SlicerX / Measured performance
The numbers, with the method.
Nine models. Slicing time and peak memory. A transparent look at one development benchmark, so you can decide what the results mean for your own work.
A measured snapshot, not a universal speed claim. These are project-run tests from 2026-10-03. Hardware, settings, versions and model geometry all affect the result. Slicing speed does not measure print quality or how quickly the physical printer finishes.
What was measured
- Date and SlicerX commit
- 2026-10-03 · 35d1e3e
- Hardware
- Apple M3 Pro (5 performance + 6 efficiency cores), SlicerX thread pool 11
- Method
- Each slicer from its command line, stock Bambu presets for the P1S 0.4 (0.20 mm Standard, Bambu PLA Basic), G-code written to disk. One warm run, then the median of 4. Peak memory is the process maximum resident size.
- Comparison column
- SlicerX with aegis walls and fixed 0.20 mm layers. The separate default-settings run is not used in the tables below.
- Baseline provenance
- SlicerX full batch on 35d1e3e (P1S); other slicers from the earlier baseline run on the same machine, a Bambu Studio rerun in this batch matched it within 1 to 4 percent
Time to slice and save G-code
Elapsed time includes writing G-code to disk. Each reported value is the median of four measured runs after one warm run.
| Model | SlicerX (aegis walls) | OrcaSlicer 2.4.2 | Bambu Studio 01.10 | PrusaSlicer 2.9.6 |
|---|---|---|---|---|
| 8-color figure | 1.73 | 3.33 | 2.74 | Not measured |
| x-mark | 0.25 | 0.53 | 0.42 | 0.45 |
| x-mark, 2 colors with tower | 0.29 | 0.75 | 0.53 | Not measured |
| tug | 0.29 | 0.86 | 0.72 | 0.74 |
| holes | 0.45 | 1.43 | 1.10 | 0.95 |
| spire | 0.38 | 0.97 | 0.69 | 0.99 |
| knot, tree supports | 2.20 | 3.68 | 3.19 | 2.63 |
| dense | 1.37 | 4.00 | 3.49 | 4.34 |
| plate of 20 parts | 0.50 | 2.95 | 2.60 | 3.65 |
Peak memory
The process’s maximum resident memory during each benchmark. This is not an estimate of total system memory use.
| Model | SlicerX (aegis walls) | OrcaSlicer 2.4.2 | Bambu Studio 01.10 | PrusaSlicer 2.9.6 |
|---|---|---|---|---|
| 8-color figure | 256 | 839 | 525 | Not measured |
| x-mark | 58 | 175 | 108 | 113 |
| x-mark, 2 colors with tower | 65 | 256 | 129 | Not measured |
| tug | 95 | 247 | 149 | 214 |
| holes | 97 | 374 | 161 | 323 |
| spire | 102 | 302 | 194 | 241 |
| knot, tree supports | 348 | 982 | 408 | 635 |
| dense | 449 | 933 | 638 | 1032 |
| plate of 20 parts | 163 | 445 | 278 | 335 |
Read the comparison carefully
- SlicerX columns: its default settings, and fixed 0.20 mm layers with aegis walls (the default wall generator) or classic walls, like for like with the others.
- Knot: all three other slicers run tree supports (Bambu Studio with tree(auto), organic), like SlicerX.
- PrusaSlicer has no Bambu profiles; it runs the same values under its own keys and has no multi-color rows.
- These are the recorded versions, not a claim about the newest version of every slicer.
- A missing result means it was not measured, not that the slicer failed or lacks the feature.
- The published dataset records timings and methodology. It is not a complete reproducibility bundle of model files, preset exports and execution logs.
Download the original benchmark data (JSON), including CPU time and the additional SlicerX settings columns.
Try a model you know in both tools, keep the printer and quality settings comparable, and inspect the toolpaths. Share your own measurements with the community.