Overheadcosts on distributed still image renderjobs

Distributed renderjobs do usually cost about 3-4 renderpoints more than single node renderjobs.

The technical reason for this is the scene preparation time, which counts as rendertime.

That preparation occurs at least a 100 times, instead of just 1 time on a single node renderjob.

(also depending on renderpasses of a job).

By lowering the number of nodes, you can achieve slightly lower rendercosts,

since the overall preparation times get reduced.

This is especially an advantage on short rendering still images (less than 5 hours e.g.)

since an overhead of 3 renderpoints (rp) on a 50 rp job won't be that crucial different, than on a 2 rp job.

Please note that the job will take longer to finish, since less nodes are assigned.