GeForce GTX Titan

Generic forum to discuss Octane Render, post ideas and suggest improvements.
Forum rules
Please add your OS and Hardware Configuration in your signature, it makes it easier for us to help you analyze problems. Example: Win 7 64 | Geforce GTX680 | i7 3770 | 16GB
User avatar
PolderAnimation
Licensed Customer
Posts: 375
Joined: Mon Oct 10, 2011 10:23 am
Location: Netherlands
Contact:

Ah ok thnx
Win 10 64bit | RTX 3090 | i9 7960X | 64GB
User avatar
smicha
Licensed Customer
Posts: 3151
Joined: Wed Sep 21, 2011 4:13 pm
Location: Warsaw, Poland

7.51 is really impressive. I think that with a waterblock 9ms/s is possible with low temps (about 40C). My 2x680 (watecooled, oveclocked) gives 8.08ms/s.
3090, Titan, Quadro, Xeon Scalable Supermicro, 768GB RAM; Sketchup Pro, Classical Architecture.
Custom alloy powder coated laser cut cases, Autodesk metal-sheet 3D modelling.
build-log http://render.otoy.com/forum/viewtopic.php?f=9&t=42540
User avatar
Refracty
Licensed Customer
Posts: 1599
Joined: Wed Dec 01, 2010 6:42 pm
Location: 3D-Visualisierung Köln
Contact:

brasco,
thanks for the great test.
One thing I would like to know is the exact energy cosumption.
Could you test one thing.
Plug in only the 580, then measure the watt consumption under Pathtracing with a watt meter (simple energy plug).
Then adding the Titan and measure again.
That would be cool.
Cheers
User avatar
brasco
Licensed Customer
Posts: 198
Joined: Wed Apr 13, 2011 2:56 pm

smicha wrote:What is the max voltage for titan ? Can you show the "voltage and tweaks" screenshot?
Nvidia say it's 1.1625v and there's a max +38Mv in the Voltage and Tweaks popout. I've not touched it though, it's all done dynamically.
Refracty wrote:brasco,
thanks for the great test.
One thing I would like to know is the exact energy cosumption.
Could you test one thing.
Plug in only the 580, then measure the watt consumption under Pathtracing with a watt meter (simple energy plug).
Then adding the Titan and measure again.
That would be cool.
Cheers
Hi, I don't currently have a watt meter, I've been meaning to order one, if or when I do, I will report back.
The EVGA 580 is a Classified I have is a very power hungry card (2x8-pin + 1x6-pin) and running both on a 1Kw supply is fine at the moment

Guru3D have an overclocking review and has a pretty good comparison:
http://www.guru3d.com/articles_pages/ge ... ide,3.html

Titan's power usage is pretty amazing for such a large chip.

cheers
brasc
Rig#1 Win 10 x64 | GTX 1080Ti | GTX 1080Ti | GTX 1080Ti | i7 7900K 4.7GHz | 64GB
Rig#2 Win 10 x64 | GTX 1080Ti | GTX 1080Ti | GTX 1080Ti | i7 3930K 4.4GHz | 32GB
Rig#3 Win 10 x64 | GTX 1070| GTX 1070| GTX 1070| i7 2600K 4.8GHz | 32GB
User avatar
smicha
Licensed Customer
Posts: 3151
Joined: Wed Sep 21, 2011 4:13 pm
Location: Warsaw, Poland

Let me add some power consumption test for GTX 680 4gb:
system at idle 177-185W,
rendering with single GTX680 - 285-300W,
rendering with 2xGTX680 - 395-467W
3090, Titan, Quadro, Xeon Scalable Supermicro, 768GB RAM; Sketchup Pro, Classical Architecture.
Custom alloy powder coated laser cut cases, Autodesk metal-sheet 3D modelling.
build-log http://render.otoy.com/forum/viewtopic.php?f=9&t=42540
mikelmnj
Licensed Customer
Posts: 47
Joined: Fri Jan 25, 2013 6:27 pm

brasco wrote:Yes, it seems Titan was made for this sort of workload, you'll still get a good price for the 690 btw, it's the top nvidia card in all the gaming benches.

