yeah, no offense meant or taken but you’re changing the way i model the club face pattern for a simplistic pattern that doesn’t count as helping or my work for my design. At best you just suggest to flatten the b input of the surface subtract which i already tried. The real problem is how the subtract module works…
i did post a file on the second reply to which you replied. Please let’s be honest about effort here. Im not wasting hours trying for nought…
seriously i dont like at all people who tell me im wrong instead of providing solutions… I was a guru in another dev environment and never liked to be bound by a ‘limiter’ kind of opinion…
I resolve this issue by disabling the resolver option. And recomputing each change. Now the CPU goes to work for 5 minutes instead of 150 minutes…
And the model came out right…
Putter crasher 1.1.gh (56.4 KB)
Yes, doing your surface subtraction from a python script is much faster than using the Solid Difference component (12 sec opposed to 1.1 minutes on my old iMac, using your geometry).
solid_diff_test.gh (73.2 KB)
-Kevin
2 Likes
Finally! GH code and geometry the rest of us can see. 704 spheres, not 200+.
1.2 mins. in R6, no toast. What a bonehead.
3 minutes compute time with 50x40 grid…
The issue is ‘resolver’ on or not on… Resolver on = 30 minutes…
No custom code…
my mistake, wrong inputs… Now works great… results tomorrow…
Thanks Kev R for the great and clear help. Now recompute takes 1.5 seconds…
Makes you wonder why some of these modules are so slow.
1 Like
You could also see a speed increase by tiling the surface you’re subtracting the spheres from. I made a test using the file @Joseph_Oster posted.
- Subtracting 625 spheres with GH SDiff component took ~ 43.8 sec.
- Subtracting 625 spheres with python BooleanDifference command took ~ 11.5 sec.
- Subtracting 169 spheres with with python BooleanDifference command + 2x2 ArrRec took + SUnion took ~ 3.6 sec.
- Subtracting 49 spheres with with python BooleanDifference command + 2x2 ArrRec took + SUnion took ~ 7.2 sec.
timing_comparison.gh (20.8 KB)
Reducing the number of spheres subtracted gives a speed increase, but if the surface is divided into a larger number of tiles the speed increase is offset by the time if takes to join the tiles back together with SUnion.
I also tested to see if the python BooleanUnion command was any faster than the GH SUnion component, and the speeds are about the same.
-Kevin
Thanks Kevin. The python you provided is really fast even chained 4 one after another.
1 Like