Can someone please try opening this? It’s a simple pulley design. A circle with a single tooth made by 4 lines, drawn on the edge and pattern copied to 160 teeth.
It takes several minutes to open/render that on my v1 ipad pro.
Little better on my Mac mini with M1 cpu. Here I can see in my iStat menu, that only one of the 8 cores are being used. And that one is maxed out, so it’s not like the math is being handed of to the GPU… No temp increase on GPU either. Only on CPU.
I suspect it also runs single core on the ipad. But there is probably some logic/code that needs to be optimised, before it’s make multicore capable.
It’s also heavy in the non-parameric version, but not nearly as much.
Hi Jesper,
On my M2 iPad, it takes about 15 seconds to load. Most of that is “processing” on the load screen. I exported the pulley in x_t format and it imported quite fast.
Then I tried recreating your sketch by projecting from the imported x_t body… This did not work! For whatever reason (maybe the Shapr3d team could determine) the projected sketch is missing most of the outer circumference??
Yeah, it’s the processing part of the load that takes time. While showing the blue icon with the white worm speeding around, trying to catch it’s tail…
Just pulled out the stopwatch.
M1 Mac 23 seconds.
1st gen Ipad pro (A9X soc) 147 seconds with the parametric beta
1st gen Ipad pro (A9X soc) less that 3 seconds in the non-parametric version.
I obviously had to redraw it in the release version, but it’s pretty simple, so I’m quite sure I did the same.
@Bob3DPO Not sure why you see those gabs. I have a solid circle all the way and then the teeth added on. Extrudes nicely too…
Anyways. 600 straight lines should not be a big ask, so there must be a congestion somewhere. Too many bits trying to get through the same hole at the same time…
I did some profiling and found that indeed there are some Parasolid operations which take excessive time for otherwise simple geometry.
HBPM handles sketch fillings in a special way to be exactly reproducible with the same parameters, and something has probably gone awry in that code path. This is the reason why you don’t see the same issue in the release application.
I have created a ticket in our engineering issue tracker to check what exactly goes wrong in this case.
On a sidenote: Is this the way/place to give this kind of feedback?? I noticed there is a “Send Beta Feedback” in testflight, but suspect that’s a standard thing. Had I gone that route you wouldn’t have gotten “test results” from the other folks.
Is this the way/place to give this kind of feedback?
Yes. HBPM’s product managers check the beta forums with some regularity, we engineers also pop up here sometimes but arguably less often.
If you post about a bug/weird performance here its most likely that someone from Shapr3D will take a look at it - but every feature/issue is subject to the same prioritization system, so I cannot make any promises about timelines
Now that’w strange. I just stopwatched it again, and it still loads in 23 in my Mini with M1.
So M2 pro is slower?? Really weird. I guess I can’t use this as a valid excuse to buy new iron…
Well. 20 some seconds once I can live with. But it’s minuttes on my ipad. So if you need to get a dimension from another design, it becomes really annoying.
Shapr admitted it shouldn’t be like this. That’s why I’m poking them. Making sure it’s not forgotten
