Printable View
I shouldn't even mention this but...
http://www.legitreviews.com/wp-conte...nz-645x362.png
12GB VRAM @ 7GHz
Dual GK110 @ 5760 cores
~8 TFLOPS FP32
TDP: lulz
Price: Don't worry about it. Let's just say it'll be cheaper to buy 3 regular Titans.
Yes, this is real, not a rumor. It was just announced by Nvidia's CEO himself at their GPU conference.
Wow. I was just reading about this announcement. They have to be testing how silly people will be. I can't image the advantage of buying one over other obvious options.
It's what I NEED :)
I ordered 4 of them, I will hook my new Box up to my 20k Watt Whole House Generator ... :o)
If you only have one PCI-e slot in your Pentium 4 motherboard and must crunch Collatz@home,
then you might want a Titan Z!;)
NewEgg will have them on Sale for $3500 w/free shipping ... :D
Looked at the uninformative YouTube vid but still felt uncomfortable below the belt...;) it made the wife jealous!
Quote "TDP: lulz"
For sure!
$3700 *after rebate* with FREE SHIPPING for ShopRunner/V.me/Premier customers! Use promo code SCRWUOUT$ at check out to unlock this user-specific *special deal* as one of our favoritest, sexiest customers!Quote:
Originally Posted by STE\/E
Better?
But yes, still want. My office uses industrial equipment so we have multiple 480VAC/3-phase drop points. SLI here I come!
lol Yes, you forgot Limit 12 Per Customer ... $-)Quote:
Originally Posted by finalfugue
My office has just one 480VAC/3-phase drop and that is more than enough to power 2x Steel Cutting Plasma tables and 4x 45HP electric motors.Quote:
Originally Posted by finalfugue
What type of toys, uh... equipment do you have over there? :)
Well as anybody can tell you, voltage means little without amperage... but we manage to suck a lotta amps too. We're probably going to be out $20K soon in upgrades because we have "only" 250A left on our 480 transformer- or about enough to power three Titan Super Black Mamba Soul Train Editions in SLI.
We make containerized water treatment systems. We shock the water with electricity to add a charge to contaminants, causing them to bind together and settle out/be easily filtered. We manufacture our systems in our facility so we have plasma cutters, welding rigs, tools, cranes, forklifts, pumps, etc. going along with any systems we're testing. Because we design our systems to all run off the same generator or power source, air compressors/pressure washers/heaters and other accessories all use the same voltage. Just don't touch anything, you don't want to be on the receiving end of an arc flash. A 45 HP pump is on the larger side for us, but we sometimes use 80-100 or more if we're moving 1000 gallons per minute+. In short, if we're putting 100A through 1000 GPM on multiple systems in test while welding/cutting/lifting/bolting other systems, we're using ungodly amounts of juice and making a terrible racket.
But I try to squeeze one dual-CPU/GPU workstation into the electronics lab, and a test engineer complains... go figure.
Me thinks that guy needs to be replaced! As he must be one of DA's lackies not to admire the power that rig represents......Quote:
Originally Posted by finalfugue
Fortunately for team SUSA he's generally considered very annoying (unlike me with my dashing charm) and 1. I already had permission for the machine from over his head. and 2. When people heard it bothered him, the only reaction was, "Can you make it louder?"
:-bd =)) =)) =)) :-bdQuote:
Originally Posted by finalfugue
I hear the sound of cooling "upgrades" in the near future... :D Streamers on the fans to make the pleasant rattling sound maybe? XDQuote:
Originally Posted by finalfugue
Why are they hyping it's cost effectiveness on the "super computer" scale (the google brain)? It would cost about 5x less to do 40 petaflops with ~30,000 GTX 750ti than it would with a titan black.
A gtx 750 ti gives you 21.8 gflops/watt @ $150
a titan black gives you 20.5 gflops/watt @ $1000
Correct me if im wrong but the price will scale faster than the performance, because for $3000 a Titan Z won't triple the performance of a Titan Black. So back to my point, first generation maxwell more cost effective in overhead and sustained power. I understand the logistics of having so many more PCBs of the GTX 750 ti than the Gk110s but when we're on the scale of 1000s PCBs... these units are configured for desktops, why are they marketing this thing as cost effective when it's clearly not compared to their most recent release prior to this?
Marketing BS...
