In a typical high-end GPU launch we’ll see the process take place in phases over a couple of months if not longer. The new GPU will be launched in the form of one or two single-GPU cards, with additional cards coming to market in the following months and culminating in the launch of a dual-GPU behemoth. This is the typical process as it allows manufacturers and board partners time to increase production, stockpile chips, and work on custom designs.

But this year things aren’t so typical. GK104 wasn’t the typical high-end GPU from NVIDIA, and neither it seems is there anything typical about its launch.

NVIDIA has not been wasting any time in getting their complete GK104 based product lineup out the door. Just 6 weeks after the launch of the GeForce GTX 680, NVIDIA launched the GeForce GTX 690, their dual-GK104 monster. Now only a week after that NVIDIA is at it again, launching the GK104 based GeForce GTX 670 this morning.

Like its predecessors, GTX 670 will fill in the obligatory role as a cheaper, slower, and less power-hungry version of NVIDIA’s leading video card. This is a process that allows NVIDIA to not only put otherwise underperforming GPUs to use, but to satisfy buyers at lower price points at the same time. Throughout this entire process the trick to successfully launching any second-tier card is to try to balance performance, prices, and yields, and as we’ll see NVIDIA has managed to turn all of the knobs just right to launch a very strong product.

  GTX 680 GTX 670 GTX 580 GTX 570
Stream Processors 1536 1344 512 480
Texture Units 128 112 64 60
ROPs 32 32 48 40
Core Clock 1006MHz 915MHz 772MHz 732MHz
Shader Clock N/A N/A 1544MHz 1464MHz
Boost Clock 1058MHz 980MHz N/A N/A
Memory Clock 6.008GHz GDDR5 6.008GHz GDDR5 4.008GHz GDDR5 3.8GHz GDDR5
Memory Bus Width 256-bit 256-bit 384-bit 320-bit
VRAM 2GB 2GB 1.5GB 1.25GB
FP64 1/24 FP32 1/24 FP32 1/8 FP32 1/8 FP32
TDP 195W 170W 244W 219W
Transistor Count 3.5B 3.5B 3B 3B
Manufacturing Process TSMC 28nm TSMC 28nm TSMC 40nm TSMC 40nm
Launch Price $499 $399 $499 $349

Like GeForce GTX 680, GeForce GTX 670 is based on NVIDIA’s GK104 GPU. So we’re looking at the same Kepler design and the same Kepler features, just at a lower level of performance. As always the difference is that since this is a second-tier card, NVIDIA is achieving that by harvesting otherwise defective GPUs.

In a very unusual move for NVIDIA, for GTX 670 they’re disabling one of the eight SMXes on GK104 and lowering the core clock a bit, and that’s it. GTX 670 will ship with 7 active SMXes, all 32 of GK104’s ROPs, and all 4 GDDR5 memory controllers. Typically we’d see NVIDIA hit every aspect of the GPU at once in order to create a larger performance gap and to maximize the number of GPUs they can harvest – such as with the GTX 570 and its 15 SMs & 40 ROPs – but not in this case.

Meanwhile clockspeeds turn out to be equally interesting. Officially, both the base clock and the boost clock are a fair bit lower than GTX 680. GTX 670 will ship at 915MHz for the base clock and 980MHz for the boost clock, which is 91MHz (9%) and 78MHz (7%) lower than the GTX 680 respectively. However as we’ve seen with GTX 680 GK104 will spend most of its time boosting and not necessarily just at the official boost clock. Taken altogether, depending on the game and the specific GPU GTX 670 has the capability to boost within 40MHz or so of GTX 680, or about 3.5% of the clockspeed of its more powerful sibling.

As for the memory subsystem, like the ROPs they have not been touched at all. GTX 670 will ship at the same 6.008GHz memory clockspeed of GTX 680 with the same 256-bit memory bus, giving it the same 192GB/sec of memory bandwidth. This is particularly interesting as NVIDIA has always turned down their memory clocks in the past, and typically taken out a memory controller/ROP combination in the past. Given that GK104 is an xx4 GPU rather than a full successor to GF110 and its 48 ROPs, it would seem that NVIDIA is concerned about their ROP and memory performance and will not sacrifice performance there for GTX 670.

Taken altogether, this means at base clocks GTX 670 has 100% of the memory bandwidth, 91% of the ROP performance, and 80% of the shader performance of GTX 680. This puts GTX 670’s specs notably closer to GTX 680 than GTX 570 was to GTX 580, or GTX 470 before it. In order words the GTX 670 won’t trail the GTX 680 by as much as the GTX 570 trailed the GTX 580 – or conversely the GTX 680 won’t have quite the same lead as the GTX 580 did.

As for power consumption, the gap between the two is going to be about the same as we saw between the GTX 580 and GTX 570. The official TDP of the GT 670 is 170W, 25W lower than the GTX 680. Unofficially, NVIDIA’s GPU Boost power target for GTX 670 is 141W, 29W lower than the GTX 680. Thus like the GTX 680 the GTX 670 has the lowest TDP for a part of its class that we’ve seen out of NVIDIA in quite some time.

