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
There wasn't a huge shift in architecture between Kepler and Maxwell. Iray was first designed on Kepler (I believe) and so as they created it, it was pretty much able to run on Maxwell with very little change.
Pascal is considerably different from Maxwell. It's a pretty significant change in architecture. While fixed-pipeline is still backwards compatible, as are a lot of standard features, CUDA level stuff it a quite a bit different.
No new rig for me until this is implemented in DAZ Studio, period! I can wait.... no worries on my end. NOW HURRY UP DAMMIT!
Me too, RAMWolff. I have an laptop with an old Geoforce 755M so I expect with a 1060 in a laptop I should be getting crazy better performance, like 4x or better. Since I keep my laptops for years (and then have given them to my daughter to monkey with for more years) I tend to try to go high end for the time. I was going to get a machine with a 960M or 965M (even better) but seeing the specs on the 1060, it seems like night and day even compared to the 965M at least for gaming. If render performance is similar I feel like I'd really regret it two or three years from now if I didn't go for that one really impactful upgrade Will cost me an extra couple of hundred bucks though, I'm afraid. Laptop prices go up way more quickly than desktops when adding better hardware unfortunately. There is ZERO chance I spend that money without knowing for sure that it will work and work well in Studio. Fingers crossed!
Why pay the laptop premium?
Oooh! Oooh! Can't wait! :)
I travel about 80% of the time for work. I hope to get that to a more reasonable ammount but for now (and the past two years) that's what it's been. So for me, it's render / model on the road or not at all. So I pay the premium. Such is life!
Can't wait to see benchmarks on this, I had a thread a month ago where I bought a 980 TI and a 1060 except I bought the 1060 before I knew DAZ wouldn't handle it until just now. I sent back the 1060 and got a second 980 TI and it hums along nicely compared to the CPU only rendering I was using which took days to complete which now only takes 30 minutes. I am curious at how a couple of 10xx series in tandem performs compared to two 980 TIs for IRAY...
From what I've read the 1060 won't do SLI so no doubling up. Move up to the next level 1070 or 1080 card and you can do it. I guess the question would then be are you better off with two 1070s or with one 1080, since SLI doesn't scale linearly (i.e. you don't get double performance with two cards).
SLI can't be used with Iray.
Good fortune in finding that perfect solution Grinch!
Yeah well imagine doing that AND paying a $1000 annual subcription fee for Iray, which we get to use for free. We have no right to complain.
Most pros do use Quadro's and invested $4000 per unit. Many more went on the cheap and bought TitanXP thinking they could save a buck, and of course these are the same guys who are whining about Iray subscription price roll-backs. Save thousands and whine about a few hundred. Typical.
I've held off on my Pascal update on my Quadro 6000's. I can tell you from first hand experience a pair of Quadro M6000 renders in Iray much faster than a pair of TitanX (Maxwell), VRAM difference notwithstanding. I rarely come close to using more than 8Gb of VRAM for a scene. However on occasion I did have some big city street scenes that did use upwards of 15Gb and speed was not noticably slower. So on this system I do not see the issue you guys mention about the M6000 24Gb. I had the 12Gb version previously as well and see no change in performance, even with many lights, materials or scene complexity.
The VCA ... well that's just stupid fast in interactive viewport.
The 1000 series cards CAN use SLI. They don't support 3 or 4 way SLI, however. Only 2 way. (except for a few benchmarks, which WILL handle 3 and 4 way SLI. And it is possible for software writers to write their software to support more than 2-way SLI, but it's very complex and most won't bother, since nVidia doesn't support it.)
However, Iray doesn't support SLI, and it can actually cause Iray to have problems or crash. Iray WILL use multiple cards, however. It just doesn't want them to be connected with SLI.
...5,000$ per card is "lotto winnings country" for most of us, though 24 GB VRAM would be nice. I have scenes that just in idle mode, top out at nearly 9 GB when opened. Doesn't leave a lot of memory for actual rendering. Unfortunately my system is a few years old and only supports a maximum of 24 GB Physical memory. So the best I could put in it would be a 12 GB M6000 or Titan-X.
I imagine working with the VCA a scene renders pretty much in realtime in full Iray view mode while you are working on it *sigh*.
With the high cost of high memory GPUs wondering if it would be more cost effective to just build a system with a boatload of quad channel memory (like 128 or 256 GB) and dual CPUs (8 core Xeons or the new AMD Zen CPUs with 32 threads) and just dealing with CPU rendering.
Yes the 10x0 series supports SLI in general but in specific I am pretty sure my statement was correct: the GTX 1060 does not support SLI, the 1070 and 1080 do. To support that, a quote from Tom's Hardware review of the boards: "Notice the lack of an SLI connector up top? Nvidia recommends a GeForce GTX 1070 or 1080 to gamers looking for more performance than a 1060 delivers (of course), and does not support SLI on the 1060." http://www.tomshardware.co.uk/nvidia-geforce-gtx-1060-pascal,review-33613.html
And this from Techradar.com: "Everything isn't rosy with the GTX 1060, however. There's one problem in particular, and is it a biggy: SLI. While opening up the beautifully designed Nvidia packaging, one thing immediately stood out to us - the lack of the usual SLI fingers littering the top of the card." http://www.techradar.com/reviews/pc-mac/pc-components/graphics-cards/nvidia-geforce-gtx-1060-1325075/review
Every review I've seen of the 1060 all say the same thing: no SLI, and I've seen quotes from NVIDIA saying it's not a big deal since it's an entry level gamer card and entry level gamers don't build SLI systems with multiple cards. You want that, get a higher end card. As you've said, not a concern for Iray users but gamers may find it important.
yeh, I'm wondering the same.