Yep it's all marketing BS. Don't forget the Titan Z is also triple slot, even though it appears dual slot at first glance
Sent from my Galaxy S4 using Tapatalk Pro
AMD is smoking crack again. Or maybe they never stopped. I can't even bring myself to post a new thread for this lol
So they just released that R9 295X2 dual Hawaii GPU. $1500. Skip down to the line that reads "Typical Board Power" and suddenly every other spec is meaningless lol
Attachment 1703
I didn't realize you could do that with only 2 x 8-pin connectors. Aren't they only supposed to be able to produce 150W? So that's 150+150+75=375, nowhere near 500W.
https://forums.geforce.com/default/t...-pin-vs-6-pin/
Quote:
the mobo provides up to 75 watts through the PCIE lanes, a 6 pin PCIE plug is rated up to 75 watts, and an 8 pin PCIE is rated up to 150 watts.
http://www.nag.co.za/2014/04/03/amds...-in-the-flesh/ -
"As for powering such a beast, that’s the tricky part. Due to the limit of the PCI-SIG standards, graphics cards can only be as long as 312mm before you begin to run into problems with fitment and strain on the main motherboard. Because of this, AMD only had space for two 8-pin PEG power connectors. The solution, then, is to simply overdraw power from the connectors, exceeding the maximum 300W power specification of the PCI-Express standard. Bus this introduces other problems because the 500W power draw will rely on the 75W delivered by the motherboard, which means that this GPU will only be able to run properly in motherboards conforming to the PCI-Express 3.0 standard."
Holy crap! No way I'd consider getting one of these. (Like I needed any reason other than the $1,500 price tag... :-)) )
I've already burned up a mother board just by using it at stock speeds because the 4-pin CPU Power connecter ended up all charred...
I just love the " simply overdraw power from the connectors" part. Yeah, simple
Artist's rendition of AMD head of product development...
Vesuvius? When even your product codename is that tongue-in-cheek, why? Just why? I call this thing Mt. Desperation. It's obvious to everyone paying attention the graphics companies are having trouble with the next die shrink leap (nVidia was oddly quiet about Maxwell at their conference. A roadmap ain't a silicon wafer buddy). It's OK. We don't hate you for it. Intel invested a small country's defense budget into making the 22nm tri-gate leap, and while that has worked reasonably well, it wasn't necessarily a miracle.
What is bothering me are these stupid, ever-more-absurd halo product releases. I understand why AMD *had* to do this. Some blog called "pwnzringn00bs.com" was going to post a graph of "t3h #1 GPUs for no-scope pwnzringn00bs" with a pure FPS descending order, and AMD had to have something up in the higher echelons. But honestly, how many people are going to buy this for gaming? The same handful of hardcore 1%ers who picked up the other four-figure GPUs I guess. These things take R&D money and manufacturing tooling. Do they actually sell enough of them to recoup those losses? I'd love to see some hard figures on how these actually sell versus their fluff one-upsmanship marketing cost. The technology is NOT going to trickle-down into mid-range popular cards they can then sell by the bushel, obviously. Those cards exist and were immediately sold to coin miners (oops). The Titan Z kiiiiinda makes sense as a "cheap Quadro" for an academics lab, but does anyone see this as a "cheap Firepro"? Maybe? Bueller?
Even if it did, that power draw is absurd. nVidia was conspicuously opaque about the Titan Z's power draw as well. I'm sure both of these cards are pushing (or exceeding) the absolute limits of even fully-compliant PCIe 3.0 mainboards and PSUs. When the tiny traces that deliver power to the PCIe slot on a less-than-perfect mainboard start to melt, are you supposed to just RMA an expensive board with a note, "lol, halo product!" and sit with your system down for weeks? Have fun explaining that to the tech who says the RMA doesn't qualify because you BBQed your equipment.
I went to Newegg. I went to Tigerdirect. A search for both a 295x and Titan Z yielded some TV mounting brackets, Titanfall, and some lamps. I would call these "paper launches" but I don't want to burn down the forums. This game is stupid and they've been playing it for 15 years. Just stop already. The cards that become beloved and legendary are the ones regular mortals can afford and that leverage the best (or a still-awesome, watered-down equivalent) of cutting-edge technology into regular systems. Put these R&D and marketing dollars into *finally* making the next die shrink, and quit pissing around down a dead-end street. Even the inclusion of Asetek (who are awesome) into the R&D process is a SCREAMING signal you're headed in the wrong direction.