Moving on, unlike the GTX 680 launch NVIDIA is letting their partners customize right off the bat. GTX 670 will launch with a mix of reference, semi-custom, and fully custom designs with a range of coolers, clockspeeds, and prices. There are a number of cards to cover over the coming weeks, but today we’ll be looking at EVGA’s GeForce GTX 670 Superclocked alongside our reference GTX 670.

As we’ve typically seen in the past, custom cards tend to appear when GPU manufacturers and their board partners feel more comfortable about GPU availability and this launch is no different. The GTX 670 launch is being helped by the fact that NVIDIA has had an additional 7 weeks to collect suitable GPUs compared to the GTX 680 launch, on top of the fact that these are harvested GPUs. With that said NVIDIA is still in the same situation they were in last week with the launch of the GTX 690: they already can’t keep GK104 in stock.

Due to binning GTX 670 isn’t drawn from GTX 680 inventory, so it’s not a matter of these parts coming out of the same pool, but realistically we don’t expect NVIDIA to be able to keep GTX 670 in stock any better than they can GTX 680. The best case scenario is that GTX 680 supplies improve as some demand shifts down to the GTX 670. In other words Auto-Notify is going to continue to be the best way to get a GTX 600 series card.

Finally, let’s talk pricing. If you were expecting GTX 570 pricing for GTX 670 you’re going to come away disappointed. Because NVIDIA is designing GTX 670 to perform closer to GTX 680 than with past video cards they’re also setting the prices higher. GTX 670 will have an MSRP of $399 ($50 higher than GTX 570 at launch), with custom cards going for higher yet. This should dampen demand some, but we don’t expect it will be enough.

Given its $399 MSRP, the GTX 670 will primarily be competing with the $399 Radeon HD 7950. However from a performance perspective the $479 7970 will also be close competition depending on the game at hand. AMD’s Three For Free promo has finally gone live, so they’re countering NVIDIA in part based on the inclusion of Deus Ex, Nexuiz, and DiRT Showdown with most 7900 series cards.

Below that we have AMD’s Radeon HD 7870 at $350, while the GTX 570 will be NVIDIA’s next card down at around $299. The fact that NVIDIA is even bothering to mention the GTX 570 is an interesting move, since it means they expect it to remain as part of their product stack for some time yet.

Update 5/11: NVIDIA said GTX 670 supply would be better than GTX 680 and it looks like they were right. As of this writing Newegg still has 5 of 7 models still in stock, which is far better than the GTX 680 and GTX 690 launches. We're glad to see that NVIDIA is finally able to keep a GTX 600 series card in stock, particularly a higher volume part like GTX 670.

Spring 2012 GPU Pricing Comparison
AMD Price NVIDIA
  $999 GeForce GTX 690
  $499 GeForce GTX 680
Radeon HD 7970 $479  
Radeon HD 7950 $399 GeForce GTX 670
Radeon HD 7870 $349  
  $299 GeForce GTX 570
Radeon HD 7850 $249  
  $199 GeForce GTX 560 Ti
  $169 GeForce GTX 560
Radeon HD 7770 $139  

 

Meet The GeForce GTX 670
Comments Locked

414 Comments

View All Comments

  • chizow - Sunday, May 13, 2012 - link



    For shame...once again you fail horribly, once you're done cherrypicking individual benchmarks you'll see Fermi does outclass Tesla by close to 100%, especially once AA is enabled:

    http://www.anandtech.com/bench/Product/317?vs=305

    Feel free to also compare to techpowerup or computerbase compilations to see similar results, since you obviously haven't been following the industry long enough to experience these revolutionary upgrades first-hand.

    This is all common knowledge to GPU enthusiasts though, including the author of this article. Please update your frame of reference before you comment further. Thanks.
  • CeriseCogburn - Sunday, May 13, 2012 - link

    Cherry picking is this very site and using the firswt bench at every release page CRYSIS your favorite amd game right now ?

    Cherry picking when the SAME VENDOR CARD nVidia is used ?

    Now nVidia is against itself in Crysis, and that's cherry picking ? I used the first game to actually look, because all we've has otherwise is you and your amd fanboys flapping lips... with ZERO evidence...

    Now I'll look at your stupid link which is ZERO EVIDENCE for now since you cannot even be beyond lazy and copy a bit of text from it or claim what you think it proves.

    LOL - so it proves nothing right ? You've got nothing. You have failed.
  • CeriseCogburn - Sunday, May 13, 2012 - link

    " ince you obviously haven't been following the industry long enough to experience these revolutionary upgrades first-hand. "

    Dude, I have been following daily for likely more years than your age.
    Good luck with that stupidity.
    80% and 75% is crap.

