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
Dell Alienware Area-51 computer with an Intel i7 Six Core, 32 GB of RAM, a GTX 970 (1664 CUDA cores) with 4GB RAM on the video card.
Intel(R) Core(TM) i7-5820K (6-cores, 15MB Cache, Overclocked up to 3.8 GHz w/ Turbo Boost)
I am completely lost as to CPU and GPU, I loaded in Studio using the default everything, it had 4,883 iterations before 5 minutes then zipped along and was done 5 minutes 27 seconds. I did manage to write down the iterations so I could see how fast it was going- at 1:53 (one minute 53 seconds) it was1403 and at 2:10 it was 1735, 20 seconds later 2137 and 28 seconds after that (at 2:58) it was 2508. At 3:20 it was 2700 iterations. Forty seconds later (4 minutes) 3470. It was 89% done at that point. Thirty seconds later (4:30) 93% and 3920 iterations.
At 5 minutes it was 96% and 4883, then it zipped and that was it around 5:27.
How do I tell about CPU and CPU? SO not technical, as most of you already know. I bought my Alienware 51 last week and am loving it, jaw dropping speed compared to my laptop that I am used to. I needed to switch to a desktop for the RAM and better video cards. This would have easily taken me 5-8 hours or more on the laptop.
System specs: Windows 7, i7 2600K @3.4GHz, 16 GB RAM; MSI GeForce GTX 980Ti Gaming 6G @1329 GHz
DAZ Studio 4.9.0.54
GPU only: 3m 32s
Optix Prime Accel (CPU+GPU): 2m 34s
CPU only (reduced scene: 2 spheres deleted): 45m 27s
Hi,
please how do you guys use 2x GTX980 if Iray doesn't support SLI? ty
No, though you can have the cards set in SLI mode and disable it for Iray.
Iray can use more than one card, but the scene has to fit on the RAM of a single card, they don't pool the RAM.
I got it. Over locking!
CPU 4790k : 4.9Ghz
GPU Matrix-GTX980 with LN2 BIOS : gpu boost clock1502MHz / men clock2000MHz(fix)
2m39s
Do we have some benchmarks comparing 4.9 to 4.8 on identical systems? How well does 4.9 work with an older 600 series gpu?
Hi,
I got 1m54s with 2x GPUs.
GPU 1: Gainward Titan X
GPU 2: MSI 980ti Gaming 6GB
and I got 2m13s with the same GPUs and an old AMD FX8320.
This is pretty disappointing for animation purposes in my opinion.
My Setup:
Daz Studio Version 4.9.0.63
EVGA GTX-970 SSC overclocked GPU to 1472 MHz, memory + 600 MHz
i7-4790K overclocked at 4.8GHz
16GB 1866MHz DDR3 RAM
OptiX Prime Acceleration = checked
DAZ 3D Results:
GPU only 90% = 2 mins 44 seconds
GPU only 100% = 3 mins 26.62 seconds
GPU+CPU 90% = 3 mins 7 seconds
GPU+CPU 100% = 3 mins 45.55 seconds
OctaneBench V2.17 Total Score: 86.74
So in my case the GPU + CPU combination made the render slower. I bought the EVGA video card around Christmas for $264 from Jet.com with a 20% discount coupon. Prices have been rising since then on the NVidia cards and so I'm glad I could afford it when I could. The OctaneBench was included just because someone brought it up on the first or second page.
GTX970 x2 OC1228MHz (SLI-off)
i7-6700K OC7%
Win10Pro RAM 32GB DDR4-2133
---------------------------------------
CPU only = 26 m 36 s
2GPU + CPU = 1 m 56 s
---------------------------------------
(OptiX ON)
Alienware Alpha i3 version
CPU: Intel i3-4130t 2.9ghz
GPU: customized GTX860m 2GB GDDR5 (dunno why Dell had to go for a custom clocked 860m but it's hardwired in and there's no slot to upgrade)
Studio: 4.9.0.63
GPU+CPU: 10min 40.93sec
GPU only: 10min 17.17sec
Laptop Asus X555LAB
CPU: Intel i3-5020u at 2.2GHz
No GPU
1hr 48min
System Stats,
AMD APU A10-7870K OC to 4.5 GHZ
16GB DDR3 2400HZ Radeon R9 Gaming RAM (11-12-12-31)
GPU NVIDIA GTX 980( Not Over Clocked)
GPU only 4 min 57 sec. 5000 iterations
APU + GPU 5 min 47 sec. 5000 iterations
after reading my log files i tried a few more things.