Now bow to your volcano god! /rant
Could you set that to music, please? :)
Yes.Quote:
Originally Posted by myshortpencil
Ha! Excellent choice :DQuote:
Originally Posted by finalfugue
For the record, assuming you didn't get a cheapass PSU that uses 18 gauge power wires, you can push up to 300W through each 8-pin connector. Most every connector is overdrawn at some point. However, overdrawing the 75W from the motherboard...i don't feel very safe about that, though i'd be more likely to do it on a Gigabyte board than any other brand.
And yes they are just paper launches. Availability of AMD's dumbassery should be the 21st and the Titan Z somewhere near that as well. Of course this thing will smoke the Titan Z for half the price, though that shouldn't be surprising drawing 500W.
Only GPU water cooled loops need apply. :)
=))=))Quote:
Originally Posted by finalfugue
That makes me nervous for all kinds of reasons. I'm not a hydrophobe, but out-of-the-box closed-loop cooling solutions have a slightly checkered past and too many points of failure for my comfort zone. I really don't want to see that become the "new normal". If it causes closed-loop solutions to become that much more robust and affordable, I'm game, but I'm trying to think of my neighbors and the wider market. It was a commoditized laptop world 7 years ago. It's an increasingly mobile/tablet/console world now. TSMC is running 20nm fab lines... and cranking out mobile SoC stuff on'em. At least server hardware repurposes very well for crunching. What increasingly exotic solutions means for the continued popular existence of my other hobby, PC gaming, is why I give these latest cards the stink eye.Quote:
Originally Posted by Mumps
Sure, this thing CAN be run in the right system. And around here are probably the few people who can do it and are willing to pay for it, but c'mon. In every engineering environment I've lived in, "MOAR POWER!" (in the bad sense) has always been an indication you're out of *good* ideas. Both companies have been at build it bigger now for quite some time (while little things like SLI/Xfire or 4K driver/game/app/OS support, which could use more resources with every new release, remain iffy). The Geforce 800 series and the swashbucklers of the Pirate Islands (wharrrever they be) can't come soon enough. If the 800 series lives up to the hints provided by the 750, it'll be a dreadnought for those of us who pay for our power.
I already had to give up on AMD CPUs some time ago. If I have to give up on ATI GPUs I will, with much grief, give them a Viking funeral. Fortunately it looks like all I have to do is plug them in.
The problem i have with closed-loop cooling being built into a GPU is it dates the hardware. Once the fluid evaporates or the pump fails, the whole thing is useless. This usually happens within 3-5 years; just beyond the warranty of the GPU. If you used it for 2 years, then tried to sell it to get some of your money back, you'd take a significant loss due to no one wanting to chance how much life may be left.
Note to self: Lowball JPM for a 295x in two years...
Sent from my HTC6500LVW using Tapatalk
I counter that with one name...Quote:
Originally Posted by John P. Myers
Swiftech
If only AMD had been that wise. And Asus too, for that matter, with their Ares and Mars silliness.Quote:
Originally Posted by Fire$torm
Is Swift's stuff really that good? Most houses in this area have no A/C, and I'd like to keep at least a couple things chugging in the Summer (that I'm not hiding at work). Is it hydron00b friendly with a little tinkering? I've got some roomy cases with the right mounting spots.
Sent from my HTC6500LVW using Tapatalk
Oh i dunno...i have a GTX 580 overclocked at 1700MHz shader and a GTX Titan overclocked at 1071MHz core, both with Swiftech waterblocks and being cooled in the same single 360mm Swiftech radiator with a single Swiftech pump running the show. My temps are 46C and 40C respectively.Quote:
Originally Posted by finalfugue
Just sayin ;)
Quote:
Originally Posted by finalfugue
Well what do you know, you can actually buy a 295x on newegg http://www.newegg.com/Product/Produc...82E16814150710. I checked on a whim and was surprised to actually see it in stock. Not about to buy it.
Haha. It actually has reviews too! Somehow I doubt one of those is a "verified owner". :D
Sent from my HTC6500LVW using Tapatalk
So the Titan Z was supposed to have been released yesterday but Nvidia couldn't do it because it was slower than AMD's R9 295X2 and the TiTan Z costs 2x more. Base clocks were only going to be 695 MHz haha! Nothing but a $3000 joke.
The only good news for Nvidia is regarding their 750 Ti which will become a water cooled true single-slot on May 13th :D EK is expecting to have their blocks ready by then.