Author Topic: Distributed rendering, same hardware; different performance  (Read 79 times)

2019-03-08, 13:04:24

noggster

  • Users
  • *
  • Posts: 1
    • View Profile
Hello!

We recently built a small render farm at work, with 4 slaves (Ryzen 2700x) with the exact same spec (Same CPU, same RAM, same motherboard, same everything).

Running Cinebench yields almost the same score on all the slaves. And all the slaves have had 24 hour of stress testing done with Prime95 without any errors or any down-throttling because of high CPU temps.

However when we use the slaves for rendering with a noise-limit, the amount of passes for each slave can differ quite a lot.
One slave could be at 60 passes, while two are on 38 and one at 30. And all slaves with the same amount of Updates sent. (Data taken from our latest render, which is still in progress. Other times the difference in a render isn't noticable).

My knowledge in how the software handles passes are limited, so this might be normal. Just thought I'd make sure by checking here, to get an expert opinion on the subject.

Thanks in advance!
« Last Edit: 2019-03-08, 13:08:26 by noggster »