Adding to Cart…
![](/static/images/logo/daz-logo-main.png)
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.You currently have no notifications.
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
I also have an offer for a 980 4gb for 280 USD.
I think Ill settle for that, it has 2k CUDA cores and memory hopefully should be okay :/
I tried using the Iray Memory Assitance tool to see the size of my renders, but it wouldnt load..
My system:i7 2700k OC 4.5GHz/ GALAX GTX 1080 Ti FE 11GB 1481MHz /RAM 16GB
Windows7 x64/Daz studio version 4.9.3.166 pro
Benchmark results:
gpu only (Optix off/Memory) Total Rendering Time: 3 minutes 17.36 seconds
gpu+cpu (Optix on/Memory) Total Rendering Time: 2 minutes 17.67 seconds
gpu only (Optix off/Speed) Total Rendering Time: 2 minutes 15.85 seconds
gpu only (Optix on/Memory) Total Rendering Time: 1 minutes 57.18 seconds
gpu only (Optix on/Speed) Total Rendering Time: 1 minutes 49.96 seconds
cpu only Total Rendering Time: 46 minutes 37.77 seconds 。。。
Ryzen 1800X @ Stock
GTX 970 x2
32GB RAM
OptiX On
CPU Only: Total Rendering Time: 18 minutes 13.51 seconds
GPU1 Only: Total Rendering Time: 4 minutes 12.5 seconds
CPU + GPU1: Total Rendering Time: 3 minutes 39.45 seconds
GPU1+2: Total Rendering Time: 2 minutes 23.47 seconds
CPU + GPU1+2: Total Rendering Time: 2 minutes 8.30 seconds
Can't wait for the ASUS Strix 1080 Ti's to come out to finish my build!
I think I saw 2 xeons get a 21 minute time, so it can beat a workstation quite easily, though those weren't the latest. Pretty nice.
The 970 solo time matches one of mine, so it did very little to enhance a gpu only score. Not surprising, though.
The cpu+gpu 1 is interesting. I am unable to test because it seems the newest drivers screwed up my render times. Now my 970 is doing 4:55 solo, and the cpu+gpu was 4:41. I lost a ton of time on these drivers (378.78.) I guess its my fault for wanting to play video games on the same same system I use Iray on...
So I'd like to ask what drivers you are on, since I assume you had to install them just now.
I've run this three times twice with only one card and gotten different results which I expected with two cards but not so much on the repeat with only one card
I did close and restart between each test
System Xenon E5 2670 x2
128 GB ram
GTX 1070 8Gb x2
First run with only one card
18 min 36.46 sec CPU only
5 min 01.00 sec GPU only
4 min 20.17 sec GPU + CPU
Second time with one card
8 min 36.03 sec CPU only
4 min 01.20 sec GPU only
2 min 37.25 sec GPU + CPU
Third time after installing second card
CPU only stayed the same
GPU only 2 min 54.33 sec
CPU + GPU 1.30.76
Those are some drastic swings.
I've got a 1080ti inbound as we speak. So my test will by 1080ti + 980ti.
Here are my Ryzen results along with Zotac GTX 1070 non-amp version (ZT-P10700I-10P) and Gigabyte G1 GTX 1070:
Using CPU in combination with two good video cards is a waste it seems as it only helped it by 3.03 seconds. Perhaps the more complicated the scene the more it would be noticed.
BENCHMARKS WERE RUN ** WITHOUT ** VIEWPORT SWITCHED TO IRAY
Perhaps those of you with Ryzen systems could post your experiences with stability and so on to the thread on the new CPUs - I'm sure a lot iof us are interested. http://www.daz3d.com/forums/discussion/149966/ot-ryzen-is-coming-out-soon/p1
1 minute 32 seconds
i7 6800k at 3.4gHz ; 64gb RAM ; 2xGTX 1080 w/SLI bridge
CPU+2xGPU+OptiX prime enabled
Intel i7-5960X, 1080 Ti, Titan X(Maxwell), 980. All components running at default clock. OptiX enabled on all renders
GTX 980 = 3 minutes 25 seconds
Titan X(M) = 2 minutes 34 seconds
GTX 1080 Ti = 1 minute 45 seconds
GTX 980 + Titan X(M) + Gtx 1080 Ti = 52 seconds
1st gen Titan X is about 33% faster than GTX 980
The 1080 Ti is about 47% faster than last year's Titan X(M) and 95% faster than the GTX 980.
Nice test @nothingmore This is helpful.
Nice test. 1080 ti looking good.
