The comparison page lets you select two or more completed simulations and view their results side by side -- with automated metric extraction, regression detection, and ranking.
The /compare page
Navigate to /compare from the sidebar to access the comparison tool. Select any combination of completed simulations from across your conversations and projects.
Select simulations
Pick 2 or more completed simulations to compare. You can select from any project or conversation you have access to.
Metric extraction
SimPilot automatically extracts key metrics from each simulation: convergence status, iteration count, residuals, and solver-specific quantities (drag coefficient, heat flux, mass flow rate, etc.).
Comparison table
Results are assembled into a side-by-side table with one row per simulation and one column per metric. Sort by any column to rank simulations.
Analysis
The AI highlights key differences, identifies the best and worst performers for each metric, and flags any regressions.
Regression detection
When comparing against a baseline simulation, the system automatically flags significant differences:
| Severity | Condition |
|---|
| Warning | Metric deviates beyond the tolerance threshold |
| Critical | Metric deviates beyond 2x the tolerance threshold |
Default tolerances are 10% for most metrics, 20% for iteration counts, and 50% for residual values. You can set custom tolerances per metric for tighter or looser control.
If no baseline is specified, the first simulation in your selection is used as the reference. You can change the baseline at any time.
Rankings
For each extracted metric, SimPilot ranks all simulations from best to worst. Rankings make it easy to answer questions like:
- Which configuration produced the lowest drag?
- Which mesh density gave the fastest convergence?
- Which turbulence model best matched the experimental data?
What you can compare
Comparisons work across different:
- Parameter values -- Same setup at different inlet velocities, temperatures, or pressures
- Mesh densities -- Coarse vs. medium vs. fine mesh results
- Turbulence models -- k-epsilon vs. k-omega SST vs. Spalart-Allmaras
- Solvers -- OpenFOAM vs. SU2 for the same problem
- Geometry variants -- Different design iterations of the same component
Export
Download comparison results for external use:
- CSV: Full metric table for spreadsheet analysis
- Report: Include the comparison in a generated simulation report (see Exports & Reports)