RUN 1:
photoreal devices (APU+GPU)
optix acceleration on
interactive devices (GPU)
3min 45 sec.
RUN 2:
Photoreal Devices (APU+GPU)
Optix on
interactive devices (APU)
4 min, 38 sec.
RUN 3:
Same as 2 with Optix off
5 min 38 sec.
Just rant the test 4 times after installing my new graphics card (same as my first, an EVGA 980 Ti. and I did not even connect the SLI strap as I know that DAZ does not use SLI and didn't want to take a chance on it causing any problems):
With both GPUs rendering (2 x EVGA 980 Ti): 2 min 15 sec
Both GPUs + Optix: 1 min 27 sec!
Both GPUs + Optix + CPU (4790K i7 4.4 OC): 1 min 37 (slower?)
Both GPUs + CPU: 2 min 29 sec (really slow!)
I should also point out that I ran similar tests on my own scenes, back when I only had one 980Ti and now I can see that as you get more CUDAs the CPU becomes not only less useful, but actually a detriment the higher CUDA you have .
My guess is the CPU actually becomes a bottleneck as you are wasting time sending data to the CPU to render when your GPUs can chew through it much more effectively. (Like asking a young one to help you do something around the house and it winds up just taking longer than if you did it yourself.)
Anyway that bottleneck is obviously somewhere between 2816 and 5632 CUDAs, as I know for a fact when I only had 2816 CUDA with the 1st GPU, that the CPU would actually cut down render times by a small amount say 10 seconds on a 2 minute render, so maybe a 8% improvement in speed back then.
But now with 5632 CUDA the CPU is actually a major slowdown. Which actually is a good thing. It means others can learn from this and don't need to waste money on i7's and expensive motherboards. Save your money for your GPUs!
Anyway, now I need to animate.
Nyghtfall,
With a GTX 980 doing the scene in about 5.5m (elsewhere in the thread) I'm finding both your 780 & 980Ti times a bit difficult to understand. Is everyone using the same settings, or is there some other factor I am missing?
With Respect,
Whorfin
Call it 1:37 - jumped about 70 iterations on that last % and closed before I could screen grab... caught it early this time.
Home Grown PC.
MSI x99a MPOWER LGA2011v3 MB
Intel Core i7-5960x (3.0 Ghz / 8 core / 20MB Cache)
64 GB Kingston DDR-4 (2133 mhz)
1300 Watt EVGA Power Supply
Kraken x61 Low Pro - Liquid Cooling Kit
Win 7 Pro (Test Build Image Designed for Software Dev - job perk)
4x 250 GB SSD Raid 1 (5.25” hot swap carriage)
2x 4 TB HDD (Mounted to Case)
2x EVGA GForce GTX Titan (12GB)
2x Msi GTX 970’s (power cables disconnect - gets too hot - need a larger case)
On my brandnew Core i7 6700 (4GHz), 32GB DDR4 RAM + MSI GForce GTX970 4GB GDDR5 RAM with Cuda 7.5 (important!), no overclocking,
the scene takes render time:
23 seconds to reach 80%
2:50 min to reach 85% (~2500 iterations)
4:08 min to reach 90% (~4000 iterations)
~5:08 min to reach 95%
I think, the GPU is much more important than the CPU performance and also, the CUDA optimisation is.
You can connect the SLI and then disable it in the Nvidia control panel before you start up daz. i have it enabled for gamming even though I dont need it and disable it for DS
I just added a 980 ti to my system (I know, just couldn't hold out for the next gen cards that are only a month away . . .)
i7 4770k 3.5ghz
GTX 970
GTX 980 Ti
2 mins 21sec to 100% (95% convergence threshold)
1 min 39sec to 100% with Optix enabled
Full scene rendered as loaded with OptiX checked.
i7-4770 @ 3.4 GHz
16GB RAM
Nvidia GeForce GTX 650 Ti
CPU + GPU = 12 minutes 46.53 sec
CPU only = 31 minutes 2.9 sec
GPU only = 14 minutes 23.87 sec
Someone alredy tried the new gtx 1080?
Thanks for bumping this thread..I was gonna go looking for it soon! :)
2x GTX Titan X
24GB Ram
Intel core i7 950 not gonna bother benchmarking this since it is a 1st gen i7 and I don't have the patience.
GPU Scene Processing took 17.5 seconds and is part of the total.