Comparison between my Ryzen and my Xenon dual systems
Ryzen 1700X Xenon E5 2670 x2
Ram 64GB 128GB
GPU Nvidia 1070 GPU Nvidia 1070
CPU only 13 minutes 38.75 seconds 18 min 36.46 seconds
GPU only 2 minutes 21.56 seconds 5 min 01.00 seconds
CPU + 1 GPU 2 minutes 13.36 seconds 4 min 20.17 seconds
CPU + 2 GPUS 1 minutes 47.3 seconds 1 min 30.76 seconds
2 GPUS 1 minutes 44.75 seconds 2 min 54.33 seconds
Interesting. That old Xeon really slows things down--even in GPU-only rendering---despite having 8 cores.
Not if you read more carefully. Best score is dual xeon + dual GPU at 1 min 30. That raises a question : Why ?
I'd expect the dual Xeon to be slightly ahead in CPU only mode but iray may be badly optimized for CPU, so that could explain the result
However that can't explain the poor result of the dual xeon with one GPU and the contrast when running two GPU + 2 CPU
@ Robert ; could you do the tests again with optix On/Off and specify optimisation for speed/memory ? Also what are your RAM speed and does your dual Xeon mobo run dual PCIe 3.0 at x16 speed ?
For reference, I am getting a 1080 Ti on Monday and wanted a reference point from my current setup:
Core i7 4770k @ 4.4 Ghz
GTX 780 3GB (overclocked)
GTX 660 2GB (stock clocked)
CPU + Both GPUs: 2 minutes 45.75 seconds
CPU + GTX 780 only: 3 minutes 25.68 seconds
Unfortunatly to run the tests again I'd have to uninstall and reinstall the graphics cards which at the moment I'm unable to do as I have about a weeks worth of video compressing on the Xenon system
I can however answer some of your questions
Mobo is capable of running 3 PCIe 3.0 slots at x16
Ram is 1600MHz DDR3 (PC3-12800) ECC REG server memory
I posted some times earlier where I had run the test three times and had some pretty drastic swings in the times
I have run it with optix on and off but saw no difference
I have also noticed that sometimes Studio has problems on the Xenon system
I'm Running 1 Xeon E5-2670 here with 2 GTX 780 ti's and 32GB 1600MHz - PC3-12800 Ram
I found my old i7-3960x blew the Xeon away in iray, I can only assume a higher clock speed and higher boost CPU is the way to go.
Robert are you making sure to exit each time you run a test before running it a second time? I do know the second time around running the test, without exiting, I personally see a much lower time result.
Just checking.
Yep exited and restarted program each time
I also did the test doing a complete shutdown and reboot for each one
I had faster times on the second and third run than first without exiting
Another thing to check is that the viewport is set to Iray before pressing render. I'm sure you already know this, but this also impacts render times significanlty.
Best to disable SLI, according to what people say.
Got Asus GTX 1070 Dual OC today
GPU: 5 minutes 1.70 seconds
GPU + Optix : 3 minutes 6.73 seconds
Has anyone done a test yet with dual 1080ti's?
marktkirby posted above:
1 minute 32 seconds
i7 6800k at 3.4gHz ; 64gb RAM ; 2xGTX 1080 w/SLI bridge
CPU+2xGPU+OptiX prime enabled
He does have the CPU enabled but I believe the difference will be around 2 seconds or less variance.
1080ti's are completely different specs than 1080s. My dual 980tis are slightly faster than the 1080s, I posted results of 1:27. But dual 1080tis should be faster, but I'd like to see actual proof before I decide if its time to upgrade.
In theory, the 1080ti should be 69% faster than the 980ti, but then going by strictly raw data the 1080 should also be 32% faster than the 980ti, but that doesn't bear out in testing, as you can see. Thats why I want real test results from dual 1080tis.
DvoraszeniaStudios, whoops sorry didn't notice the "ti" portion. Yeah, I'd love to see what people are getting with that GPU as well.
i7-6950X (10 Core) at 4.2 GHz, 64 GB, 2 x Titan X Pascal SLI Bridge
2017-03-23 15:15:05.722 Iray INFO - module:category(IRAY:RENDER): 1.0 IRAY rend info : CUDA device 1 (TITAN X (Pascal)): 2242 iterations, 11.139s init, 61.925s render
2017-03-23 15:15:05.722 Iray INFO - module:category(IRAY:RENDER): 1.0 IRAY rend info : CUDA device 0 (TITAN X (Pascal)): 2435 iterations, 11.139s init, 61.809s render
2017-03-23 15:15:05.722 Iray INFO - module:category(IRAY:RENDER): 1.0 IRAY rend info : CPU: 323 iterations, 11.018s init, 61.798s render
2017-03-23 15:15:05.744 Saved image: O:\Test\test.jpg
2017-03-23 15:15:05.880 Finished Rendering
2017-03-23 15:15:05.924 Total Rendering Time: 1 minutes 14.6 seconds
Disable SLI in your NVIDIA control panel. DS won't use SLI and it can actually slow down your render. I think if you disable SLI, hopefully you can get closer to 1:00 minute flat.