OTOY Forums

Hello all

This is my first post so please go easy on me.
I am very interested in your product and may build a machine just for your render software
(2x480's). First off my goal is to load interiors into Octane (original format 3ds or max, but I understand
that you can convert 3ds/max to obj. files using polytrans for maya, is this true?). Then, I want
to add characters to the mix and be able to animate them in the scenes. I have heard from someone
that you can only load one obj. file into octane at one time, thus making complex animations harder (is this true?).
My main 3d software will be maya (getting it this week).

Thank you for reading my post, I look forward to your reply.

as of now, and till beta 3 that they'll implement instances, we can only load one obj file per scene. Right now you would have to have all the scene in your host app (maya in your case), make there the animation, and then using the exporter transfer it to octane.
Welcome
PS: you can convert max files from within max with the internal obj exporter, or use the nice plugin to do so.

Thanks for your reply Kubo!

I think I understand So ultimately I would be able to work with complex animations with many
things going on at the same time but the main limiting factor would be the amount of Ram my graphics
card has since I would have to load everything in maya and then save the whole scene and animation as
an obj? Then export to octane? Sorry if I'm missing something obvious but I have never used Maya before tbh.
Thank you for this useful information! Oh, and I don't have max, will only use maya. From what I have learned however,
this task should not be a problem if I use polytrans?

Hi there. Yes the big limiting factor is the available ram on the gpu. Be aware that using 2 gtx480's will still only give you 1.5gb of ram as the whole scene needs to reside on both cards.
I don't find this all very limiting though. I've done full animations and have had no problem with it not fitting on the card (gtx470). If you do find that you are running out of memory, it just means that it is time to go back into your host app and start optimizing things a little better. It takes more time, but in the end it makes things run smoother anyway.Another big thing is your texture images. PNG, or other lossless formats, take up a whole lot more memory then basic jpegs. That will save you a lot.

Thank you havensole.

Speaking of ram limitations, the Tesla C2050 has 3gb of ram (coming out next week). It looks like a great card but with a price tag of 2200 Euro I can buy a dedicated rendering machine with 2x480's...How would you compare the rendering speed between these two options?

Sure, the Tesla C2050 would be great for future-proofing, but its just soo damn expensive! Yet, I could just put it in my current machine. If you could get either a brand new machine with i7 920 + 2x480's or a Tesla C2050, what would you choose? I have no idea exactly how much ram my scenes and animations are going to be taking up. I plan on making 1080p videos. Thanks for your input.

btw, you can check out the specs for the C2050 here: (I hope this is allowed)
http://www.nvidia.com/object/product_te ... 70_us.html

Its a tough call. If you are in need of a ton of ram then the tesla is the way to go, but render speed would be compatible to 1 gtx480. Plus you can't plug your monitor into it. If I were making a rig dedicated to octane for my own purposes then I would go with the 2 gtx 480's in an xpander box and not build a whole new system. But that is me and I already have too many full systems lying around.

havensole wrote:PNG, or other lossless formats, take up a whole lot more memory then basic jpegs. That will save you a lot.

I'm not sure if that is correct, since the images need to be uncompressed when they are loaded into memory. I would think that a 1000x1000 JPG would take up the same memory as a 1000x1000 TIF, but I haven't tested this yet.

Thanks for replying to my thread guys, I really appreciate it. Yeah I just had a look at the stats of both cards (after studying the terms) and indeed it does seem like they would be about as fast. That's a real shame for that price...slack. I might be forced into buying two of them, as there are no decent options for the firepro v8800 I was considering getting. Would be a better choice for opencl but hey, not mature enough yet. There is machstudio but IMO the render does not look anywhere near as good as Octane

I wish there was a 3gb 480 release, but it doesn't look like it because they are supposedly getting busy for the gtx 485, which is supposed to only have 1.5gb of ram...

Sorry for getting off topic ^.^ Lets get back to PNG/TIF/JPG formats in real time work;and ram usage. I would like to use lossless unless it take a huuge amount of ram in medium to big environments...
And, if you are running out of ram, does it somehow make animation slugish?

I speak from experience about the images. I use to run out of ram a lot with my old 8800gt 512mb with PNG formatted images and then I started using jpeg. Instantly I noticed a lot more memory available in octane. As far as making the render slower I don't think so. If so I would imagine it would be so minute that it wouldn't really matter.
For animations, in the current workflow of Octane and animation, a fast system (hard drive and ram) would make things faster as the bottle neck right now is the exporting of each frame and constant re-launching of Octane at each frame render. The render itself might only take a few seconds to a couple of minutes, but the exporting each frame and whatnot add another few seconds to a few minutes. Speeding up that process is what will really help you. I know Radiance mentioned using a RAM disk (making a section of your system ram appear as a physical disk) to do the exporting process would be a lot faster, but I haven't tried it yet. I will probably try it when I start rendering my animation for the competition.

Awesome information! Thanks. So you would benefit from an ssd? I may get an one then, didn't see the need before but if it speeds up the render processing time then I'm all for it. Also, do you think that the greater amount of ram is worth it when seeing a little quality difference when using jpeg? I have talked to a few people and they say that gpu rendering cannot "post ram", I am a little confused about this subject. From the way I understand it, post (when talking about ram) means the re-using, or recycling of ram to be used for the next batch of information. Does that mean you keep building up ram packets of info until its full say 2000 frames during an animation sequence at 1080p?