    33% down to 20% or so is fine, but we have another problem - the GTX580 was an enormous movement in total as it was tacked on improved after the 480 and was just prior distorting your tiny moaning brain, as you add on your ten years blabbering talking point with your long time amd fanboyism to bolster it.

    It took me a few minutes to prove you lied. :)

    You've got a bit of a point, but not much, and the added features of this new nVidia release ( adaptive v-sync, frame rate target (precisionX) ) negates all of that anyway - not to mention the driver add going all the way back to the 8 series.

    You're just moaning for no reason and lying too much while doing it.
  • Iketh - Thursday, May 10, 2012 - link

    your logic is flawed beyond belief
  • chizow - Thursday, May 10, 2012 - link

    Really? How so? If it was so flawed it should be simple for someone so clever as you to poke it full of holes.

    I'm waiting. :/
  • CeriseCogburn - Saturday, May 12, 2012 - link

    Here's a hole so large you'll be moaning when I'm done, and not about pricing or performance increase.

    A lot of people may wait a couple generations to upgrade, or go from a top tier card 2 or 3 gen back to a secondary card now new - or vise versa, cross over the big two, etc. etc.

    In your retentive and specialized moaning, you've restricted end user reality to a single specific instance you've handily outlined as your only metric, and have declared your single path to be the only qualifying upgrade doctrine to use.
    Now there's a zeal of rectal tightness one can easily surmise no end user gaming enthusiast has ever adhered to in their purchasing history, in the entire world, not once, ever.

    So what we really have is a much varied user base in the card(s) they currently run, and a quite varied distane and jump, node, architecture, two cards to one, one card to two, using a current card as a PhysX boost for a hot and cheap upgrade, etc.

    Thus, a person can wait out the $499 nVidia flagship launch or one or two, or some in between node shrink G80 to G92b, 280 to 285, 470 or 480 to 580, etc, and make the jump NOT when your choice choking and frankly stupid single choice only stroke my moaning firebrand demanded purchase scenario rears it's stupid dead head.

    In other words, the $499 you complain about is not the second $499 the real gamer and end user customer spent, they've been sitting a round, and are only spending once, not on your miniscule upgrade single purchase own only before and after rant line....

    So people figure it out in spite of your complaining, and make an enormous jump in their upgrade, or sell off a sli of cf set and barely spend a dime for a good "reset" for a future dual card perf bump on the cheap, or take the second or third or prior tier for a spin with a healthy discount from the release you hate with passion so much.

    You see, you've become a one trick pony, the one trick an amd fanboy can rage about and pretend to have a point - now I wouldn't mind so much if your 75% and 80% crap wasn't so obviously a doubly inflated lie - but on the other hand the initial constraint you introduce is near worthless for any current end user your hoped for perfectly having a fit scenario would apply to !. -

    NOTE: I'm so close to current performance because in the last ten years of those wonderful and enormous increases chizow has so adeptly been gassing the entire room about in hyper ventilation, that I think I'll keep my recently purchased flagship(s) that enjoyed not long ago that great and gigantic leap of power chizow loves in his tiny red heart so much ! Thanks chizow ! I can sit here a big fat winner with all my money in my pocket and it's such a poor increase I am win still for zero dollars !

    See how that works genius ? :-)
  • SlyNine - Saturday, May 12, 2012 - link

    Your constant use of twisted circular logic is amusing.
  • CeriseCogburn - Sunday, May 13, 2012 - link

    That's no rebuttal at all. We'll go with you and your chizow pal's upgrade path right - the one and only you and he allow for your argument ... that's not twisted..

    (rolls eyes)
  • CeriseCogburn - Sunday, May 13, 2012 - link

    You're actually a person who did exactly what I said SlyNine, you're perfect personal proof, as you have more than once stated you went from 5870 and jumped 2 flagship releases and bought the 680.

    Now, even after personally doing this, you attack my explanation calling it circular logic.

    Look in the mirror amd fanboy. I am sorry your amd fanboy base lifestyle took an upset this round, and you personally decided 7970 sucked compared to 680, and jumped from your 5870.

    You obviously couldn't bring yourself to move to the small performance increase the just prior 6970 was, slapping chizow with that brick unconsciously, you attack me, the person who correctly outlined what actually occurs, that you actually did, by your own words, elsewhere more than once, in these posts.

    ROFL - you really, really, really blew it badly that time Slynine.
  • BulletSpongeRTR - Thursday, May 10, 2012 - link

    Exactly, I'm a lowly line cook for a large restaurant chain making $10/hr. But I have SAVED my pennies for this card and will be ordering one today. If an individual cannot reign in their expenses and put away a little here and there to buy what they want (and lets be honest, a 670 is a want NOT a NEED) then they should not be complaining. I'm nearly done acquiring parts for my first build and will be glad when it's done. One more "Summer if Ramen" is all I can do.

Log in

Don't have an account? Sign up now