.64 µs
(35.80%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[True]__lprep_adv[True]__is_iau_active[True]]📈 view plot
🚷 view threshold
17,636.17 µs
(-58.72%)Baseline: 42,727.81 µs
47,000.60 µs
(37.52%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[False]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
21,598.70 µs
(-46.45%)Baseline: 40,336.52 µs
44,370.17 µs
(48.68%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
24,520.51 µs
(-47.55%)Baseline: 46,754.22 µs
51,429.64 µs
(47.68%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
144,864.59 µs
(+216.73%)Baseline: 45,738.15 µs
50,311.96 µs
(287.93%)

🐰 View full continuous benchmarking report in Bencher

@gridtoolsjenkins
Copy link
Collaborator

@gridtoolsjenkins
Copy link
Collaborator

🐰 Bencher Report

Branchfix/fix-bench-infra-for-new-gt4py
Testbedci-runner:santis-gh200:gtfn_gpu:icon_benchmark_regional

🚨 1 Alert

BenchmarkMeasure
Units
ViewBenchmark Result
(Result Δ%)
Upper Boundary
(Limit %)
TestInitCellKdimFieldWithZeroWp[compile_time_domain]Latency
microseconds (µs)
📈 plot
🚷 threshold
🚨 alert (🔔)
25.78 µs
(+19.36%)Baseline: 21.60 µs
23.76 µs
(108.51%)

Click to view all benchmark results
BenchmarkLatencyBenchmark Result
microseconds (µs)
(Result Δ%)
Upper Boundary
microseconds (µs)
(Limit %)
TestApplyDiffusionToVn[compile_time_domain]📈 view plot
🚷 view threshold
105.95 µs
(+2.47%)Baseline: 103.40 µs
113.74 µs
(93.15%)
TestApplyDiffusionToWAndComputeHorizontalGradientsForTurbulence[compile_time_domain]📈 view plot
🚷 view threshold
95.78 µs
(-1.91%)Baseline: 97.65 µs
107.41 µs
(89.17%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]📈 view plot
🚷 view threshold
185.47 µs
(-3.00%)Baseline: 191.20 µs
210.32 µs
(88.18%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]📈 view plot
🚷 view threshold
177.10 µs
(-1.65%)Baseline: 180.06 µs
198.07 µs
(89.41%)
TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]📈 view plot
🚷 view threshold
150.76 µs
(+3.73%)Baseline: 145.34 µs
159.87 µs
(94.30%)
TestCalculateNabla4[compile_time_domain]📈 view plot
🚷 view threshold
88.10 µs
(+3.00%)Baseline: 85.53 µs
94.08 µs
(93.64%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]📈 view plot
🚷 view threshold
183.16 µs
(+2.15%)Baseline: 179.30 µs
197.23 µs
(92.87%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]📈 view plot
🚷 view threshold
150.90 µs
(+0.04%)Baseline: 150.83 µs
165.92 µs
(90.95%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[False]]📈 view plot
🚷 view threshold
526.52 µs
(+2.74%)Baseline: 512.48 µs
563.73 µs
(93.40%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[True]]📈 view plot
🚷 view threshold
462.16 µs
(-1.88%)Baseline: 471.00 µs
518.10 µs
(89.20%)
TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]📈 view plot
🚷 view threshold
336.75 µs
(-1.05%)Baseline: 340.32 µs
374.35 µs
(89.96%)
TestComputeHydrostaticCorrectionTerm[compile_time_domain]📈 view plot
🚷 view threshold
22.55 µs
(+9.58%)Baseline: 20.57 µs
22.63 µs
(99.62%)
TestComputePerturbedQuantitiesAndInterpolation[compile_time_domain]📈 view plot
🚷 view threshold
298.04 µs
(-2.60%)Baseline: 305.99 µs
336.59 µs
(88.55%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]📈 view plot
🚷 view threshold
990.37 µs
(-2.26%)Baseline: 1,013.24 µs
1,114.57 µs
(88.86%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[True]-compile_time_domain]📈 view plot
🚷 view threshold
998.63 µs
(-1.97%)Baseline: 1,018.65 µs
1,120.52 µs
(89.12%)
TestFusedVelocityAdvectionStencilVMomentum[compile_time_domain]📈 view plot
🚷 view threshold
352.88 µs
(-1.74%)Baseline: 359.14 µs
395.05 µs
(89.33%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[False]]📈 view plot
🚷 view threshold
162.77 µs
(-2.46%)Baseline: 166.87 µs
183.56 µs
(88.68%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[True]]📈 view plot
🚷 view threshold
236.64 µs
(-1.64%)Baseline: 240.58 µs
264.64 µs
(89.42%)
TestInitCellKdimFieldWithZeroWp[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
25.78 µs
(+19.36%)Baseline: 21.60 µs
23.76 µs
(108.51%)

TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]📈 view plot
🚷 view threshold
138.06 µs
(+1.06%)Baseline: 136.61 µs
150.27 µs
(91.88%)
TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]📈 view plot
🚷 view threshold
51.08 µs
(+4.52%)Baseline: 48.87 µs
53.76 µs
(95.02%)
TestUpdateMassFluxWeighted[compile_time_domain]📈 view plot
🚷 view threshold
60.34 µs
(+0.57%)Baseline: 60.00 µs
66.00 µs
(91.43%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
601.78 µs
(-1.37%)Baseline: 610.11 µs
671.12 µs
(89.67%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
605.90 µs
(-0.81%)Baseline: 610.83 µs
671.92 µs
(90.17%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
590.52 µs
(-1.10%)Baseline: 597.08 µs
656.78 µs
(89.91%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[True]__lprep_adv[True]__is_iau_active[True]]📈 view plot
🚷 view threshold
602.99 µs
(-1.10%)Baseline: 609.70 µs
670.67 µs
(89.91%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[False]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
614.76 µs
(-0.43%)Baseline: 617.44 µs
679.19 µs
(90.51%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
582.76 µs
(-0.46%)Baseline: 585.46 µs
644.00 µs
(90.49%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]📈 view plot
🚷 view threshold
597.25 µs
(-0.05%)Baseline: 597.55 µs
657.31 µs
(90.86%)
🐰 View full continuous benchmarking report in Bencher

@gridtoolsjenkins
Copy link
Collaborator

🐰 Bencher Report

Branchfix/fix-bench-infra-for-new-gt4py
Testbedci-runner:santis-gh200:dace_gpu:icon_benchmark_regional
Click to view all benchmark results
BenchmarkLatencyBenchmark Result
microseconds (µs)
(Result Δ%)
Upper Boundary
microseconds (µs)
(Limit %)
TestApplyDiffusionToVn[compile_time_domain]📈 view plot
🚷 view threshold
101.15 µs
(+0.15%)Baseline: 101.00 µs
111.10 µs
(91.04%)
TestApplyDiffusionToWAndComputeHorizontalGradientsForTurbulence[compile_time_domain]📈 view plot
🚷 view threshold
129.68 µs
(+1.46%)Baseline: 127.82 µs
140.61 µs
(92.23%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]📈 view plot
🚷 view threshold
209.44 µs
(-2.65%)Baseline: 215.15 µs
236.66 µs
(88.50%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]📈 view plot
🚷 view threshold
259.25 µs
(-2.21%)Baseline: 265.11 µs
291.62 µs
(88.90%)
TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]📈 view plot
🚷 view threshold
122.51 µs
(-1.58%)Baseline: 124.48 µs
136.93 µs
(89.47%)
TestCalculateNabla4[compile_time_domain]📈 view plot
🚷 view threshold
87.43 µs
(-0.06%)Baseline: 87.49 µs
96.24 µs
(90.85%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]📈 view plot
🚷 view threshold
195.78 µs
(-3.43%)Baseline: 202.74 µs
223.02 µs
(87.79%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]📈 view plot
🚷 view threshold
143.87 µs
(-5.32%)Baseline: 151.95 µs
167.15 µs
(86.08%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[False]]📈 view plot
🚷 view threshold
443.00 µs
(-1.60%)Baseline: 450.19 µs
495.21 µs
(89.46%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[True]]📈 view plot
🚷 view threshold
396.62 µs
(-3.10%)Baseline: 409.32 µs
450.25 µs
(88.09%)
TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]📈 view plot
🚷 view threshold
304.50 µs
(-5.09%)Baseline: 320.84 µs
352.93 µs
(86.28%)
TestComputeHydrostaticCorrectionTerm[compile_time_domain]📈 view plot
🚷 view threshold
16.02 µs
(-11.11%)Baseline: 18.03 µs
19.83 µs
(80.81%)
TestComputePerturbedQuantitiesAndInterpolation[compile_time_domain]📈 view plot
🚷 view threshold
268.29 µs
(-5.02%)Baseline: 282.47 µs
310.72 µs
(86.35%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]📈 view plot
🚷 view threshold
715.72 µs
(-1.25%)Baseline: 724.79 µs
797.27 µs
(89.77%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[True]-compile_time_domain]📈 view plot
🚷 view threshold
730.24 µs
(-1.04%)Baseline: 737.91 µs
811.70 µs
(89.96%)
TestFusedVelocityAdvectionStencilVMomentum[compile_time_domain]📈 view plot
🚷 view threshold
242.41 µs
(-1.77%)Baseline: 246.77 µs
271.45 µs
(89.30%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[False]]📈 view plot
🚷 view threshold
159.86 µs
(-2.75%)Baseline: 164.37 µs
180.80 µs
(88.41%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[True]]📈 view plot
🚷 view threshold
217.20 µs
(-2.06%)Baseline: 221.77 µs
243.95 µs
(89.04%)
TestInitCellKdimFieldWithZeroWp[compile_time_domain]📈 view plot
🚷 view threshold
19.99 µs
(+0.32%)Baseline: 19.92 µs
21.91 µs
(91.20%)
TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]📈 view plot
🚷 view threshold
125.45 µs
(-5.74%)Baseline: 133.08 µs
146.39 µs
(85.69%)
TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]📈 view plot
🚷 view threshold
43.35 µs
(-0.99%)Baseline: 43.79 µs
48.17 µs
(90.01%)
TestUpdateMassFluxWeighted[compile_time_domain]📈 view plot
🚷 view threshold
56.00 µs
(-4.71%)Baseline: 58.77 µs
64.64 µs
(86.63%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
536.57 µs
(-1.66%)Baseline: 545.63 µs
600.19 µs
(89.40%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
562.73 µs
(-1.53%)Baseline: 571.46 µs
628.60 µs
(89.52%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
551.38 µs
(-1.81%)Baseline: 561.56 µs
617.72 µs
(89.26%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[True]__lprep_adv[True]__is_iau_active[True]]📈 view plot
🚷 view threshold
586.95 µs
(-2.26%)Baseline: 600.53 µs
660.58 µs
(88.85%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[False]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
537.21 µs
(-0.42%)Baseline: 539.49 µs
593.44 µs
(90.52%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
581.56 µs
(+0.68%)Baseline: 577.61 µs
635.38 µs
(91.53%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]📈 view plot
🚷 view threshold
596.56 µs
(-0.76%)Baseline: 601.11 µs
661.22 µs
(90.22%)
🐰 View full continuous benchmarking report in Bencher

