iray hardware benchmarks

For those of you looking into using iray inside Maya or through future solutions like [0x1] integration, the friendly developers at migenius have put together a useful comparison for hardware (GPU) setups.

Take a look at their comparison here using Reality Server: Iray Benchmarks

iray benchmark scene from Evermotion

About David

I am a VFX artist that specializes in Lighting and Rendering. I spend a fair amount of my time supplying clients with artistic solutions as well as technology solutions. With a background in fine art and technical animation training, I strive to bridge the divide between the artist and technologist.

Posted on November 8, 2013, in Uncategorized. Bookmark the permalink. 12 Comments.

  1. I’m using a 680 SC of of late I was thinking about upgrading. Recently this GPU benchmark was posted http://www.cgchannel.com/2013/11/group-test-amd-and-nvidia-professional-gpus-2013/

    As I was saying, I was thinking about upgrading, all benchmarks do not include a SuperClocked GPU, therefore I’m staying with my card, it’s solid, I don’t know if it’s faster then a 580 or on par, regardless the 680 vanilla out preforms a k6000 and a Telsa in iRay.

    I’m in a pretty comfortable position with this card, I could upgrade to a Titan, I really can’t find any justification; neither can I find upgrading to a k5000, which I was also thinking about.

    • Most common reasons to upgrade would be: Newer architecture that requires less power or more available GPU memory.

      GTX cards are fine for iray and should outperform equivalent Quadro/Tesla cards because GTX cards are for raw performance and not underclocked. Quadro/Tesla should last longer than an GTX and provide more memory for a compute solution.

      For occasional rendering if I don’t need the memory, I would stick to the GTX cards. I’m curious how a 690X would perform personally given the speed and number of available cores.

      Keep in mind these decisions will mean more and more as GPU features make their way to mental ray as well, like the forthcoming Global Illumination on the GPU.

  2. What about dual cards, isn’t the 690 just two cards ? I couldn’t tell you which is better Titan or 690.

    • GTX 690 is two GPUs as you say, very similar to the Tesla K10 in specifications with less memory per GPU. I haven’t been able to get my hands on one for benchmarking however I would expect it to be a little bit faster than the Tesla K10 given its clocked a bit higher.

  3. Viewport is a big issue for me, I don’t use alot of GPU rendering, at this point I’d weight viewport speed over GPU rendering as my main focus.

    • Well, other than iray, there’s not a lot of GPU rendering going on unless you use the AO feature for frames. In which case it can certainly shorten overall render times for sequences.

  4. I read what benchmarks tell me is the best GPU for viewport speed. What are anyone opinions on what they think are good GPU cards for viewport enhancement ?

    There are a few GPU rendering options, Octane and as you mentioned iRay, that is a big list 🙂 What I’m getting tired of, and I know it’s a good render is everyone showcasing their 10 second frame shots with Arnold, arghh it’s like ‘everyone has to do the arnold thing’

    • 10 seconds a frame is a sphere on a plane or a machine with 128 cores. Arnold makes no claim to speed in their documents or their marketing for reason. Only ease of use (speed of setup, which isn’t the same as render speed)

  5. How come people claim, I rendered millions of points in seconds with Arnold ?

    • Because you can, Arnold does a great job with lots of data at once. But rendering lots of data and rendering quickly are two different things.

      mental ray and Vray both have the advantage of adaptive algorithms and available biased lighting solutions. Either will outperform for frame quality for speed on most scenes. Arnold is not adaptive and does not offer a biased lighting solution so it’s very easy to use. But this ease of use also makes it slow.

      The render time curve for mental ray and Unified Sampling is logarithmic. For Arnold it’s exponential.

      • Arnold is easy to use, which is why it’s popular. I’ve never used Arnold, why all the questions 🙂 MentalRay can be slow, my knowledge of it is quite strong, I just sometimes forget the order of some things, with all the layers in 3D trying to remember it all can cause blackouts 🙂

      • “Slow” is relative until you try something else. . .

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: