Sign in with
Sign up | Sign in

Nvidia Kepler GK104: More Leaks, Rumors on Specifications

By - Source: 3DCenter.org | B 66 comments

Nvidia's Kepler goes big with its 1536 CUDA cores, a vast improvement over the Fermi's 512 CUDA cores.

As reported on February 7th, we got a first glimpse of the rumored specifications for Nvidia's Kepler based graphics cards. The leaked specifications were met with both "wow.. can't wait" and "wow... those are so fake why even post" from both sides of the comment fence. Now, we are starting to get more pieces of information on the upcoming Kepler series. Based on information coming out of German-based 3dcenter.org, we may have a clearer picture of the true specifications for Kepler GK104.

Outside of the switch to the 28 nm process, one of the major changes in the Kepler architecture is to allow for more CUDA cores. This is achieved by no longer having shader frequency, just GPU frequency. Each Stream Multiprocessor will contain 96 CUDA cores, unlike the 32 - 48 that Fermi had. This change in layout of the CUDA cores will have the GK104 sporting up to 1536 CUDA cores, which is a big boost from GF110 and GTX 580. The number of texture units have doubled from 64 to 128 on GK104. The GK104 will only have 32 ROPS versus 48 in GF110 but it shouldn't affect performance compared to the Fermi.

*BSN: Iillustration of GK104 CUDA Core Arrangement*BSN: Iillustration of GK104 CUDA Core Arrangement

The above GK104 architectural overview comes from Bright Side of News.

Nvidia Kepler GK104:

  • 28nm production at TSMC,
  • Die size 340mm²
  • 4 Graphics Processing Clusters (GPC)
  • 4 Streaming Multiprocessors (SM) per GPC = 16 SM
  • 96 Stream Processors (SP) per SM = 1536 CUDA cores
  • 8 Texture Units (TMU) per SM = 128 TMUs
  • 32 Raster OPeration Units (ROPs)
  • Chip clock (top model): 950 MHz
  • 1250 MHz actual (5.00 GHz effective) memory, 160 GB/s memory bandwidth
  • 256-bit DDR memory interface (up to GDDR5)
  • 2048 MB (2 GB) memory amount standard
  • 2.9 TFLOP/s single-precision floating point compute power
  • 486 GFLOP/s double-precision floating point compute power
  • Elimination of Hotclocks

  

The GK104's performance is expected to exceed the GTX 580 at the $350 to $400 price range. In addition, it is expected to outperform AMD's HD 7950 at similar price point and challenge the HD 7970 for the performance crown. The GK104 looks to be the similar to the current generation GTX 560 Ti with regards to price to performance in its category. 

Please keep in mind, of course, that these specifications are from 3dcenter's supposed reliable source. We won't know for sure until Nvidia shows its hand. Stay tuned!

Discuss
Display all 66 comments.
This thread is closed for comments
Top Comments
  • 13 Hide
    EDVINASM , February 14, 2012 9:12 AM
    $%ing boring already. Give us GPU a real world benchmark suite to compare with existing GPUs. Not even a test product in the bench. We need one and need one now not in April. Unless gamers can hibernate the wait would simply ruin NVidia in short term. Am NVidia user but boy am I starting to shop around for AMD.
  • 12 Hide
    edwinjr , February 14, 2012 10:33 AM
    1000+ cores & no shader clock.
    Thats looks like amd gpu.
  • 10 Hide
    RazorBurn , February 14, 2012 8:48 AM
    It doesn't matter if the source is reliable or not, because whats important is that all of them tells there's a Huge improvement in the GPU..
Other Comments
  • 5 Hide
    sarcasm , February 14, 2012 8:47 AM
    billybobserGiven the lies (ahem, rumours) they were shilling out, the Nvidia was expected to beat the 7970 for half the price!I see that's been downgraded to 'challenge', which is wholly disappointing, seeing as AMD are milking people with their new pricing strategy, it doesn't look like nVidia are going to convert anyone.


