fix: better init of spergel HLR and tabulate xValue#258
Draft
beckermr wants to merge 19 commits into
Draft
CodSpeed HQ / CodSpeed Performance Analysis
completed
Jul 7, 2026 in 0s
3 benchmarks regressed
⚠️ 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.
⚡ 4 improved benchmarks
❌ 3 regressed benchmarks
✅ 29 untouched benchmarks
🆕 2 new benchmarks
Warning
Please fix the performance issues or acknowledge them on CodSpeed.
Performance Changes
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ❌ | WallTime | test_benchmark_spergel_init[run] |
7.4 µs | 40.3 µs | -81.77% |
| ❌ | Simulation | test_benchmark_spergel_init[run] |
104.1 µs | 267.2 µs | -61.02% |
| ❌ | Simulation | test_benchmarks_lanczos_interp[xval-no_conserve_dc-run] |
828.5 µs | 1,248.5 µs | -33.64% |
| ⚡ | Simulation | test_benchmark_spergel_xvalue[run] |
34,862.4 ms | 310 ms | ×110 |
| ⚡ | WallTime | test_benchmark_spergel_xvalue[run] |
1,819.6 ms | 37.6 ms | ×48 |
| ⚡ | WallTime | test_benchmark_spergel_calcfluxrad[run] |
225.7 µs | 121.6 µs | +85.6% |
| ⚡ | Simulation | test_benchmark_spergel_calcfluxrad[run] |
1,090.6 µs | 626.3 µs | +74.13% |
| 🆕 | Simulation | test_benchmark_spergel_init_float[run] |
N/A | 104.2 µs | N/A |
| 🆕 | WallTime | test_benchmark_spergel_init_float[run] |
N/A | 7.5 µs | N/A |
Tip
Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.
Comparing test-spergel (ef1603d) with main (9b25b7d)1
Footnotes
Loading