NVIDIA RTX 2080 Ti
chrbou2001
Posts: 120
Does any body know if this card speeds up rendering time (at least when combining two cards the memory REALLY doubles, not like before).
Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
Do you have a link to that combined memory feature? That sounds really useful if true.
It's called NVLINK:
https://www.nvidia.com/en-us/geforce/graphics-cards/rtx-2080-ti/#sli
In upper right click on the NVLINK button it will drop you to that section.
I think that it will decrease rendering times moreso than just traditional sli due to the increased speed with the NVLINK though it's difficult to say when that will be. Once the cards drop and initial drivers it may be awhile until it is able to be utilized with DAZ rendering.
Nvidia has not released any info about how the new cards/features perform with Iray. As far as SLI, NVidia recommends NOT using SLI
with Iray.
NVLINK is far from SLi
And yet there is no information from Nvidia about NVLink and Iray.
Conceptually, the fully implemented NVLink is far from SLI. However, in practice, depending upon the various implementations with the various architectures, it may end up being very similar in some cases. In the professional/high end Quadro cards, the price alone for an NVLink connector is around $600, while for the gaming versions it's closer to $70. Which implies a major functional difference. Presumably, the low end RTX cards have much less functionality enabled compared to the high end cards, thereby justifying the huge price difference. I wouldn't expect the low end cards to have stuff like combined memory stacking enabled unless developers can go the extra mile to implement stuff like "Unified Memory" across the GPU's.
nvidia does not support IRAY on 2080. officially. ITs up to DAZ if they want to self code support for it, but NVIDIA has discountinued IRAY as a product.
What is your source for that? The plug-ins for various high-end apps, which were produced by nVidia, have now been passed back to the application vendors for support but that does not have anything to do with Daz and DS since it never was one of the nVidia produced plug-ins. The core Iray is still nVidia's as far as I am aware.
The 2080 Ti has more CUDA cores, so the inclination is that it should speed up Iray rendering. More CUDA cores = faster renders. The real question should be, is the performance boost worth the extra $$$ versus buying something like a 1080 Ti? At this point in time, I would say no. Of course, if you have money to throw away, then this wouldn't matter. You're always going to pay more for "the latest" but it doesn't necessarily mean its "the greatest" in regards to price vs performance; especially if the product hasn't been out very long, tried and tested.
I just wanted to comment my two cents on the thread topic:
First, the NVLink "DOES NOT" double the memory or "Pool" the memory on the RTX cards, the ONLY cards where you can combine the memory by a NVLink are the Quadro Cards. Here is a excellent link explaining the differences between the two NVLink, and note that the NVLink for Quadro card is about $600.00 vs $79.00 for the RTX cards. This is very telling. REF: https://www.pugetsystems.com/labs/articles/NVLink-on-NVIDIA-GeForce-RTX-2080-2080-Ti-in-Windows-10-1253/ ;
Second, to the person that posted NVidia is discontinuing IRay, Please do not post such statements unless you can provide a reputable source ( Such as NVidia themselves ) this only starts rumors.
I frequent the official NVidia developement forums and believe me there is no talk or RUMORS of stopping any support for IRay. There are too many HIGH end ( Beside's Daz3d ) programs that incorporate IRay such as 3DS Max And Maya 2019. OK, i am done and that's my 2 cents.
If you were a Lightworks developer you would know. Nvidia stopped development of IRAY along time ago, and all those IRAY implementations you mention are being developed by the companies that create those applications themselves. Nvidia even made a public statement over a year ago to let integrators know.
https://www.nvidia.com/en-us/design-visualization/solutions/rendering/product-updates/
. If you really think Nvidia is still developing IRAY you are in for a rude awakening. It is now up to DAZ to integrate and update IRAY just like everyone else who decided to keep it going after Nvidia stopped. Since you seem to be so upset about this, I will not share what that answer to that is that DAZ emailed me because that would probably make you even more upset :(
Just a point to the OP. Increasing the memory capacity of your GPU(s) will have no effect on the render speed of your scene unless that scene is bigger than the capacity of the GPU. Currently the whole scene is loaded into each GPU, so to speak. Increasing the number of GPUs (increasing the total number of CUDA cores) or having faster GPUs will increase render speed. Whether using 2 GPUs rather than one will double the speed is a question of computer and software architecture, I believe. However 2 GPUs should be significantly faster than one.
Hi Wolf,
does this mean, that now finally the DAZ development has to fix the refraction bug, I already analysed and reported 3 years ago here on the service desk, has to be fixed by DAZ?
Which is exactly what i thought you meant:
these ar the plug-ins that the high-end suites use to communicate with Iray, they are not Iray itself. Daz has always been responsible for the DS plug-in, so there is absolutely no change there.
Richard I think we are actually saying the same thing :)
The punchline is that for further development of Iray for daz , daz needs to do all the work.
I don't think we are - nothing has changed, Daz has always been responsible for the plug-in and nVidia has provided Iray as a .dll to which DS links and passes data. Daz continues to handle the DS side, nVidia continues to handle (and develop) the Iray .dll - just as heretofore.
ah yes.. thats where the problem lays... what happens when those .dll's no longer are going to come :) ( hint, they havn't come for a very long time )
NVIDIA OFFICIAL STATEMENT: The first is what it calls “deep learning super sampling” (DLSS), which is powered by the presence of Nvidia’s tensor cores. These are processors designed specifically to run AI. The application in gaming is effectively an AI-driven anti-aliasing solution. These core are found in the new RTX Graphic Cards.