...add to that the issue with W10 in which it "reserves" a portion of your VRAM leaving you with less than the maximum for rendering. The more GPU memory you have the bigger the percentage W10 reserves.
Crikey I can get 128 GB of quad channel DDR4 memory for about the price of a GTX 1080. Yeah, the CPUs are a bit more expensive (the 8 core Xeons I am looking at are around 690$ ea.). If I bite the bullet on the UI and move to W8.1 pro, I can have up to 512 GB of physical memory. 8.1 Home supports up to 128 GB.
the problem is that iray is way much faster with GPUs and wayyyyyyy much cheaper. It really depends on your memory needs. but with 12GB of vram gpu, you can achieve a lot of renders.
A 980Ti will be faster than a dual 8 cores xeon system. and you can add 3 more of them on a 500usd mother board.
Past the 8 cores, the price of xeons go sky rocket hight.
...980 Ti's are difficult to find since the 10xx series has been released. I've seen them priced almost as high as a Maxwell Titan-X (good luck finding one of those which isn't used or "refurbished"). That is for only 6 GB which would maybe handle about 50 - 60% of my scenes at best, so a good part of the time, I would still be rendering on the CPU anyway and the multiple cards would be a waste of money save for improving viewport response. And with three of those you not only need some serious cooling but a hefty PSU as well..
For the price of those three hard to find 980 Ti's alone, I could pretty much build an entire system with a single 3.2 GHz 6 core I7, 128 GB of Quad channel DDR 4 memory, Dual SSDs (250 GB boot and 1 TB Library/Runtime), on an X99 MB, a 2 TB Storage HDD, using a less beefy (and much less expensive) GPU to just run the displays.
Indeed.
I have a 980ti, and a 970 for display (it drives two 2560x1440 displays), so doesn't add much to render times, and makes the computer unusable if I do add it.
Although more than half of my scenes fit onto the 980, I'm left performing tricks to get larger ones to render. Those tricks take a cosiderable amount of time; so much so, that I have projects on the back-burner, whilst I consider the matter.
So, I'm looking at alternatives; I think (yes my opinion) that Nvidia are greedy, and whilst their cards are good, their actions show what's bad about limited competition, expecially when coupled with a closed-loop rendering process. I like IRAY, but am not blinded by fanboyism. (If I had the time, I'd be looking at alternatives.)
..wel said.
If Reality4 didn't have the instabilities I ran into, and Lux didn't render in geologic time in CPU mode, I would go back in that direction. At least with Iray, if a scene exceeded the VRAM, it defaulted to CPU mode. With Lux I've had it just crash.
Talking about Xenon 8cores I purchased one on ebay 2.6ghz ($78 dollars with shipping) for a spare single socket 2011 X79 motherboard I had and wow what a difference. My main system is a i7 3930K 6core and I like my Xenon eight core better. I may upgrade my i7 to Xenon when I get a little more money. I'm a little in the red this month with all the purchases I have made.
...for the new system I am planning I need a Xeon v-3 that uses the LGA 2011-v3 Socket so I can take advantage of quad channel DDR4 memory. That requires an X99 MB with 288 pin memory slots. If I'm going to be primairly doing CPU/physical memory based rendering in Iray and Carrara, I want the fastest combination I can get and enough memory overhead to render large scenes in ultra high quality (as in "gallery") so the process does not drop into swap mode.
.
I'd be curious what results you get; I was looking at an 8core i7; then I looked at a couple of xeons, and was surprised about the difference in price of the two, not that much. And the 8core i7 would need (sooner or later) more graphics card horsepower, so increasing the price, well above the xeon. What stopped me was no relaible data on the performance I could get.
... Well that and the cost. a big chunk of change either way.
..OK Daz just released the 4.9.3.117 Public Beta a short while ago and looks like Iray and the 10xx series cards may be a "go".
Downloading it now. We shall soon know.
EDIT:
My GTX 1080 seems to work.
-I'll go cry from joy now-
Could you please make some iray render of the character with hair and the clothes and post the render times, please.
Arabella 7, Jade Scalp Hair, Something Wicked Set. No other items in.
W:1600 H:1200 dimensions.
Base Render settings, no Optix. HDRI map.
Reached 98% convergence in 3 minutes 12 seconds.
And While I was using the PC and had 2 monitors active.
A quick setup.
...OK, so this changes the situation., No more having to look for Maxwell GPUs as now the 8 GB 1070 and 1080 are useable. That would maybe handle 75% - 80% of the scenes I create (my Railway Station Scene still would need a Titan X). Two 1070s would cost about 300$ to 350$ less than a single Pascal Titan X (which is only available direct from Nvidia) and actually offer more CUDA cores total. The one caveat, 4 GB less VRAM to work with.
Really was hoping the Titan X was going to be uprated to 16 GB like the Quadro P5000 was.
Oh crap .. just as well I'm seeing my financial adviser on Thursday ....
It appears that from my initial testing, the GTX 1080 is around 6~10 times faster than CPU rendering with an i7-6700K.
Is there some proper way to test this? Like a certain scene I could render and we'd check how fast it reaches 1000 Iterations?