now for some Nvidia Pr0n

in The Commons
You currently have no notifications.
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
Yeah...Santa better make that a large box of cash...a very large box.
My credit card just stopped complaining about that I spend so much in the DAZ store. I think it's scared...
Thanks for that, I actually was looking for recommendations for motherboards and a few other parts for a potential new build.
Wonder what was so render-intensive about that Nvidia logo at the end that still took 20 minutes to render with 3 Titans though? ;)
Greetings,
This!
That made no sense... I expected to see some kind of really awesome thing, but a bolted (self-lit) Nvidia on a planar background? The heck? Unless he massively over-did the polys, I feel like Iray would chug that out for a snack using that stack of cards.
Still, my partner and I drooled watching the build...not just the three Titan X's, but the 2 *TERABYTES* of SSD drives, 64G of high-speed RAM... _sigh_ Someday. :)
-- Morgan
WANT!!!!
I hate to say it, he's still got something wrong 21min for that, my twin 980Ti would do it in 5min maybe less are you sure that right ???
Unless that 21 minutes included building the logo...from scratch...with one hand tied behind his back and the othe smaked with a hammer every 5 minutes...
21 minutes on Cinema4D with that setup... on Octane render... sounds like it's not worth the cash that was forked out.
http://prntscr.com/9f83h9
Now here's the pr0n. :P
My dream unit: Customized Boxx Workstation Apexx 4 (Yes, there is an Apexx 5, I know) and i'll only ever get it if I win the damn lottery... but it'll render anything in no-time.
The 980TI for the display, the Titans for Iray/Octane.
This came at just the right time, as yesterday evening I struck oil in my backyard and found a huge cache of pirate gold under the sofa... It is an enlightening video though.
There is something seriously wrong with this test. Google for Blender Titan nVidia and see how those card perform on the benchmark BWM scene (hint, it is takes more than 30 seconds, you have nothing to brag).
1) at 3:58 they appear to be installing the SLI connector which I thought was not advised for Iray
2) The actor/animator guys hat tilt is deliberately trying to be ironic but smacks of being "little rascals-esque", I can't take it seriously. Also his shirt is not as cool as the tech guys (do not google toolless t-shirt btw it's disheartening)
3) during the fast motion build sequence the guitar sounds like a bumble bee.
4) $5K in computers parts and it still caps out at a total of 12GB VRAM for Iray.
Being a gamer you want SLI and if you jump over to work in Iray, then it's a quick simple trip to the nVidia control panel to turn off SLI mode.
gianni pron. how fast can nvidia render a naked gianni6 with strand based hair?
depends on amount of hair
Thor hair
But that was being build for Octane....
well tried a couple in OR4C on my 980ti
can show with anatomy as it vanished!!!!!!
I know, but for anyone going this path for Iray this is a very expensive project with little return on investment for the cost.
I'm still most disappointed with the hat. it's trying to hard.
Just one Titan X has 12 GB of VRAM, and the cheapeset one on Newegg is going for $900. So, my guess is that Iray users going down this path will want, at most, one card.
Chewbacca!!!
what do you give a wookie for christmas, when he already owns a comb
yooo givve him luv and understandinnng
http://www.pcworld.com/article/2898175/nvidias-next-gen-pascal-gpu-will-offer-10x-the-performance-of-titan-x-8-way-sli.html
OK this just got hot and steamy
Those dev boxes look interesting too... but $15k a pop....?
I was about to say "I could build that rig for much cheaper" but when I thought about it, Much of the other internal parts have got to be custom ordered and custom built as I've yet to see an off the shelf motherboard with 16 card slots (8x pci-e) as that's what it will need for 8 double-wide cards. The most pci-e slots you can search for on newegg is 6 and even then you only get 1 result.
Their quad setup wouldn't be imposible to mimic though... Now I'm going to have to go on newegg and gather up the parts and price it out LOL.
...I'd opt for a case with a side panel that had a huge fan instead, never can have too much cooling when running multiple beast GPUs. Never understood the attraction to seeing the guts of the system all lit up. Guess it's a gamer thing.
One thing they didn't mention is whether the memory was dual or quad channel.
The next generation workstation I've been developing will be a true powerhouse which will take advantage of the new 16 GB Pascal GTX GPUs due out next year, along with dual CPUs and twice the DDR4 memory of the build in the video. Effectively it will contain a "deskside" version of Nvidia's VCA using half the number of GPUs (at a fraction of the cost and power consumption) as the Pascal cards are expected to have between 5,000 - 6,000 cores.
Now that's raw rendering power.
The 980 ti is by far the better value for any CUDA-based render engine at the moment. Depending on the settings, it could make sense for that scene to take 21 minutes, especially if SLI is on. I didn't see anything about the resolution of the final output, but if it was QFHD then that would also really increase the render time. You could technically make anything take any amount of time if you turn the samples up high enough.
If anyone is really looking to do something similar, stick with the Z97 architecture. 4 980tis is still cheaper than 3 Titan Xs, the 4790K is about $700 less than the 5960X. 32GB RAM will be plenty, and DDR3 is cheaper than DDR4. Plus, from what I can tell, clock speed is more important for transferring the geometry to the renderer than core count. The scene populates much faster with the 4.0+ GHz of the 4790K than my other, higher core count system. And the 6GB VRAM on the 980 tis is still plenty, especially in Octane where you can change the resources each image file consumes.
Also, the reference cooler is more than adequate if you tweak the fan curve with something like EVGA Precision X or MSI Afterburner and can tolerate some noise. After market coolers, like the ACX 2.0 from EVGA, will suffocate each other and run much hotter than stock coolers if you have them stacked right next to each other.
Finally, if you must use neon green...don't let a client see it. This was a show-off system more than a practical, dignified home rendering station.
My current specs:
Z87-Pro Mobo
3.5 GHz Core i7-4770K
32 GB DDR3
6GB Geforce GTX 780
Windows 7 Pro 64-bit
I'm currently saving up for a 12GB 980ti. Once I have that, I imagine I'll be set for the next several years, until I decide I want more rendering power.
If you want to be set for a long time, wait for the nvidia new gen pascal architecture: from 28nm to 14nm, new archi, double transistors, new meory channel, more ram etc... For a long time, the jump has not been this important.
If you cant wait, just sell your 780 and get a used 980ti, wich just double render perfs. And then resell it to get the Pascal one.
I can see my guts
my side is off to keep it cool
"I'm currently saving up for a 12GB 980ti. Once I have that, I imagine I'll be set for the next several years, until I decide I want more rendering power."
A 12GB 980ti? When is that coming out?
...for the scenes I often create, 6 GB of video memory is not enough to prevent the process dumping to the CPU/Memory using Iray. Once that happens, it doesn't matter how many 980s and CUDA cores I'd have. A 980 TI retails for about 650$, the Titan X about 950$. 6 GB would maybe be enough to render 50 - 60% of my scenes in Iray GPU mode, 12 GB would cover 90% of them.