Network Render License Options Poll

[rappet]

"License per GPU" sounds great! It's a very dynamic and scaleable model. Another Monthly Payment model for big farm installations is imaginable.

Sorry not to agree.. Imho License per gpu does not make sense...
To start with a Titan overclocked is not to compare with a 580 or older card.
license per cudacore calculated with renderspeed would be more fare then just the number of GPU, but that does not sound practical.
And To compare it does not make sense if the OR standalone licence was per gpu...
Imo License is for software not for hardware to be used for the software.

I do not like the idea of limitation of numbers of GPU to be used for renderslaves. Why? I probably never will use more then 12 GPU as renderslaves, but I just think limits is not a positive concept
Maybe A License per slave is most reasonable. And second slave cheaper, third more cheaper, etc... No limits

I do not know what is best, but my guess is Otoy will come up with something that is thought through well (is that correct english?)
Just gave my five cents

[ingostan]

[quote="gabrielefx"]today you can configure these gpu based workstations:
8 gpus on Tyan barebones: 8 Teslas or 8 Titan black edition.
8 gpus on Nvidia grids, same of Tyan barebones.
4 gpus on several semi pro motherboards: 4 Titan Z by EVGA water cooled
many of these systems are certified by Nvidia.
external expansions aren't certified.

/quote]

As far as i understand - and with trial and error - you cant configure any slavenode with only one standalone 2.0 license as of NOW
Or have i misunderstod everything completely?

Regards

[gonzohot]

I voted for "2 free render nodes with 6 gpu limit per node" but i just saw Grimm option right now.
I think grimm proposition better. Can i re-vote ?

[Goldorak]

I like Grimm's option too, but there are technical reasons why it is hard to implement within the current licensing system - we have to charge something per node, as running the commercial engine without a license key for the machine is not an option today. We are working towards something that may come closer to his model, but I don't have further details right now.

Thank you for the feedback. We'll announce something once it's ready.

[kavorka]

As long as an additional full license will add onto the max, I'm cool with that. Would be horrible if I had 3 comps each with a separate license and then have to by a slave for each one so I can network them together.

[Seekerfinder]

I like Grimm's option too, but there are technical reasons why it is hard to implement within the current licensing system - we have to charge something per node, as running the commercial engine without a license key for the machine is not an option today. We are working towards something that may come closer to his model, but I don't have further details right now.

Thank you for the feedback. We'll announce something once it's ready.

Goldorak,
Please also reconsider the GPU count. I have posted about this a few times but have seen no response from you guys on that request. Please increase it for individuals / small studios from the current 12 GPU limit and make it at least 20... That way we can use all our power locally. I know you're working on the model. Just, please don't lock us out.

Thanks,
Seeker

[lucioing]

I'm sorry... Someone could, please, explain to me what does 'Grimm' means?
Thank you

L.

[rappet]

I'm sorry... Someone could, please, explain to me what does 'Grimm' means?
Thank you

L.

first page of htis thread.. Grimm posted:
"I was thinking about a slightly different take on this than the options you have in the poll. I don't know how well it would work with the current licensing scheme, but here is what I would like to see for the network licensing. The first standalone license allows the use of a specific number of GPU's, lets just say that it is the current limit of 12 like it is now. The network demons are not licensed at all and the number of GPUs are tracked by the master computer (the one that has the standalone licensed). When the number of GPUs reaches the limit (12) then any additional GPUs are ignored and not used. To use additional GPUs you then have to purchase another block of 12 more GPUs. This price could be the same as standalone or could be reduced? This scheme allows the greatest flexibility I think as it wouldn't matter if you have 6 computers with 2 GPUs each or 4 computers with 3, etc."

[Seekerfinder]

I'm sorry... Someone could, please, explain to me what does 'Grimm' means?
Thank you

L.

GRIMM: Graphical Render Internal Memory Module

Just kidding. What Rappet said - it's a user name.

Best,
Seeker

[lucioing]

I'm sorry... Someone could, please, explain to me what does 'Grimm' means?
Thank you

L.

first page of htis thread.. Grimm posted:
"I was thinking about a slightly different take on this than the options you have in the poll. I don't know how well it would work with the current licensing scheme, but here is what I would like to see for the network licensing. The first standalone license allows the use of a specific number of GPU's, lets just say that it is the current limit of 12 like it is now. The network demons are not licensed at all and the number of GPUs are tracked by the master computer (the one that has the standalone licensed). When the number of GPUs reaches the limit (12) then any additional GPUs are ignored and not used. To use additional GPUs you then have to purchase another block of 12 more GPUs. This price could be the same as standalone or could be reduced? This scheme allows the greatest flexibility I think as it wouldn't matter if you have 6 computers with 2 GPUs each or 4 computers with 3, etc."

Thank you, rapper!

Here's my thought.
there are two figure:
- otoy
- user

Of course, otoy wants to sell a license per GPU
Of course, user wants free license for unlimited GPU

well, we have to reach a break even point

I think WE HAVE TO PAY otoy, because if we pay otoy, otoy could spend in research and development... but at the same time, we have to maximize our productivity.

For me, I think it's good to pay a node license for a x numbers of GPU. The cost of each license could be around 99$ (if Octane s.a. costs 299$).
Well, Otoy could share for free the first one. So if you buy a license, you get a free license for x GPU.
Now, how much 'x'?
I think it doesn't make sense to spend 5K-8K $ for video cards and complaining at a cost of 99$...
So, I think that 4GPUs could be the honest compromise between us (the users) and otoy (the company). Be honest, how many cards in your PC?
The maximum I saw is 4xTitans... At the moment, I could buy up to two Titan black 6GB: they costs (in Italy) about 1000-1100$= 2000,2200$.
To get 4 cards I have to spend 4000-4400$... I think, at the moment, we can't spend so much money lightly...
And if we could spend for 8 video cards, we can also spend 99$ more.
That's my thought.

L.