I have a pre-built NZXT with i-7 12700k CPU and RTX 3070 GPU with 32 GB of RAM. There are a lot of instance objects, cloth simulation with self-collision enabled, a rigid body, and collision object, transparent objects, screen-space refractions enabled. I am wondering if this kind of lag is normal given all i have in this scene or if I am simply over-working my PC. I can not attach the blend file because it keeps failing to upload to google drive. I can wetransfer it to anyone willing to help and would be very appreciative.
$\begingroup$
$\endgroup$
4
-
$\begingroup$ Lag is normal, especially with physics. Bake physics for dramatically improved viewport performance. Reduce vertex count of soft body and collision objects, use primitive (sphere etc) collision for rigid body to improve unbaked viewport performance. Apparently you're using Eevee; reduce sample count to improve viewport or rendering performance. This is kind of too vague to answer as it stands, because there is no file, and because you haven't said what you're willing to give up on-- the things that cost performance all exist for reasons. $\endgroup$– NathanOct 16, 2022 at 22:46
-
$\begingroup$ Sorry for being vague. Ideally, I wouldn't want to give up on anything. I appreciate the suggestions. Also, If you are willing, I can email you the blend file to see if there's anything else you recommend. I will try to reduce the sample count. Also, there is no soft body, but there is cloth. Should I reduce the vertex count on cloth simulation? Also, would a large umber of instance objects cause lag? If so, what can be done? $\endgroup$– BananasmackedOct 17, 2022 at 16:38
-
$\begingroup$ No, I'm not interested in sharing email. Reducing the vertex count of anything will improve performance, but reducing the vertex count of (unbaked) cloth will especially impact performance, in the same way as soft body (they're almost the same thing.) Large numbers of anything will cause lag, but if you need large numbers, then instances are already the best way to do it. $\endgroup$– NathanOct 17, 2022 at 16:45
-
$\begingroup$ Okay, thank you. $\endgroup$– BananasmackedOct 17, 2022 at 17:02
Add a comment
|