This is a migration of volesti#423.
Problem Statement: As discussed in volesti issue #423, there is currently no automated way to quantify the performance overhead or convergence differences (ESS, Mixing Time) introduced by the dingo Python interface compared to pure volesti C++ executions.
Goal: Establish a repeatable benchmarking suite within dingo to ensure that the Python wrapper remains performant and numerically consistent with the C++ core.
Proposed Implementation:
• Benchmark Script: A new module (e.g., tests/benchmarks/) that:
1.Generates standard high-dimensional test geometries (unit cube, simplex).
2.Executes the same sampler (e.g., Coordinate Hit-and-Run) through both the dingo interface and a direct volesti call (if possible via the existing bindings).
3.Calculates and reports:
• Time per sample (Wall-clock time).
• Effective Sample Size (ESS) comparison.
• Variance/Numerical stability metrics.
##
• CI Integration: (Optional) Explore adding a "Performance regression" check to the CI pipeline to catch future overhead spikes.
I am a prospective GSoC 2026 contributor. This task will help me understand the bridge between dingo's Pythonic layer and the high-performance C++ backend.
This is a migration of volesti#423.
Problem Statement: As discussed in volesti issue #423, there is currently no automated way to quantify the performance overhead or convergence differences (ESS, Mixing Time) introduced by the dingo Python interface compared to pure volesti C++ executions.
Goal: Establish a repeatable benchmarking suite within dingo to ensure that the Python wrapper remains performant and numerically consistent with the C++ core.
Proposed Implementation:
• Benchmark Script: A new module (e.g., tests/benchmarks/) that:
1.Generates standard high-dimensional test geometries (unit cube, simplex).
2.Executes the same sampler (e.g., Coordinate Hit-and-Run) through both the dingo interface and a direct volesti call (if possible via the existing bindings).
3.Calculates and reports:
• Time per sample (Wall-clock time).
• Effective Sample Size (ESS) comparison.
• Variance/Numerical stability metrics.
##
• CI Integration: (Optional) Explore adding a "Performance regression" check to the CI pipeline to catch future overhead spikes.
I am a prospective GSoC 2026 contributor. This task will help me understand the bridge between dingo's Pythonic layer and the high-performance C++ backend.