OTOY Forums

First of all big congratulations on pulling Octane together. Its a massive achievement and I can imagine the work you've put in has been time consuming and complex.

Its win win for us renderheads, we have a new horse in the unbiased race and the interest that Octane is generating is spurring on development in the other unbiased camps. Knowing which horse to back and stick your money on at the moment is a tough one though.

As it stands at the moment we can buy beta at 99 Euros which will allow us to upgrade to 1.0 when its happens. 99 euros is worth a punt but myself and I'm sure other users might be willing to buy into a longer timeframe upgrade path if given the early choice. Keep the 99 euros for beta to v1.0 but have you considered an increased price point for a 3.0 or 4.0 early adopter price? Example 250 euros for upgrades to v3.0. I wouldn't think anyone would complain? Those that have the cash are spurring development for Octane.

I also have a question regarding Octane and if its right for me. As an illustrator I don't tend to render small. 4000 x 3000 is a minimum and I whilst I try to optimise my scenes they can sometimes be very big. I'm wondering if Octane being locked into the video memory is less than ideal for this sort of work? Maybe I should consider Octane for simple work or as an unbiased animation creator?

I'm wondering if Octane being locked into the video memory is less than ideal for this sort of work ?

Well depends on how much money you can spend in video cards
Current mainstream cards (gtx260) can be configured with 1792mb
Current pro cards (quadro FX) can be configured up to 4096mb
Next gen of mainstream cards (gtx480) all gonna have 1500mb minimum
Next gen of pro cards (Fermi) gonna have 3000mb or 6000mb

Well when you can buy 8GB of system ram for less than a 2GB card then you have to think that this next gen of cards will be price heavy.

A 4GB quadro FX is stupid money...

Sure but TESLA and FERMI are not quadro FX
A Tesla C1060 with 4GB is 1200$ and its like a GTX285 (same speed) but twice the memory
You also need to take into account the speed increase vs CPU for the total cost

Just think about the cost of 30 CPU (with 30 motherboards, and 30 time 4GB RAM) vs 1 GPU at 1200$

Hi,

large resolutions are difficult with current GPU's, due to limited video ram.
however, this will soon change as there are already video cards apearing with larger video memory (1.8GB),
and soon 1,5GB and 3GB consumer cards.

We're currently starting with the 99 EURO beta program and we'll see where we go

Radiance

@Radiance

I apologize for my ignorance on this subject but why not utilize the system RAM instead of GPU RAM?

I dont want to Jack the thread, but in light of the above post, just curious if the c1060 tesla is good for octane as this is what I'm planning on using as it seems to have the best price/performance ratio as far as I can see. I assume if its CUDA then happy days?

Sorry for the digression....

because the connection between system ram and the GPU is way too slow.

Radiance

because the connection between system ram and the GPU is way too slow.

I'm also a renderer user not a programmer but couldn't it be like normal renderers. I fly along when my scene fits in ram but if I overspill I go slow because the software is managing my resources and using the hard disk as virtual ram. Would it ever be possible to use the GPU ram as the ideal, if its full then system ram comes in. We take the hit performance wise but at least we still render the project?

radiance wrote:
and soon 1,5GB and 3GB consumer cards.



Radiance

I hope were talking about fermi base GeForce's and not Quadro/Tesla(way over price to be consider "consumer card"). In fact i would kill for a 2-3GB Geforce,that is on a single GPU and single PCB(not the "dual" card; GTX 295).


edit: Currently, the only consumer card that comes with 2GB is the GTX 285.