Parallel path is computed faster on Mac M1 Pro than i7 9700f

[PureWasabi]

As title says when generating a parallel pass for a 2.5d carving the process is much faster on a M1 Pro than a i7 9700f

On the Mac takes 240 seconds

On the PC I've waited more than 10 minutes without success

I don't think it's normal even if the processors are different

[PureWasabi]

this is the path

[a.laasW8M6T]

Hi

I wouldn't imagine this would take too long on PC, Parallel generally is one of the fastest toolpaths to calculate.

The I7 might be thermally throttling though, what is the clock speed during calculation?

Can you share the File here so we can test?

Goto File>Export> Save as .f3d/.f3z

[PureWasabi]

Sure. Here you are the tasks performance and the f3d file.

I've launched the generation 1h ago also on an i9-14900kf and it's still computing

[a.laasW8M6T]

Hi

I am also seeing very slow calculation times, The CPU is running at 100% on all cores but doesn't really seem to be doing anything, and normally my cooling fans spin up to full speed but are just idling along.

The slow performance appears to be related to the use of avoid surfaces, if we turn these off it generates in only 58.9 seconds vs over an hour.

You don't really need the avoid surfaces in this case as you are already containing the toolpath with boundaries and heights and the result ends up being the same anyway.

I wouldn't use rest machining here either as it has little to no effect on the toolpath after adaptive, but would be useful if you wanted to use a smaller finishing tool to lean up some areas.

Same with the flat toolpath, there is no need for rest machining to be turned on.

None of this explains the massive performance difference between the M1 and PC though.

@seth.madore are you able to check with someone to see why there would be such a big difference?

[PureWasabi]

On i9-14900kf it took 6112.5s I think there are some serious bugs/memory leaks here

[seth.madore]

Gee whiz, that first Parallel toolpath generated in 28.9s on my Mac, it's going on 10 minutes on my PC (which is 4 years old). Definitely something going wrong that needs investigating!

[seth.madore · Accepted solution]

After 7700 seconds, my PC finally generated the toolpath. 😧

Once that was done, I took a look in our internal build and the same toolpath (again, on PC) generated in 34 seconds. So, it looks like whatever problem was occurring, we've resolved it in a future release!

[oleg.tikhomirov]

@PureWasabi Thanks for feedback. Most lickely it's a known issue (Slow Parallel with enabled machine/avoid surfaces), which should be fixed in July relase. Please try it again, when new version is available.

@seth.madore I think it's the same problem as in CAM-50289.

[dmealer]

I have a toolpath that takes 45 min to regenerate on my I7 Gaming laptop.

My M3 Pro Mac book does it in 12 min. So i would say your results are accurate