Bit off topic, does anyone have a K20X at hand to compare? I'd be curious to see if Nvidia have protected the Teslas with their drivers or the like. I know OpenCL is broken in the current Titan drivers (Photoshop crashes a lot...), but they're supposed to be working on a fix.

cheers
brasc
Hey Brasco,

See my post earlier in the thread (Page 4) -- Tesla's seem to be more for brute force simulation than rendering.

Mikel
Windows 8.1 | OctaneRender® v2.0 | Intel Core i7 4770K | 32GB RAM | x2 GeForce GTX 780 Ti SC (3GB/ea.)

Twitter: MikelMNJ
darkline
Licensed Customer
Posts: 291
Joined: Wed Mar 07, 2012 11:56 pm

sorry to be slightly off topic, but am I the only one who is getting virtually no speedup from v1 - 1.11?

running a gtx580 with the 310.90 drivers

v1 3.55 Ms/s
v1.11 3.65 Ms/s

I really need a titan methinks
Windows 7 64bit/ Intel 3930K/ ASUS Rampage IV/ GTX980ti x 2/ 64GB system RAM
kavorka
Licensed Customer
Posts: 1351
Joined: Sat Feb 04, 2012 6:40 am

darkline wrote:sorry to be slightly off topic, but am I the only one who is getting virtually no speedup from v1 - 1.11?

running a gtx580 with the 310.90 drivers

v1 3.55 Ms/s
v1.11 3.65 Ms/s

I really need a titan methinks
I think the speedup for the 5.xx series is less than the 6.xx series.
I got the same speedup for my 580
Intel quad core i5 @ 4.0 ghz | 8 gigs of Ram | Geforce GTX 470 - 1.25 gigs of Ram
darkline
Licensed Customer
Posts: 291
Joined: Wed Mar 07, 2012 11:56 pm

interesting, so the 680 may have caught up with the speed of the 580 now. Might be time to switch to kepler finally.
Windows 7 64bit/ Intel 3930K/ ASUS Rampage IV/ GTX980ti x 2/ 64GB system RAM
User avatar
brasco
Licensed Customer
Posts: 198
Joined: Wed Apr 13, 2011 2:56 pm

mikelmnj wrote: Hey Brasco,

See my post earlier in the thread (Page 4) -- Tesla's seem to be more for brute force simulation than rendering.

Mikel
Great stuff, thanks Mikel.

Regarding the mention of the Titan not being great for VFX sims, I use Turbulence FD for Cinema 4D and I've done some extensive testing over the weekend, results were interesting.
In quite a few scenarios I've seen 50%+ improvement on my 580, in some where the sim isn't highres enough to occupy all the cores, it's only on par, if not worse than the 580.
The base clock only boosts when it sees the card at a higher usage, with 50% of the cores or less, it doesn't boost at all so I bet the lack of Hyper-Q causes that.

Octane seems to have been tuned for 100% occupancy so as soon as I let octane suck up the GPUs the Titan boosts to max clock immediately.
darkline wrote:sorry to be slightly off topic, but am I the only one who is getting virtually no speedup from v1 - 1.11?

running a gtx580 with the 310.90 drivers

v1 3.55 Ms/s
v1.11 3.65 Ms/s
Hi Darkline,
I get a tiny boost with the GTX580 also:

1.10 - 3.77 Ms/s
1.11 - 3.80 Ms/s

So it looks like 1.11 has concentrated on Kepler optimisation.
And yes, you need a Titan in your life :) - I'm already eyeing up a second!

cheers
brasc
Rig#1 Win 10 x64 | GTX 1080Ti | GTX 1080Ti | GTX 1080Ti | i7 7900K 4.7GHz | 64GB
Rig#2 Win 10 x64 | GTX 1080Ti | GTX 1080Ti | GTX 1080Ti | i7 3930K 4.4GHz | 32GB
Rig#3 Win 10 x64 | GTX 1070| GTX 1070| GTX 1070| i7 2600K 4.8GHz | 32GB
Post Reply

Return to “General Discussion”