    Who cares, I'm loving the competition. I love how they keep trying to out do each other over and over because the consumers end up winning. It's like Intel vs AMD except on a different front. When it comes to cpu recommendations, its always "Intel Intel Intel." But with GPUs, its a really huge toss up between the two which gives consumers more options but still able to get their money's worth.
  • 10 Hide
    RazorBurn , February 14, 2012 8:48 AM
    It doesn't matter if the source is reliable or not, because whats important is that all of them tells there's a Huge improvement in the GPU..
  • 10 Hide
    welshmousepk , February 14, 2012 9:03 AM
    I just hope there are plenty of chips to go around. When quantities are low, backwater countries like mine (new zealand) get totally ripped off with pricing. I recently upgraded my GPUs, and while I had really wanted a 7970, they were going for 1200 dollars here. I ended up just getting a GTX 580 for a little under 700. that's almost have the price. where's the logic in that?!
  • 8 Hide
    nohode , February 14, 2012 9:04 AM
    I just want the prices to drop since recently AMD 79xx pricing scheme which is bullsh1t
  • 13 Hide
    EDVINASM , February 14, 2012 9:12 AM
    $%ing boring already. Give us GPU a real world benchmark suite to compare with existing GPUs. Not even a test product in the bench. We need one and need one now not in April. Unless gamers can hibernate the wait would simply ruin NVidia in short term. Am NVidia user but boy am I starting to shop around for AMD.
  • 7 Hide
    viridiancrystal , February 14, 2012 9:23 AM
    Looks to my like each CUDA core is going to be much less powerful than on Fermi.
  • -3 Hide
    Anonymous , February 14, 2012 9:38 AM
    No, it seems the cores have similar performance with Fermi cores, only the frequency is different.
  • 1 Hide
    SchizoFrog , February 14, 2012 9:42 AM
    nVidia always said that their cards would launch in March/April and as such AMD rushed a couple of their cards out to be first for this current generation. Unless your PC just blew up, anyone that can't wait an extra few of weeks to see exactly where they stand is either an idiot or has far too much money on their hands.
    It would not surprise me if nVidia has a solid launch with a good mid range card and a high end card that just about tops what AMD has to offer. I bet Kepler will be able to do far more than initial cards will suggest though and we'll see many, many iterations of virtually the same cards with different clock speeds to cover various price points, much like the current 560/560Ti/560Ti(448)...
  • 1 Hide
    rmpumper , February 14, 2012 10:09 AM
    dragonsqrrl
    rmpumper7950/7970 should be priced ~$50+ of 6950/6970 prices. So as it is now, if nvidia's gtx680 will be better than 7970 they will price it at >$600? That's a load of crock.
    Every rumor and leak I've seen so far on gk104 pricing seems to indicate otherwise...http://www.guru3d.com/news/nvidia- [...] -299-230-/According to Nvidia's AIB partners the initial price set for the first gk104 based graphics card is $300. Of course this can go up or down based on the competition. Unfortunately, I have the feeling it'll be going up.


    So much for that then. $300 my ass.
  • 4 Hide
    Burodsx , February 14, 2012 10:15 AM
    I'm interested in seeing how the temperatures will compare.
  • 3 Hide
    Anonymous , February 14, 2012 10:29 AM
    User "mosu" has some faulty calculations there...

    It is actually more like:

    750 watts * 0,6 (from 40nm -> 28nm) * 0,63 (950 Mhz / 1500 Mhz) = 285 watts.

    You see, if you drop clockrate by one percent, power density drops three percent...

    So, when you drop clock from 1500 Mhz to 950 MHz, you can actually use about three times (3x) as much cores and still consume about the same amount of power.




  • 12 Hide
    edwinjr , February 14, 2012 10:33 AM
    1000+ cores & no shader clock.
    Thats looks like amd gpu.
  • -3 Hide
    photonboy , February 14, 2012 10:48 AM
    COOLING:
    I bought my HD5870 from AMD solely because the GTX4xx cards ran hot and loud. I've been very happy with it overall. I can even run many games at maximum settings (1920x1080 @ 60Hz) including the new Amalur game (awesome game).

    However, the die sizes seem similar between AMD and NVidia so cooling should be similar too.

    Therefore, it's likely I'll buy an NVidia GTX670 or GTX680. Now that PhysX can run without dropping a game below 60FPS that's a plus, but NVidia also is a bit better with the driver support.

    Games I put on HOLD waiting for a better card than the HD5870:
    - Witcher 2
    - Crysis 2 (High-Def pack)
    - Assassin's Creed Brotherhood
    - Metro 2033
    - Grand Theft Auto IV
  • -2 Hide
    ewood , February 14, 2012 10:53 AM
    mosusimple math: 512 CUDA cores=250watts 1536CUDA cores=750Watts, assuming that 28nm tech gives them a 40% reduction on power usage, will consume at least 500 watts...not feasible.


    750*(6/10)= 450 not 500
  • -1 Hide
    cronik93 , February 14, 2012 10:59 AM
    I can see consoles doing that around 2014.
  • 1 Hide
    maxinexus , February 14, 2012 11:14 AM
    Competitiveness is awesome. The greatest inventions were driven mostly by wars. So bring it on Nvidia we want price wars!!!
  • 9 Hide
    vitornob , February 14, 2012 11:36 AM
    mosusimple math: 512 CUDA cores=250watts 1536CUDA cores=750Watts, assuming that 28nm tech gives them a 40% reduction on power usage, will consume at least 500 watts...not feasible.


    You couldn't be more wrong.
    Simple math: 512 cuda core, each at 1544mhz or 1536 cuda core, each at 950mhz
    This last one would consumes 1,84 times the wattage, but with 40% reduction it would consume 1,1 times the wattage. With a few optimizations it will be easy to make it consumes the same or lower.

    Feasible.
  • 0 Hide
    alvine , February 14, 2012 11:42 AM
    dayum if their new flagship will have 1500 cuda cores and will not cost an arm and a leg I will deff upgrade from my two gtx 280 ( 240 cuda cores x2 lolll)
Display more comments