I would like to point out to the miners on the cluster that putting 1KH on in one instance is the same as putting 100 hash on in one instance.
If you are going to mine the cluster and have big CPUs you need to run more instances.
EG... A HP DL580 G7 should have 16-20 instances open. a G8 would need 18-22 and a G9 would be 25-35 instances open.
so if running 20 instances your rig needs to have 20 GB of ram just for mining.
The way the cluster is made is so small cpus can get a chance to mine a block. The cluster splits the work to the instance like it would in a cpu to the other cores.
Each instance works as a core. more instances better chance of the cluster winning a block. If your rig gets the winning nounce you win 100% of the block and it will go in to your wallet just like if you was mining alown.
If you have a laptop that can do 200 HS on a pool run 2 instances to the cluster if you have enough ram.
Last April 2 - 3 2018 i'm running 2 Servers (2x xeon x5650 - 24 threads ), 1 core i7 3770(4 threads) , 1 amd fx8350 (4 Threads) , 1 server xeon (3 threads ) and 1 corei3 (3 threads) averaging 1000 kh/s running for 24 hrs on cluster mining for a test but no luck to find a block. All together can mine 170 to 190 roicoin using pool mining averaging 1000 kh/s for 24 hrs. Running all cpu with affinity selected for power saving.