@gridtoolsjenkins
Copy link
Collaborator

🐰 Bencher Report

Branchfix/fix-bench-infra-for-new-gt4py
Testbedci-runner:santis-gh200:gtfn_cpu:icon_benchmark_global

🚨 19 Alerts

BenchmarkMeasure
Units
ViewBenchmark Result
(Result Δ%)
Upper Boundary
(Limit %)
TestApplyDiffusionToWAndComputeHorizontalGradientsForTurbulence[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
37.33 ms
(+16.07%)Baseline: 32.16 ms
35.37 ms
(105.52%)

TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
120.33 ms
(+14.38%)Baseline: 105.21 ms
115.73 ms
(103.98%)

TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
57.25 ms
(+13.93%)Baseline: 50.25 ms
55.27 ms
(103.57%)

TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
20.16 ms
(+46.79%)Baseline: 13.74 ms
15.11 ms
(133.45%)

TestCalculateNabla4[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
9.44 ms
(+33.71%)Baseline: 7.06 ms
7.77 ms
(121.55%)

TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
23.34 ms
(+35.11%)Baseline: 17.27 ms
19.00 ms
(122.83%)

TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[False]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
338.92 ms
(+12.28%)Baseline: 301.86 ms
332.05 ms
(102.07%)

TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
84.57 ms
(+14.19%)Baseline: 74.06 ms
81.47 ms
(103.81%)

TestComputeHydrostaticCorrectionTerm[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
2.48 ms
(+20.60%)Baseline: 2.05 ms
2.26 ms
(109.64%)

TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
258.49 ms
(+15.71%)Baseline: 223.40 ms
245.74 ms
(105.19%)

TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[True]-compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
264.70 ms
(+14.17%)Baseline: 231.85 ms
255.03 ms
(103.79%)

TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[False]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
55.70 ms
(+10.09%)Baseline: 50.59 ms
55.65 ms
(100.09%)

TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[True]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
64.74 ms
(+13.16%)Baseline: 57.21 ms
62.93 ms
(102.87%)

TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
4.88 ms
(+45.86%)Baseline: 3.34 ms
3.68 ms
(132.60%)

TestUpdateMassFluxWeighted[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
6.81 ms
(+22.90%)Baseline: 5.54 ms
6.09 ms
(111.73%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
243.58 ms
(+13.64%)Baseline: 214.33 ms
235.77 ms
(103.31%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
242.64 ms
(+11.14%)Baseline: 218.32 ms
240.16 ms
(101.03%)

TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[False]__divdamp_type[32]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
250.53 ms
(+13.63%)Baseline: 220.47 ms
242.52 ms
(103.30%)

TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
249.73 ms
(+11.09%)Baseline: 224.80 ms
247.28 ms
(100.99%)

Click to view all benchmark results
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)
Upper Boundary
milliseconds (ms)
(Limit %)
TestApplyDiffusionToVn[compile_time_domain]📈 view plot
🚷 view threshold
14.54 ms
(+3.78%)Baseline: 14.01 ms
15.41 ms
(94.35%)
TestApplyDiffusionToWAndComputeHorizontalGradientsForTurbulence[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
37.33 ms
(+16.07%)Baseline: 32.16 ms
35.37 ms
(105.52%)

TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
120.33 ms
(+14.38%)Baseline: 105.21 ms
115.73 ms
(103.98%)

TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
57.25 ms
(+13.93%)Baseline: 50.25 ms
55.27 ms
(103.57%)

TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
20.16 ms
(+46.79%)Baseline: 13.74 ms
15.11 ms
(133.45%)

TestCalculateNabla4[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
9.44 ms
(+33.71%)Baseline: 7.06 ms
7.77 ms
(121.55%)

TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]📈 view plot
🚷 view threshold
22.93 ms
(+9.72%)Baseline: 20.90 ms
22.99 ms
(99.75%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
23.34 ms
(+35.11%)Baseline: 17.27 ms
19.00 ms
(122.83%)

TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[False]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
338.92 ms
(+12.28%)Baseline: 301.86 ms
332.05 ms
(102.07%)

TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[True]]📈 view plot
🚷 view threshold
326.64 ms
(+9.06%)Baseline: 299.52 ms
329.47 ms
(99.14%)
TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
84.57 ms
(+14.19%)Baseline: 74.06 ms
81.47 ms
(103.81%)

TestComputeHydrostaticCorrectionTerm[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
2.48 ms
(+20.60%)Baseline: 2.05 ms
2.26 ms
(109.64%)

TestComputePerturbedQuantitiesAndInterpolation[compile_time_domain]📈 view plot
🚷 view threshold
62.46 ms
(+4.32%)Baseline: 59.87 ms
65.86 ms
(94.83%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
258.49 ms
(+15.71%)Baseline: 223.40 ms
245.74 ms
(105.19%)

TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[True]-compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
264.70 ms
(+14.17%)Baseline: 231.85 ms
255.03 ms
(103.79%)

TestFusedVelocityAdvectionStencilVMomentum[compile_time_domain]📈 view plot
🚷 view threshold
173.25 ms
(+7.82%)Baseline: 160.68 ms
176.75 ms
(98.02%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[False]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
55.70 ms
(+10.09%)Baseline: 50.59 ms
55.65 ms
(100.09%)

TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[True]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
64.74 ms
(+13.16%)Baseline: 57.21 ms
62.93 ms
(102.87%)

TestInitCellKdimFieldWithZeroWp[compile_time_domain]📈 view plot
🚷 view threshold
1.01 ms
(+7.85%)Baseline: 0.93 ms
1.03 ms
(98.05%)
TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]📈 view plot
🚷 view threshold
24.00 ms
(+7.92%)Baseline: 22.23 ms
24.46 ms
(98.11%)
TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
4.88 ms
(+45.86%)Baseline: 3.34 ms
3.68 ms
(132.60%)

TestUpdateMassFluxWeighted[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
6.81 ms
(+22.90%)Baseline: 5.54 ms
6.09 ms
(111.73%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
243.58 ms
(+13.64%)Baseline: 214.33 ms
235.77 ms
(103.31%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
242.64 ms
(+11.14%)Baseline: 218.32 ms
240.16 ms
(101.03%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
236.05 ms
(+8.62%)Baseline: 217.32 ms
239.06 ms
(98.74%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[True]__lprep_adv[True]__is_iau_active[True]]📈 view plot
🚷 view threshold
238.22 ms
(+7.39%)Baseline: 221.82 ms
244.00 ms
(97.63%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[False]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
248.08 ms
(+7.98%)Baseline: 229.75 ms
252.73 ms
(98.16%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
250.53 ms
(+13.63%)Baseline: 220.47 ms
242.52 ms
(103.30%)

TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
249.73 ms
(+11.09%)Baseline: 224.80 ms
247.28 ms
(100.99%)

🐰 View full continuous benchmarking report in Bencher

@gridtoolsjenkins
Copy link
Collaborator

🐰 Bencher Report

Branchfix/fix-bench-infra-for-new-gt4py
Testbedci-runner:santis-gh200:dace_cpu:icon_benchmark_global

🚨 8 Alerts

BenchmarkMeasure
Units
ViewBenchmark Result
(Result Δ%)
Upper Boundary
(Limit %)
TestCalculateNabla4[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
7.10 ms
(+38.55%)Baseline: 5.12 ms
5.64 ms
(125.96%)

TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
35.10 ms
(+78.53%)Baseline: 19.66 ms
21.63 ms
(162.30%)

TestComputeHydrostaticCorrectionTerm[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
53.01 ms
(+124.44%)Baseline: 23.62 ms
25.98 ms
(204.04%)

TestComputePerturbedQuantitiesAndInterpolation[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
89.90 ms
(+67.92%)Baseline: 53.54 ms
58.89 ms
(152.65%)

TestInitCellKdimFieldWithZeroWp[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
1.14 ms
(+11.31%)Baseline: 1.02 ms
1.12 ms
(101.19%)

TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
28.91 ms
(+37.25%)Baseline: 21.06 ms
23.17 ms
(124.77%)

TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
6.53 ms
(+14.29%)Baseline: 5.71 ms
6.29 ms
(103.90%)

TestUpdateMassFluxWeighted[compile_time_domain]Latency
milliseconds (ms)
📈 plot
🚷 threshold
🚨 alert (🔔)
14.65 ms
(+126.46%)Baseline: 6.47 ms
7.12 ms
(205.87%)

Click to view all benchmark results
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)
Upper Boundary
milliseconds (ms)
(Limit %)
TestApplyDiffusionToVn[compile_time_domain]📈 view plot
🚷 view threshold
6.70 ms
(-21.90%)Baseline: 8.58 ms
9.43 ms
(71.00%)
TestApplyDiffusionToWAndComputeHorizontalGradientsForTurbulence[compile_time_domain]📈 view plot
🚷 view threshold
25.22 ms
(-94.25%)Baseline: 438.70 ms
482.57 ms
(5.23%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]📈 view plot
🚷 view threshold
73.31 ms
(-97.58%)Baseline: 3,023.69 ms
3,326.06 ms
(2.20%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]📈 view plot
🚷 view threshold
40.19 ms
(-98.98%)Baseline: 3,928.51 ms
4,321.36 ms
(0.93%)
TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]📈 view plot
🚷 view threshold
14.12 ms
(-1.85%)Baseline: 14.38 ms
15.82 ms
(89.23%)
TestCalculateNabla4[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
7.10 ms
(+38.55%)Baseline: 5.12 ms
5.64 ms
(125.96%)

TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
35.10 ms
(+78.53%)Baseline: 19.66 ms
21.63 ms
(162.30%)

TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]📈 view plot
🚷 view threshold
15.24 ms
(-25.25%)Baseline: 20.39 ms
22.42 ms
(67.95%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[False]]📈 view plot
🚷 view threshold
146.42 ms
(-82.67%)Baseline: 844.94 ms
929.43 ms
(15.75%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[True]]📈 view plot
🚷 view threshold
204.52 ms
(-74.08%)Baseline: 789.08 ms
867.99 ms
(23.56%)
TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]📈 view plot
🚷 view threshold
33.40 ms
(-22.64%)Baseline: 43.18 ms
47.49 ms
(70.32%)
TestComputeHydrostaticCorrectionTerm[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
53.01 ms
(+124.44%)Baseline: 23.62 ms
25.98 ms
(204.04%)

TestComputePerturbedQuantitiesAndInterpolation[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
89.90 ms
(+67.92%)Baseline: 53.54 ms
58.89 ms
(152.65%)

TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]📈 view plot
🚷 view threshold
136.29 ms
(-93.18%)Baseline: 1,998.43 ms
2,198.27 ms
(6.20%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[True]-compile_time_domain]📈 view plot
🚷 view threshold
118.78 ms
(-94.53%)Baseline: 2,170.11 ms
2,387.12 ms
(4.98%)
TestFusedVelocityAdvectionStencilVMomentum[compile_time_domain]📈 view plot
🚷 view threshold
47.15 ms
(-97.79%)Baseline: 2,131.21 ms
2,344.33 ms
(2.01%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[False]]📈 view plot
🚷 view threshold
25.39 ms
(-98.33%)Baseline: 1,523.89 ms
1,676.28 ms
(1.51%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[True]]📈 view plot
🚷 view threshold
42.54 ms
(-97.28%)Baseline: 1,565.24 ms
1,721.76 ms
(2.47%)
TestInitCellKdimFieldWithZeroWp[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
1.14 ms
(+11.31%)Baseline: 1.02 ms
1.12 ms
(101.19%)

TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
28.91 ms
(+37.25%)Baseline: 21.06 ms
23.17 ms
(124.77%)

TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
6.53 ms
(+14.29%)Baseline: 5.71 ms
6.29 ms
(103.90%)

TestUpdateMassFluxWeighted[compile_time_domain]📈 view plot
🚷 view threshold
🚨 view alert (🔔)
14.65 ms
(+126.46%)Baseline: 6.47 ms
7.12 ms
(205.87%)

TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
168.29 ms
(-93.74%)Baseline: 2,686.80 ms
2,955.48 ms
(5.69%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
174.16 ms
(-92.09%)Baseline: 2,201.81 ms
2,421.99 ms
(7.19%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
207.44 ms
(-94.97%)Baseline: 4,125.90 ms
4,538.49 ms
(4.57%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[True]__lprep_adv[True]__is_iau_active[True]]📈 view plot
🚷 view threshold
209.47 ms
(-88.55%)Baseline: 1,829.96 ms
2,012.95 ms
(10.41%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[False]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
222.53 ms
(-88.30%)Baseline: 1,901.49 ms
2,091.64 ms
(10.64%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
497.75 ms
(-85.91%)Baseline: 3,532.44 ms
3,885.69 ms
(12.81%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]📈 view plot
🚷 view threshold
368.41 ms
(-84.20%)Baseline: 2,332.03 ms
2,565.23 ms
(14.36%)
🐰 View full continuous benchmarking report in Bencher

@gridtoolsjenkins
Copy link
Collaborator

@gridtoolsjenkins
Copy link
Collaborator

🐰 Bencher Report

Branchfix/fix-bench-infra-for-new-gt4py
Testbedci-runner:santis-gh200:gtfn_gpu:icon_benchmark_global
Click to view all benchmark results
BenchmarkLatencyBenchmark Result
microseconds (µs)
(Result Δ%)
Upper Boundary
microseconds (µs)
(Limit %)
TestApplyDiffusionToVn[compile_time_domain]📈 view plot
🚷 view threshold
685.21 µs
(+0.33%)Baseline: 682.98 µs
751.28 µs
(91.21%)
TestApplyDiffusionToWAndComputeHorizontalGradientsForTurbulence[compile_time_domain]📈 view plot
🚷 view threshold
828.51 µs
(-0.67%)Baseline: 834.07 µs
917.48 µs
(90.30%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]📈 view plot
🚷 view threshold
1,368.50 µs
(-1.56%)Baseline: 1,390.16 µs
1,529.18 µs
(89.49%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]📈 view plot
🚷 view threshold
1,270.69 µs
(+0.26%)Baseline: 1,267.42 µs
1,394.16 µs
(91.14%)
TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]📈 view plot
🚷 view threshold
1,037.74 µs
(+0.11%)Baseline: 1,036.61 µs
1,140.27 µs
(91.01%)
TestCalculateNabla4[compile_time_domain]📈 view plot
🚷 view threshold
508.23 µs
(-1.50%)Baseline: 515.96 µs
567.55 µs
(89.55%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]📈 view plot
🚷 view threshold
1,080.98 µs
(-2.18%)Baseline: 1,105.06 µs
1,215.57 µs
(88.93%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]📈 view plot
🚷 view threshold
929.44 µs
(-1.85%)Baseline: 946.97 µs
1,041.67 µs
(89.23%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[False]]📈 view plot
🚷 view threshold
3,659.75 µs
(+6.61%)Baseline: 3,432.76 µs
3,776.03 µs
(96.92%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[True]]📈 view plot
🚷 view threshold
3,070.26 µs
(-0.64%)Baseline: 3,089.99 µs
3,398.99 µs
(90.33%)
TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]📈 view plot
🚷 view threshold
2,188.26 µs
(-0.97%)Baseline: 2,209.69 µs
2,430.66 µs
(90.03%)
TestComputeHydrostaticCorrectionTerm[compile_time_domain]📈 view plot
🚷 view threshold
100.62 µs
(-12.13%)Baseline: 114.50 µs
125.95 µs
(79.89%)
TestComputePerturbedQuantitiesAndInterpolation[compile_time_domain]📈 view plot
🚷 view threshold
2,012.21 µs
(-1.00%)Baseline: 2,032.48 µs
2,235.73 µs
(90.00%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]📈 view plot
🚷 view threshold
4,231.40 µs
(-2.61%)Baseline: 4,344.69 µs
4,779.15 µs
(88.54%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[True]-compile_time_domain]📈 view plot
🚷 view threshold
4,296.63 µs
(-2.29%)Baseline: 4,397.41 µs
4,837.15 µs
(88.83%)
TestFusedVelocityAdvectionStencilVMomentum[compile_time_domain]📈 view plot
🚷 view threshold
2,558.28 µs
(-0.64%)Baseline: 2,574.84 µs
2,832.32 µs
(90.32%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[False]]📈 view plot
🚷 view threshold
1,300.16 µs
(-1.39%)Baseline: 1,318.44 µs
1,450.29 µs
(89.65%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[True]]📈 view plot
🚷 view threshold
1,787.00 µs
(-1.37%)Baseline: 1,811.75 µs
1,992.92 µs
(89.67%)
TestInitCellKdimFieldWithZeroWp[compile_time_domain]📈 view plot
🚷 view threshold
72.26 µs
(-3.37%)Baseline: 74.78 µs
82.26 µs
(87.84%)
TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]📈 view plot
🚷 view threshold
926.05 µs
(-1.68%)Baseline: 941.84 µs
1,036.03 µs
(89.39%)
TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]📈 view plot
🚷 view threshold
514.09 µs
(-0.67%)Baseline: 517.56 µs
569.32 µs
(90.30%)
TestUpdateMassFluxWeighted[compile_time_domain]📈 view plot
🚷 view threshold
363.35 µs
(-2.94%)Baseline: 374.34 µs
411.77 µs
(88.24%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
3,649.99 µs
(-0.47%)Baseline: 3,667.26 µs
4,033.98 µs
(90.48%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
3,696.12 µs
(-0.63%)Baseline: 3,719.53 µs
4,091.48 µs
(90.34%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
3,549.20 µs
(-0.74%)Baseline: 3,575.50 µs
3,933.04 µs
(90.24%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[True]__lprep_adv[True]__is_iau_active[True]]📈 view plot
🚷 view threshold
3,698.74 µs
(-0.63%)Baseline: 3,722.12 µs
4,094.33 µs
(90.34%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[False]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
3,853.09 µs
(-0.53%)Baseline: 3,873.75 µs
4,261.13 µs
(90.42%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
3,710.80 µs
(-0.55%)Baseline: 3,731.18 µs
4,104.29 µs
(90.41%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]📈 view plot
🚷 view threshold
3,821.86 µs
(-0.64%)Baseline: 3,846.35 µs
4,230.98 µs
(90.33%)
🐰 View full continuous benchmarking report in Bencher

