GPU render node build

[richard_rosenman]

The Quadro M6000 has 24GB of onboard RAM and that seems to be the highest to date so I think that’s the max you’d be able to get: Nvidia's $5,000, 24GB Quadro M6000 sets a new record for video card memory - Digital Trends

-Rich

[KhyeKading]

@krassi

Hii!

@KhyeKading Well we use Forest pack and tons of proxys ..
The main Problem for me is the increasing resolution and the resulting need for more ram.

i am not sure how i could optimize this with tons of bitmaps and proxies loaded for Rendering. 32 gig is done pretty easy and my 32 gig render cube is 50:50 not rendering because it has not enough ram.

It sound impossible for to render an exterior scene with 4k and only 11gigs of ram that i would have on a 1080ti
So thats why i asked.. It would benefit a lot when working and using RT but it should be able to render.

So thats my Point.. Can i use GPus for Ram intensiv Exterior 4k stuff?

Thanks for all the inputs!

Are you using vray instancer?

[jironomo]

I may be wrong but when using vray RT gpu in the new hybrid mode once the gpu runs our of vram (11gb for a 1080ti) doesn’t the rest get offloaded to the cpu ram?

So if your scene needs 20mb or ram… the gpu will fill up with its 11gb and the remaining 9gb gets offloaded onto the cpu ram?

So your scene will still render instead of crashing like it used to (before hybrid)?

[krassi]

Hi KhyeKading!

Nope!
So i should check this out! Thanks for the hint. It just freaks me out that i am at 32gig when rendering a 1280 pixel preview!!!

We use Forest pack with the newest Version.. And i often think forest pack is the thing thats crashing max all the time and that kills ram.

I just checked for treadripper systems with 2 x 1080ti and those are at around 4500 euros.. thats nice for that firepower.. just need to know if i should go for 4x 1080ti.. !
A single m6000 is around 3000 euros! …Nope. i get a full threadripper system with 64 gig ram for 2500 euros (ok no fast GPU but for a render node its ok).

Seeya Daniel

[biochemical_animations]

@JezUK

Hmmmm, I’m not so sure about this - I pretty sure (from memory) that my performances increased linearly when I went from 2 to 4 gpus…

Could it be an issue with riser cables or the pci-e lanes ?

I’ve gotten all the parts for my open rig (but still haven’t gotten around to building it) - and the plan is to have 7 gpus - so I’m hoping my memory is right… (actually, I can try a test later this week as I am currently back to two cards temporarily).

I don’t think it’s risers, and my lanes are running at 8x 2.0 gen. I’ve read others mention the ROI is not entirely linear.

it maybe because I’m rendering using the denoiser element and the generation of the elements is not streamlined yet…not sure.

however, when the gpus are working on the noise reduction step of the render…that is where I get linear performance…that is where the multi gpu setup really shines. But there are other slowdowns I don’t understand like the elements mentioned above and the rays per pixel reduction step…those are slow and not efficient yet.

This is using BF+BF which is fastest for me.

nonetheless, I still highly recommend the multi gpu rig…and my tests may be workflow and scene dependent…so take this with a grain of salt. I still think anyone going the multi gpu route will be pleasantly surprised.

[biochemical_animations]

@KhyeKading

I’m on the verge of doing a similar build. Currently I’m debating about the PCIe lane speeds. I know that that it will impact the speed at which the scene can be sent to the card. For instance, for a x1 PCIe slot transferring at 250GB/s is it really as easy as saying 4 seconds per 1GB of scene data? For instance will a 5GB scene take 16 seconds to transfer?

Those are theoretical numbers, so yes, it will take an amount of your render time to transfer…but once the scene is transferred to the gpu, it doesn’t transfer again for each frame…just once at the start of a job, to the best of my knowledge. My first frame is always a bit longer than subsequent frames. Try to get max pcie gen and lanes in use.

[biochemical_animations]

@krassi

Hi!
Well we also want to leave our Stone age old i7 Logs and invest in new Hardware.

My main question would be if i could use 1080ti Gpus with 11gig Ram to render my ram intensive Archviz Stuff in 4k!
I get often on my cpu based System greepy ram Usage beyond 32gigs of ram.

Is it actually possible to render stuff like this on 1-4 GPUs with RT? ( i read about using Vray scene files and some stuff i didnt use before in my daily workflow)

The systems will be Threadrippers with 1-x GPUs
So we think of putting at least 2x1080ti in 2 Workstations and then the question would be to get for all together 20k Budget as much threadrippers with 64gig ram that we can kick in our server darkroom ..or more GPUs till the Mainboards are exploding.

So we want to invest the full 20k for more firepower.
Thanks for your help and i will make a similar post here in the hardware section!

Seeya, Daniel

Gpu render ram usage is dependent on the gpu ram, not system…so 11gb is all you get. It’s not combined either…four 1080 ti cards only get 11 gb total to render the frame.