GPUs only: 1:39 sec
Doubtful, as only a few cards went out to attendees at the debut, and a few sent out to reviewers. They aren't generally available until May 27th. 1070 GTX comes out on June 10th.
All pc parts are standard factory settings and not OC.
i7 3770k
24gb Ram
x1 GTX 970
Time took 3 mins 40 sec with OptiX Enabled
I know this is an old pc, but I upgrading within the next week and I thought this would make for a nice before and after.
Amd Phenom II X6 1100T, 16GB ram, Nvidia Geforce GTX 460 v2
Without Sphere 8 and 9:
CPU: No point
GPU: 18 minutes 22.86 seconds
BOTH: 16 minutes 3.43 seconds
With Sphere 8 & 9:
CPU: 5000 iterations after 3705.739s (61 minutes 45.739 seconds)
GPU: 19 minutes 42.15 seconds
BOTH: 16 minutes 39.6 seconds
I have the system pretty much bought, with an i7-6700K and 64gb of ram. After reading this thread, I'm unsure if I want to go with one 980 ti or two or three cheaper cards. I'm curious to see if adding my 460 to the new system will help or hurt it. Either way, It will be a long time before I can save up another $600+ to put on a video card. Render times are key, but It
So, I just got my 980 ti and couldn't wait for my cpu to get here to before I ran a test on it.
About the same specs...
CPU: Amd Phenom II X6 1100T, 16GB ram
GPU 1: Nvidia Geforce GTX 980 ti
GPU 2: Nvidia Geforce GTX 460 v2
CPU, GPU 1: 3 minutes 5.23 seconds
CPU, GPU 1 & 2: 2 minutes 46.79 seconds
GPU 1: 2 minutes 53.57 seconds
GPU 1 & 2: 2 minutes 35.68 seconds
I'm very happy... and little stocked that my 600 watt PSU could run two video cards.
Silly and important question for most of the posted results I've seen that forgot to mention it. Is that 'time' with or without sphere 8 and 9 in the render?
I'm curious about some of the lower priced cards 970, 960, tho some posts are a bit vague regarding the benchmark settings.
dryice####, yea, I was surprised some of the newer cards are not as bad of a watt-hog as the former 9 years of cards. I'm still ripping the inter-web apart looking for some watt figures, tho the 'Maxwell' chips looks very good compared to some former cards.
OK, for a base line, on a card that is also running the desktop on two monitors, so it is not dedicated exclusively to Iray. And because it is no longer way back in the Studio Iray beta days, I decided to look at the test again on my budget minded system (aka, "Willfully Inadequate For Rendering Computer").
AMD 8350 (8 real cores no fake core hyperthreading junk, 4GHz no OC or throttling) with 32GB system Ram. Zotec GT 730 graphics card with 4GB VDDR3, Windows 7 64bit. None of that has changed at all. Now testing with Studio 4.9.2.70, that is drastically newer then before.
All spheres in the test scene, GPU only, OptiX Prime enabled, with nothing other then Studio running.
Maximum number of samples reached. Total Rendering Time: 35 minutes 34.96 seconds.
The time there is drastically better then the test "Timing out" at 60 minutes back in the beta days. It isn't much better then the face plant times of AoA on HD figures in 3delight, between excruciatingly painful to impossible to work with for setting up scenes. The card will technically do Iray, I didn't really get it just for that, rather I got it to get the desktop out of my system ram and get the desktop out of a second to-be dedicated to Iray card. The GT 730 has no problem with videos and program OpenGL interfaces on the desktop, and it only uses 23 watts. It just is not an Iray or Gaming card, and I don't play games.
Some time ago, I posted some screen-caps that appeared to indicate that the GPU was not firing on all cylinders. The one thing I know for sure, is the card is not thermal-throttling from heat, not at a peak of 51c during the entire test. I can think of a million other reasons I do not have the tools to test for that may be the reason for the less then 100% load on the GPU in Iray. The 4GB card only has a 64bit memory bus instead of the 128bit memory bus the smaller ram (2GB) siblings have, the card is stuttering between CUDA and Desktop refreshing tasks, It may be Watt throttling if it even has that, etc. Whatever it is with Iray, it works fine for everything else I use it for without wasting a few hundred watts on my electric bill.
As for the potential second card for Iray crunching, I'm not thrilled with only having 4GB for the scenes, and I'm still looking around.
I warmed the new system up with Just Cause 3 before I ran the test.
i7 6700k, 32gb, gtx 980 ti
2 minutes 38.13 seconds
This is a hair slower than both the 460 and 980 ti only