@gridtoolsjenkins
Copy link
Collaborator

🐰 Bencher Report

Branchfix/fix-bench-infra-for-new-gt4py
Testbedci-runner:santis-gh200:dace_gpu:icon_benchmark_global
Click to view all benchmark results
BenchmarkLatencyBenchmark Result
microseconds (µs)
(Result Δ%)
Upper Boundary
microseconds (µs)
(Limit %)
TestApplyDiffusionToVn[compile_time_domain]📈 view plot
🚷 view threshold
677.73 µs
(+0.03%)Baseline: 677.52 µs
745.27 µs
(90.94%)
TestApplyDiffusionToWAndComputeHorizontalGradientsForTurbulence[compile_time_domain]📈 view plot
🚷 view threshold
733.62 µs
(-0.79%)Baseline: 739.44 µs
813.38 µs
(90.19%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[False]__second_order_divdamp_factor[0.032]]📈 view plot
🚷 view threshold
1,490.34 µs
(-0.83%)Baseline: 1,502.76 µs
1,653.04 µs
(90.16%)
TestApplyDivergenceDampingAndUpdateVn[compile_time_domain-divdamp_order[24]__is_iau_active[True]__second_order_divdamp_factor[0.012]]📈 view plot
🚷 view threshold
1,817.55 µs
(-0.99%)Baseline: 1,835.72 µs
2,019.29 µs
(90.01%)
TestCalculateNabla2AndSmagCoefficientsForVn[compile_time_domain]📈 view plot
🚷 view threshold
886.00 µs
(-0.55%)Baseline: 890.92 µs
980.02 µs
(90.41%)
TestCalculateNabla4[compile_time_domain]📈 view plot
🚷 view threshold
568.98 µs
(-0.90%)Baseline: 574.16 µs
631.58 µs
(90.09%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[False]]📈 view plot
🚷 view threshold
1,161.67 µs
(-1.70%)Baseline: 1,181.74 µs
1,299.91 µs
(89.37%)
TestComputeAveragedVnAndFluxesAndPrepareTracerAdvection[compile_time_domain-prepare_advection[True]__at_first_substep[True]]📈 view plot
🚷 view threshold
934.82 µs
(-1.21%)Baseline: 946.32 µs
1,040.95 µs
(89.81%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[False]]📈 view plot
🚷 view threshold
2,984.77 µs
(+3.44%)Baseline: 2,885.65 µs
3,174.21 µs
(94.03%)
TestComputeDerivedHorizontalWindsAndKEAndHorizontalAdvectionofWAndContravariantCorrection[compile_time_domain-skip_compute_predictor_vertical_advection[True]]📈 view plot
🚷 view threshold
2,674.53 µs
(-0.54%)Baseline: 2,689.05 µs
2,957.95 µs
(90.42%)
TestComputeHorizontalVelocityQuantitiesAndFluxes[compile_time_domain]📈 view plot
🚷 view threshold
2,072.81 µs
(-1.45%)Baseline: 2,103.28 µs
2,313.61 µs
(89.59%)
TestComputeHydrostaticCorrectionTerm[compile_time_domain]📈 view plot
🚷 view threshold
102.02 µs
(-6.50%)Baseline: 109.11 µs
120.02 µs
(85.00%)
TestComputePerturbedQuantitiesAndInterpolation[compile_time_domain]📈 view plot
🚷 view threshold
1,847.98 µs
(-0.76%)Baseline: 1,862.19 µs
2,048.41 µs
(90.22%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[False]-compile_time_domain]📈 view plot
🚷 view threshold
3,486.54 µs
(+0.23%)Baseline: 3,478.59 µs
3,826.45 µs
(91.12%)
TestComputeThetaRhoPressureGradientAndUpdateVn[is_iau_active[True]-compile_time_domain]📈 view plot
🚷 view threshold
3,566.02 µs
(+0.41%)Baseline: 3,551.60 µs
3,906.76 µs
(91.28%)
TestFusedVelocityAdvectionStencilVMomentum[compile_time_domain]📈 view plot
🚷 view threshold
1,551.01 µs
(+1.19%)Baseline: 1,532.81 µs
1,686.09 µs
(91.99%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[False]]📈 view plot
🚷 view threshold
1,090.37 µs
(-1.01%)Baseline: 1,101.49 µs
1,211.64 µs
(89.99%)
TestFusedVelocityAdvectionStencilsHMomentum[compile_time_domain-apply_extra_diffusion_on_vn[True]]📈 view plot
🚷 view threshold
1,462.58 µs
(-0.48%)Baseline: 1,469.59 µs
1,616.54 µs
(90.48%)
TestInitCellKdimFieldWithZeroWp[compile_time_domain]📈 view plot
🚷 view threshold
71.18 µs
(-2.96%)Baseline: 73.35 µs
80.69 µs
(88.22%)
TestInterpolateRhoThetaVToHalfLevelsAndComputePressureBuoyancyAcceleration[compile_time_domain]📈 view plot
🚷 view threshold
896.47 µs
(-1.29%)Baseline: 908.16 µs
998.98 µs
(89.74%)
TestMoIntpRbfRbfVecInterpolVertex[compile_time_domain]📈 view plot
🚷 view threshold
302.41 µs
(+0.32%)Baseline: 301.46 µs
331.61 µs
(91.20%)
TestUpdateMassFluxWeighted[compile_time_domain]📈 view plot
🚷 view threshold
357.94 µs
(-2.96%)Baseline: 368.85 µs
405.74 µs
(88.22%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
3,789.40 µs
(+1.73%)Baseline: 3,724.90 µs
4,097.39 µs
(92.48%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[False]__at_last_substep[True]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
3,928.90 µs
(+0.11%)Baseline: 3,924.73 µs
4,317.20 µs
(91.01%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[False]__lprep_adv[True]__is_iau_active[False]]📈 view plot
🚷 view threshold
3,852.69 µs
(+0.16%)Baseline: 3,846.58 µs
4,231.23 µs
(91.05%)
TestVerticallyImplicitSolverAtCorrectorStep[compile_time_domain-at_first_substep[True]__at_last_substep[True]__lprep_adv[True]__is_iau_active[True]]📈 view plot
🚷 view threshold
4,116.18 µs
(-0.27%)Baseline: 4,127.17 µs
4,539.89 µs
(90.67%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[False]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
3,793.71 µs
(-0.07%)Baseline: 3,796.48 µs
4,176.12 µs
(90.84%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[False]__divdamp_type[32]]📈 view plot
🚷 view threshold
4,038.67 µs
(-0.18%)Baseline: 4,046.04 µs
4,450.64 µs
(90.74%)
TestVerticallyImplicitSolverAtPredictorStep[compile_time_domain-at_first_substep[True]__is_iau_active[True]__divdamp_type[3]]📈 view plot
🚷 view threshold
4,146.85 µs
(-0.14%)Baseline: 4,152.72 µs
4,567.99 µs
(90.78%)
🐰 View full continuous benchmarking report in Bencher

