**Title:** Slicing crashes immediately on a multi-object plate when E-cores are
in the CPU affinity mask (i9-13900K) — works with P-cores only
## Environment
| | |
|---|---|
| Snapmaker Orca | 2.4.0, build 01.10.01.50 |
| Orca Web | flutter 2.3.38, build 20260915154633 |
| Profile | 02.03.03.03 |
| OS | Windows 11 build 26300 (10.0.26300), 64-bit |
| CPU | Intel Core i9-13900K — 8 P-cores / 16 threads + 16 E-cores = 32 logical |
| Printer preset | Snapmaker U1 (0.4 nozzle) |
| Process preset | 0.20mm High Quality @Snapmaker U1 (0.4 nozzle) |
| Filaments | 3 slots configured |
## Summary
On a plate of 9 objects totalling ~130,000 triangles, pressing Slice terminates
the process in 1–2 seconds. No error dialog, no entry in the log — the process
simply exits.
Restricting the process's CPU affinity to the P-cores makes it slice normally
and repeatably. The crash requires the E-cores to be in the mask.
## Steps to reproduce
1. Import 9 STLs onto one plate (~130k triangles total; largest single object
42,396 triangles, present twice).
2. Auto-arrange.
3. Slice.
Process exits within 1–2 seconds.
## Affinity results (same plate, same session, repeatable)
| Affinity mask | Logical processors | Result |
|---|---|---|
| default | 0–31 (16 P-core threads + 16 E-cores) | **crash in 1–2 s** |
| `ffff` | 0–15 (all 8 P-cores, no E-cores) | slices normally |
| `ff` | 0–7 (4 P-cores) | slices normally |
Set via Task Manager → Details → Set affinity, or at launch with
`cmd /c start "" /affinity ffff <exe>`.
Note that 16 P-core threads at full load is a *higher* power and thermal load
than the default mask, and that case succeeds. The variable is core type, not
total load — which is why I'd expect a race condition in the slicing thread
pool, exposed by the large timing disparity between P-cores and E-cores, rather
than a resource limit.
## Ruled out
- **Geometry.** Every object slices individually. Pairs and triples of them
slice. A 5-object plate of trivial boxes (2,124 triangles) slices. All meshes
were independently verified closed, manifold, consistently wound and free of
degenerate facets before testing.
- **Object count.** A different 13-object plate at ~64,000 triangles slices
fine on the default affinity; this 9-object plate at ~130,000 does not.
- **Prime tower.** Disabling it (Process → Global → Multimaterial) changes
nothing. All objects are on extruder 1 in any case.
- **Build volume and placement.** Tallest object 178 mm against 270 mm; total
footprint ~26,000 mm² against 72,900 available.
## Secondary issue, probably unrelated but reproducible
Every launch logs this twice:
[error] Slic3r::GUI::PartPlate::calc_exclude_triangles: Unable to create exclude triangles
Traced to `system/Snapmaker/machine/fdm_toolchanger.json`, which the U1 profiles
inherit from:
“bed_exclude_area”: [“0x0”]
A single vertex cannot be triangulated. By comparison the A250/A350 profiles
carry four-point polygons. This leaves the plate with no valid exclusion
geometry and affects every U1 user.
.How about this one: 
Snapmaker_Orca.zip