Dear,
I was testing the deadline software.
I don’t have any nodes to test so i was only looking at the interfaces,…
I did a test render on my own system with a few scenes.
I noticed the render time via deadline (But only my own computer as slave) was almost the double.
Can someone explain me why? Will this effect be gone with multiple nodes?
Another question:
I saw my computer resources (ram + CPU) where only used for ±40% max. 8 % AVG.
How can you adjust this? (If i have some nodes they need to render at 100% )
Last question:
Is it possible to chare GPU power via deadline? I need this to simulate realflow scenes.
Kind regards and all the best 
Hello,
Sometimes different methods(network rendering vs GUI) can result in different times, but double seems a bit on the high side, assuming that double is a significant addition in time spent. There is also the load time of the application, which isn’t usually accounted for in GUI based tests. Can you advise the average time in the GUI and the average time in Deadline? Can you share a job report so we can try to identify places where things lagged?
For resources not being as high as you’d like, we generally recommend using concurrent tasks, which will increase the resource load on the system assuming nothing else is using the resources.
As for GPU rendering, I do not believe we have GPU rendering as something we support in Realflow, but if you can share the workflow you are using, we could advise if it might still work without explicit code to do so.
Dear,
Thank you for your feedback. Sorry i took so long for me to aswer.
I did a test with a more complex scene. Render time was now less when I used deadline! So I’m happy 
I think the problem was that my scen wich took only 1 min to render took more time to load then to render (relatively)…
So I you use concurrent task ( at 1) then your node will only render? And stop all other tasks? (give more priority to the render process).
About the GPU: It wasn’t about rendering but about simulation realflow.
Hello,
Yes, so that seems to show that the majority of the original times you were seeing was start up, so multiple tasks should help as the machine can go from task to task within the same job fairly easily.
Concurrent tasks allows a machine to render more than one task at a time, which should help raise the resources being used. It will not have an impact on any other processes on the machine, nor on any other machines doing tasks for the same job.
I do apologize for poor wording, I tend to call any work being done by the slave a ‘render’ when in fact they can be DBR, simulations, tile jobs, encoding, or actual still image renders. The remainder of what I said, though, still applies.
Cheers,
Dwight