Fix the issue of avoiding excessive invalid computations that can cau…

[dongzhongshu]

Fix the issue of avoiding excessive invalid computations that can cause the program to freeze when support is enabled and either support_bottom_distance or support_top_distance is very large.

[dongzhongshu]

Fix the issue of avoiding excessive invalid computations that can cause the program to freeze when support is enabled and either support_bottom_distance or support_top_distance is very large.

[SoftFever]

Thank you @dongzhongshu
Could you please explain a bit more about the issue and its fix?

[dongzhongshu]

Thank you @dongzhongshu Could you please explain a bit more about the issue and its fix?
When I enable the support and set top_z_distance to 50mm (or less), the slicing process gets stuck at 50%. I found that in the trim_support_layers_by_object function within SupportMaterial.cpp, there is no validation of valid values when calculating the number of intersecting layers with the current layer, leading to a large amount of invalid calculations. A similar situation occurs in the calculateCollision function in TreeModelVolumes.cpp.

[vovodroid]

May be just limit this value in PrintConfig.cpp to something reasonable like 1mm?

[SoftFever]

agree with @vovodroid here
It's better to limite the max value for the parameter in PrintConfig(via def->max).
Setting 10mm is fine though

[vovodroid]

Does anyone need 10mm gap?

[SoftFever]

Just in case for those using pellets extruders, they use large nozzles

[dongzhongshu]

Is it possible to implement a feature in the future where automatic support generation can be supplemented with manual support drawing? In such a usage scenario, this modification is also necessary.

[dongzhongshu]

agree with @vovodroid here It's better to limite the max value for the parameter in PrintConfig(via def->max). Setting 10mm is fine though

Thank you for your reply, but setting 10mm will also cause the system to freeze.

[vovodroid]

So let's limit it (or check) corresponding to nozzle diameter or maximum layer height.

[Noisyfox]

I'm wondering if this improves the performance with large models too, such as this one #6098

[discip]

@Noisyfox

I'm wondering if this improves the performance with large models too, such as this one #6098

I did test it but unfortunately could not observe any improvement.

[dongzhongshu]

@Noisyfox

I'm wondering if this improves the performance with large models too, such as this one #6098

I did test it but unfortunately could not observe any improvement.

Perhaps my description was not accurate. The reason for the modification is that when the modified value is too large, it causes unnecessary calculations during the computation of support and the intersection of slicing layers, as subtracting an excessively large z_threshold results in no possible intersection with the slicing layers. This issue you described is unrelated to the one mentioned here. 😊