Performance: Subtractive helix & Polar Pattern

About the development of the Part Design module/workbench. PLEASE DO NOT POST HELP REQUESTS HERE!
Post Reply
Posts: 30
Joined: Mon Apr 23, 2018 2:57 pm

Performance: Subtractive helix & Polar Pattern

Post by PJtraveler »

I'm working on knurling a cylinder. It took about 20 minutes with 24 instances to 2 subtractive helixes. The GPU was at 0% or close to it, the cpu was running at 2 % and only 17% of memory was in use. I have a reasonably powerful machine (24 cores) 128 GB mem, 3060 GPU, and lots of disk.
Please could we work on the algo to improve the speed?

Posts: 582
Joined: Mon Mar 30, 2015 2:30 pm

Re: Performance: Subtractive helix & Polar Pattern

Post by jnxd »

One thing I can think of is regarding parallelization. There's not much work to do, but whatever is is being done sequentially. I expect knurling to have a fair bit of intersections, which may be slowing down each of the steps.

As of now I don't know how well parallism is supported in OCCT itself, so what can be done may very well be limited by that.

Are you in the position to show us the model, or any other example where it takes a long time? It can be iaed for future benchmarks.
Post Reply