Anyone know when R15 is supposed to hit?
It’s always been early september for the previous few releases (can’t state for earlier than R12). I don’t think it will take long from now!
From Maxon’s site:
Got to admit, this time of year drives me nuts. Do my plugins work in R15? Which ones don’t? Have my clients switched? Have my friends?
Though admittedly the prospect of hooking up 38 cores via Team Render, has got my interest levels up.
There’s something about python plugins:
This is why I occasionally launch 11.5 and 12. Every generation seems to lose a handful of useful tools.
Same here! - cant wait to utilise the 42 I have at my disposal on one render!
i completely know what you mean. the plugin thing is such a time sink.
though my “excitement” of the high number of cores under TR has yet to hit… it’s more of a worry still since the one post a bit back about TR crashing with more then 5 nodes and the maxon fan club raves about it working so well with 1 laptop (though demo’s with up to 4 in mash’s video)
i just HOPE they are REALLY TESTING TR with a realistic number of nodes and 1 to 5 is not realistic in a “team” render to me
but looking forward to seeing it firsthand myself too
Well 1-5 is clearly their market for a team
I guess there’s a difference between a team and a farm.
I don’t have 1500 teammates to manage.
It was also hinted that the decreasing performance some have experience above 4 was probably due to traffic congestion. We will soon all have the answer to our concerns…
This doesn’t solve your farm issues, and it’s not a replacement for NET yet, but it’s a very cool additionnal plug and play tools for many studios and freelancers.
Can anyone verify if TR will work (properly) with more than 5? I currently use 7 machines when I’m in a crunch, and we are adding to that every 6 months.
We used up to 10 machines in the US office, and it was fairly stable, at least as stable as beta software can be.
That was awhile ago though, and I can’t comment on the current situation, but I can ask Rick today when the US office opens.
My copy is already on its way …
That’s the kind of info I needed. Thanks Chi!
For animation it should all be largely as efficient as it was before, afterall its just copying a project to a machine, rendering and sending it back. For stills though, it depends how big the job is; the smaller the job, the more the network speed will weigh against you. If the image takes 60 seconds to render, then on a gigabit network, anything more than half a dozen machines and it wont be worth using them, as it will take longer to send the project out and get the results back than the rendering takes.
A larger project though, say a 1 hour render for a still, I doubt anyone would hit any realistic limit for rendering a still of that sort of magnitude.
thanks for weighing in here mash, i appreciate the insight.
while also not wanting to take this VRAY thread away into a team render direction i can’t help but ask one question about your reply.
say you do have lots of machines set up in TR, and it works fine with animations and the full frame rendering - no performance issues etc, can you or how can you then set off a distributed render to your list but not send to the whole TR clients list?
basically if say the performance of DR on gigabit ethernet peaks at say 10 TR clients, can you or how can you send a DR to just 10 of the TR nodes? and not bog everything down by having 40 nodes all churning up the ethernet (or worse crashing the TR nodes)
i know i will be testing this first hand soon enough, but curiosity has got me and if you can share the info that is welcomed
Skip to 5:00. That list is a list of all render nodes, to the left is a checkbox for which machines you want to use. So you could pick half and set off a still image, then pick the other half and set off another project, in theory.
I dont have any machines here right now so I cant guarantee you can send off two renders at the same time, but at a minimum you can limit the render to certain clients.
just installing R15 now, but so far no sign of a separate serial number/way to install a stand alone TR control a la NET Server, so seems a $4K expense for those who want to run TR on a central computer - ugh.
and/or a floating license at $1k and then the third party render manager to run it (at whatever $ expense that is) and then the reality that you still can’t render DR under the third party render manager so really it goes back to above $4K if that is needed (which is what was being pointed to as the reason for killing NET Server)
also 5GB for a TR client only install?! really?
have not launched it yet, but may tonight or tomorrow i suppose and check it out
Most of that will be the content library which projects may reference.