@iomaganaris
Copy link
Collaborator

cscs-ci run default

@iomaganaris
Copy link
Collaborator

cscs-ci run distributed

Copy link
Collaborator

@iomaganaris iomaganaris left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks a lot for the fixes 👍

1 if skip_stenciltest_verification else 2
2 if skip_stenciltest_verification else 3
)
assert (
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not really an issue, but remember that we run the benchmark CI pipeline with PYTHONOPTIMIZE=2 so the asserts are disabled.

@egparedes egparedes merged commit 0918c3d into C2SM:main Mar 16, 2026
54 of 55 checks passed
jcanton added a commit that referenced this pull request Mar 18, 2026
* main: (29 commits)
  Scheduled Halo Exchange (#980)
  Add missing metrics fields to `test_parallel_grid_manager.py` test (#1114)
  Muphys: Lowering with single precision (#1101)
  Add single-rank lsq pseudoinv factory test (#1099)
  Cleanup Diffusion config (#1060)
  Fortran bindings: fix numpy allocation and cleanups (#1112)
  fix: fix gt4py metrics extractor in the StencilTest benchmarking (#1111)
  py2fgen: don't recompile if unchanged (#1110)
  CI for standalone_driver (#1070)
  Update mpi4py and pymetis groups to make them optional (#1100)
  Bump mshick/add-pr-comment from 2 to 3 (#1109)
  Use inout fields for full_muphys as well (#1108)
  Update GPU configuration for graupel (#1104)
  Move the mask of _q_t_update outside in graupel (#1093)
  Update gt4py to v1.1.7 (#1105)
  cleanup for ugly if condition of single node default in lsq coeffs (#1103)
  Domain decomposition and halo construction (#540)
  Muphys: Add flag to wait for graupel completion (#1095)
  Give each gt4py program a return type hint (#1087)
  Turn data download off for distributed CI (#1092)
  ...
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants