Enable up to 6 job threads. I would recommend not to use 6, but if you know what you do, you are free to use 6.
New config settings for the number of Nvidia GPUs installed.
As well as a speed comparison and an information flag if the GPU is set into TCC mode (works within a service)
And define which GPU is used for which job thread.
A new config setting to set room and floor. They do not have any effect, you can use them for anything. E.g. even to specify different client types for your python scripts
A new config setting. User X does not count as artist, so the client will not get into working hours mode.
Decide which job thread will be disabled during your working hours. E.g. disable just the main 3D render thread and allow Nuke and execute jobs.
There is a new setting to choose one of 4 pre-defined config presets: Workstation, Renderfarm, Mixed, Cloud.
If you choose a preset, then the client config is using that preset and does not have its own config any more.
Similar to the config presets, you can now choose which list of drive mappings a client should use.
This way you can map a drive in a section of your render farm to a different file server.
Which might be a syncing caching server for your texture files to reduce traffic between departments.
By default, RR uses the Ghz of a client and the Ghz*h cost setting to sum the cost of a job render.
As this does not work work for high GPU machines, you can override this and set a custom cost per hour for each rrClient.
By default, local scenes and textures are copied to the local RR_Localdata folder on the HDD of the client.
If you have a separated network with a slow connection like a cloud render group, it might be better if the Clients use a shared fileserver on the cloud.
Then the files are copied once by one client to the fileserver instead that each cloud client copies the same file to its local drive.
The render processes are started with low process and I/O priority by default.
To debug issues, you can disable this.