Fusion gets stuck forever computing shell

[Bjanders]

Edit: I previously said it was the chamfer causing the issue, but it is the shelling operation. I updated the title and the text below to reflect this.

I have a parametric storage box design where you define, for example, the box dimensions. The box is grid based (like 25mm spacings).  

Two days ago when I made it, I had no issues changing the dimensions. Everything would compute within seconds. Since yesterday, when I change the box grid to something like 10 units, it gets stuck computing a chamfer. I even waited 30 minutes and it was still stuck.

The only thing that I can think of what I did in between it working fine and getting this issue is moving all my projects from Single User Storage to a Team. Of course, my problem can be completely unrelated to this, but I thought I'd mention. I switched to Team specifically for this design where I want to utilize Configurations, but I have not yet taken that into use.

I'll attach the design. In the user parameters, if I change the y_count from say 4 to 10, then it gets stuck computing.

here in Shell1 (which is before Chamfer2 shown below in the popup):

I'm using a Mac M2 Pro. I have not yet tested it on my (more powerful) Windows computer.

Is there something in the design that I'm doing inefficiently? Even so, I wonder why I started having these issues when it initially worked fine.

[Bjanders]

Changing y_count to 5 updates immediately. 6 takes a few seconds; 7 takes a minute or so. I waited for 8 for over 10 minutes and then force quit Fusion.

[Bjanders]

Just to be sure, I recreated it from scratch up to the shell command, and I still have this issue. Is it normal? Any ways to work around it?

It would be nice if there was a time limit for computational operations after which it open a dialog saying something like "This operation is a taking a long time. Would like to abort or continue?". Or even better, if there was a "Cancel" button you could press any time and then it would abort and revert. Now I have to kill Fusion when setting a too high number (8), potentially losing work.

[jeff_strater]

Thanks for sharing the design, @Bjanders - I can reproduce this performance issue.  One small point:  It is not "forever".  I have a brand-new M3 Macbook Pro (so the comparison is probably not fair), but I started the shell before I went into a meeting just now, and it did eventually finish (I estimate about 20 minutes).  But, there does seem to be some kind of exponential behavior at play here that we need to investigate.  I'll have the geometry team look into it.  I'll follow up here if I see anything interesting.

[jeff_strater · Accepted solution]

Thanks for reporting this, @Bjanders - the geometry team verified that this is a problem, and are looking at possible solutions.  I don't think this is a case we have seen before, so it is very valuable that you reported it.

I'm not sure what to tell you to do in the meantime, unfortunately.  If you have any freedom in the design, it might be good to experiment with different spacing between the protrusions on the bottom to see if that affects the outcome at all.

[Bjanders]

Thanks for looking into this!

I already worked around it by making a flat inside bottom (just doing a cut down into the solid body with an offset from the outside walls, instead of shelling), and that will do perfectly fine for now. In the long run I'd be happy to do the shell as well, however.

Yes, I never thought it would literarily take forever -- how could I even verify that 😉 -- but anything that takes several minutes to compute feels like forever. 😀