brodie_geers wrote:
If it's similar to the current rendering engine (which seems a reasonable assumption) then based on my tests number of cores and higher ghz matter pretty much equally. If you multiply the ghz by the number of cores and compare cpu to cpu the results tend to be quite linear. Only when other factors come in do the results seem to get thrown off (like hyper threading).
Exactly, it scales in a similar way to the main render engine so any conclusion you get using your current hardware can be extrapolated to this preview technology.
For the people who have been asking about other plugins, as I said we can't say yet which plugins will integrate it and which will not, it depends on specific barriers of each SDK. We are doing our best and hopefully soon we will update more videos showing the integration in more plugins. I am sure you understand I cannot talk about things that are not 100% guaranteed :)
This first implementation does not use the network for accelerating the interactive preview. Local networks introduce latencies that don't fit too good with interactivity, in some specific cases there could be some boost but not in general. (I am talking only about this first implementation only, we might research more on interactivity over a network in the future)
Juan