Hi guys,
I'm currently rendering a project where I'm at about 7:30 minutes per frame using my workstation and one slave. I discovered the following behaviour:
Until reaching about 1280 of 1350 samples, the estimated render time is about 6 minutes. Then my 3 GPUs in my workstation stop rendering and the last 70-50 samples are proceeded only by the one card in my network slave. This leads to a total rendertime of nearly 8 minutes.
Is there room for improvement? I guess the slave has to finish his part of the job when the workstation already finished all the parts. So I understand the logic behind it but maybe there is a way to improve it.
Hard and Software Setup:
Host:
1080 Ti
2080 ti
1080
Slave:
1070
R19.068
V4.01.1-R2
Driver: 417.71