perf: better fitting approximations for spergel and moffat #205
CodSpeed HQ / CodSpeed Performance Analysis
completed
Mar 11, 2026 in 0s
Performance Regression: -46.07%
⚠️ Unknown Walltime execution environment detected
Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.
For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.
⚠️ Different runtime environments detected
Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.
⚡ 2 improved benchmarks
❌ 2 regressed benchmarks
✅ 32 untouched benchmarks
⚠️ Please fix the performance issues or acknowledge them on CodSpeed.
Performance Changes
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ⚡ | WallTime | test_benchmarks_lanczos_interp[xval-no_conserve_dc-run] |
101.4 µs | 68.2 µs | +48.77% |
| ⚡ | WallTime | test_benchmark_moffat_init[run] |
83.1 µs | 62.4 µs | +33.07% |
| ❌ | WallTime | test_benchmark_spergel_xvalue[run] |
1.6 s | 2.3 s | -30.21% |
| ❌ | WallTime | test_benchmarks_metacal[run] |
18.4 ms | 34.1 ms | -46.07% |
Comparing spergel-calc-max-rad-eff (441abe5) with main (d2f2d94)
Loading