Performance issue with simple sops - Bug?

   961   0   1
User Avatar
Member
21 posts
Joined: 1月 2016
Offline
Hey all,

Just been working on a quick infection solver setup for an hda. I'm finding some really strange results though whereby just adding a simple vex wrangle which calculates the max value of an attribute and normalises it to 0-1 range, and then colourising by that result, is increasing sim time by 3 times.

You can see in the screen shot that the solver, although should be the heaviest part of the scene, its actually the colour node and the vex wrangle which is taking the longest.

I included my scene file, and added in some extra tests - optionA, B and C, which make even less sense.

And Of course i'm resetting my sim on each of these tests so theres no cache from the previous one influencing.

OptionA - just the output from the solver - takes 31 seconds to do 500 frames on my pc.
OptionB - adds colour based on the infect attrib - takes 19 seconds (though is technically doing extra work!)

Option C - my original setup, takes an infection time attribute (frame the point got infected), divides by the maximum to give a 0 to 1 range and then colours by this. - This takes 1minute 38seconds.

Seems crazy to me, does anyone have any insight into why this might be happening?

Thankyou very much
Edited by richnosworthy - 2021年7月6日 19:23:58

Attachments:
performancemonitor.png (1.4 MB)
RN_InfectRandomChance_B.hiplc (200.1 KB)

  • Quick Links