Hi,
We’ve been playing around with the Pool Management in Deadline 6 and were wondering how this is set to develop, personally we’d like to be able to be able to set the graph up in a few different ways:
Option to take into account Slave Normalisation so we can get an actual render power distribution.
Option to see distribution by priority, ie 1st priority and then distribution of the pools. That way we can establish exactly what our priority of spread between the pools is.
Cheers
Nick
Hey Nick,
We’ll be looking at adding more graphs to Deadline 6.1.
We could see adding a graph option that groups slaves based on cpu power, but the normalization index is really meant to only factor into render times. If we did take that index into account, doesn’t that actually give you less of idea of your overall farm power? If I’m misunderstanding something here, let me know!
We’ll add this to the wish list.
Cheers,
Hi Ryan,
To be honest this is not a big thing and maybe there’s a better way of doing it but I just find the normalised time very useful because we’ve calculated it from actual render tests whereas the CPU doesn’t seem to relate perfectly to actual render times. An example of this is that some of the better machines CPUs are 24@2.7Ghz - 64.8Ghz (hyperthreaded) and the older machines have 8@2.5Ghz - 20Ghz (not hyperthreaded). Based on CPU alone the difference should be about 3.2x, however when I’ve run render tests on the machines it actually comes out at around 2.6x.
It’s not necessarily that important but when I’m splitting the farm up I generally make a guess at how much power each pool is getting and a nice graph that showed it might be useful, whether that’s CPU/normalised rendertime, either would be useful.
Nick