NVIDIA OFFICIAL STATEMENT: The next version of the NVidia Iray SDK (2018) will include this DLSS access for RTX Graphic Cards.
Doesn't sound to me like NVidia has stopped developing the IRay SDK. Both of the statements were made within the last 60 days and are documented.
Let us not forget that NVidia IRay is a RAY TRACING application, and that the new RTX series of cards are specificially designed to compute REAL TIME RAY TRACING -
I use IRay+ In 3DS Max 2019 for REAL TIME RAY TRACING.
So you've got NVidia developing a AI version of IRay - while developing a series of graphics cards that use AI and you think NVidia is not going to further develope the IRay SDK
Not a chance.
The nVidia Iray libraries? They have been updating fairly regularly - check the DS beta thread for posts from (edited versions of) the nVidia change logs for Iray updates issued since DS 4.10.0.123.
I think another point needs to clarify: NVIDIA has transferred the sales and support of the Iray plugin products.....
..This statement in NO WAY implies that NVidia has stopped developing the IRay SDK they have simply shifted the integration of IRay to 3rd party developers as they have a number of other popular SDK library's.
QUOTE NVidia: To bring AI and further GPU acceleration to graphics, NVIDIA continues to significantly focus on developing SDKs and technologies for software development partners who create professional ray tracing products. With this emphasis, NVIDIA has made product development changes around the Iray and Mental Ray plugin products.
My understanding is that they are cocnentrating on Iray, and not devoting resources to the integration plug-ins (the applications specific plug-ins that give access to Iray). So, if anything, it means more resources for Iray/Mental Ray not less.
All good guys you must know something i don't ;) . Or perhaps it is the other way around.... regardless, lets hope that whatever comes next either an updated iray plugin or Nvidia X , that it is supported by Daz and we can all get back to rendering on these insanely fast new cards.
You can already render on the 20x0 cards, using the version of Iray in the 4.11 Public beta. I thought that had already been said, but this is a new thread on the topic (there have been quite a few).
Nvidia definitely continues development and support of Iray. They'd get sued out of business if they suddenly stopped. 3DS and Maya are both utterly dependent on Iray and those are not little bobbyist programs. Those are the day to day income of graphic artists all over the world.
What Nvidia is no longer producing is the Iray plugins for specific programs like 3DS, Maya and Rhino. DS does not use such a plugin. They dumped those in the laps of the developers of the specific programs. Since, as I already pointed out, DS doesn't access Iray through such a plugin DS doesn't have anything to worry about.
On NVLink and the RTX cards. The RTX cards have the NVLink connector but do not otherwise implement NVLink. All they implement is just an improved version of SLI. I cannot see any reason it would be of any benefit to rendering. If you were to get two or more Quadro cards with NVLink connectors then yes, you could pool video memory, but when would you ever need 20 or 30 Gb's of VRAM for a single render? I've got a 1080ti and I've never used up 11Gb.
QUOTE: Nvidia definitely continues development and support of Iray
Again: NVIDIA OFFICIAL STATEMENT: The next version of the NVidia Iray SDK (2018) will include this DLSS access for RTX Graphic Cards
YES. And when the 3rd party developers implement the new 2018 IRay SDK with support for tensor cores and DLSS we will have a OUTSTANDING and likely extremly fast IRay rendering program. As I stated I am also a 3DS Max User/Artist and I currently use IRay+ that does real time rendering in Max and i can't wait for this to get into production/release.
I have put off buying a RTX card, because i am not one to jump on the bandwagon, i'd rather wait until things mature a bit.... ( and I am pretty dam happy with my Titan ) ...and the fact that if you read the Geforce Forums you will discover that alot of people are having significiant problems with the new RTX cards, in particular the RTX 2080TI.
The future in regards to IRay / GPU Rendering look really exciting to me.... IRay for Max just might give VRay ( which Is actually my "go to" rendering engine for Max ) a run for the $$...
I saw nothing on the change log, and when its out im sure it will be a lot of threads comparing 1080 vs 2080
There are already a few benchmarks, using both the SickleYield scene that has become the standard and a more demanding scene posted by Daz_Rawb.
Hi, I'm a relative noob to Daz (a few months in). I saw Kenshaw's note above that most people never even get a scene that is greater than 11GB. I have a scene using Urban Future 5 with Four characters and it seems to be pushing me out of GPU rendering and into CPU rendering. I have a RTX 2080 ti. I just picked up a 2nd RTX 2080 ti w/ NVLink this evening on an open box sale thinking I would get 22GB of space. However, the Daz scene still gets pushed to CPU rendering.
Two question: Would you expect Urban Future 5 plus 4 charcters to be greater than 11GB (I'm sure it depends on how the characters are clothed + emissives etc - is there a way to tell how big the scene is within Daz. Note: that the Windows Task Manager says I have about 10GB GPU memory in use).
Also, are we sure that NVLinking 2080 ti's won't pool memory to provide 22GB of vram?
thx for you help
Hollaback