Kepler Before Xmas

[ocre]

Dont shoot the messenger!

http://www.tweaktown.com/news/21432...epler_gpus_to_hit_before_christmas/index.html

I have actually heard this might happen a little while ago and was wondering why we havent seen at least one place saying it. This is the first time i seen it in an article. But what i want to remind everyone about is the 580 surprise. Nvidia was very proud of that effect and would sure like to reproduce it. This is the way I imagined it when i first heard of the possibility.

Anyway, what do you guys think, plausible? Is there any chance of kepler before xmas?

 

[Annon2255]

Interesting... I was reading that the 7000 series could be pushed back to February 2012. This would be a good time for NVidia to strike if that's the case.

 

[ocre]

I actually heard of this nearly 2months ago, i was waiting to see if any
other leaks would come out. By yesterday i started doubting the info. Today i seen it....finally!

It can still go either way. I know that Nvidia wanted the element of surprise again. Anything couldve went wrong but i know they were close. there at least is a chance this could happen.

 

[Genx87]

I would be surprised. But the 500 series cards showed up pretty quick out of nowhere. We suspected they may show up before christmas last year. But no concrete info until about 2 weeks before the launch. Even then many didnt believe Nvidia could get out the 500 series that quick.

 

[TheUnk]

Would be awesome. For the first time ever I could use the EVGA StepUp program haha

 

[thilanliyan]

I'm sure they can whip up something with a bunch of 2 by 4s by then

 

[ocre]

I'm sure they can whip up something with a bunch of 2 by 4s by then

freak'n LoL!!

 

[tviceman]

I seriously doubt they will get any high end Kepler parts out this year. The rumors have been pointing to die shrinking GF116 and GF118 for notebooks (the gtx560m and certain gt555m gpu's are powered by GF116), which make the most sense becaues GF116 and GF118 are smaller sized chips and are known quantities which would make manufacturing them on 28nm easier and quicker than new designs with big cores, providing engineers with insight and technical knowledge of the 28nm process as soon as possible.

There was the story a while back on semiaccurate (gotta take it with a grain of salt though because Charlie spins everything Nvidia in a negative light) and Fudzilla that said Nvidia had taped out certain Kepler parts. If true, I still think it will be one of the smaller derivatives - despite Nvidia's track record of launching the big chip first - just because of how many problems they had getting GF100 initially out of the door. Make the smaller part first, fix it's problems which will be fewer than the bigger part, and scale it up to size.

I'd bet anyone here that we don't see 28nm cards this year that can beat out a gtx560ti / hd6950.

 

[atticus14]

i would love to see both cards keeping prices in check but they'll like be a supply:demand price premium anyway for awhile...but at least it'll eventually calm down.

The only thing i wonder about all this is whats going to be price points for all these cards and how will it effect previous gen cards as well.

 

[Idontcare]

Nvidia launching before AMD would violate the Temporal Directive or something, wouldn't it? I thought it was verboten by the laws of physics.

 

[boxleitnerb]

Nvidia launching before AMD would violate the Temporal Directive or something, wouldn't it? I thought it was verboten by the laws of physics.

Why? Just because it happened once with the HD5000 series? Last I checked Nvidias GT200 was out a week before HD4000 and 8800GTX was out very early compared to HD2000. 6800 Ultra was launched a tad before X800XT PE. Rest of the launches were a bit out of sync.

I don't see anything out of the ordinary here.

 

[ocre]

I seriously doubt they will get any high end Kepler parts out this year. The rumors have been pointing to die shrinking GF116 and GF118 for notebooks (the gtx560m and certain gt555m gpu's are powered by GF116), which make the most sense becaues GF116 and GF118 are smaller sized chips and are known quantities which would make manufacturing them on 28nm easier and quicker than new designs with big cores, providing engineers with insight and technical knowledge of the 28nm process as soon as possible.

There was the story a while back on semiaccurate (gotta take it with a grain of salt though because Charlie spins everything Nvidia in a negative light) and Fudzilla that said Nvidia had taped out certain Kepler parts. If true, I still think it will be one of the smaller derivatives - despite Nvidia's track record of launching the big chip first - just because of how many problems they had getting GF100 initially out of the door. Make the smaller part first, fix it's problems which will be fewer than the bigger part, and scale it up to size.

I'd bet anyone here that we don't see 28nm cards this year that can beat out a gtx560ti / hd6950.

Was these rumors of a 28nm kepler/fermi hybrid? I heard a bird singing. told me of a chip in the works that isnt quite a full kepler but more of a fermi/kepler mix. Kinda sharing traits from both. I heard this around the same time and it could be related to cards that show up before xmas. Not sure of the application of this hybrid if its true. There is a good chance of it being true in my view.

But what it will be used for? Perhaps there will be die-shrink fermi cards with mix of kepler in them. Or perhaps this is a mobile design.

Its interesting to me. What if the hybrid is Nvidia's step to a full kepler. Sort of like the 6800 series. I am not sure, but that would be cool. I do believe there is a hybrid in the works none the less, what it could be used for is anyones guess

 

[tweakboy]

Dont shoot the messenger!

http://www.tweaktown.com/news/21432...epler_gpus_to_hit_before_christmas/index.html

I have actually heard this might happen a little while ago and was wondering why we havent seen at least one place saying it. This is the first time i seen it in an article. But what i want to remind everyone about is the 580 surprise. Nvidia was very proud of that effect and would sure like to reproduce it. This is the way I imagined it when i first heard of the possibility.

Anyway, what do you guys think, plausible? Is there any chance of kepler before xmas?

Heck no, the source is bunk. I would be shocked if nVidia sells kepler in 2011. Their 2011 chip was 580 or 590. Let them milk it a bit. It will be released my guess is Feb or March. gl,
I know for fact as do many, that ATI/AMD 7xxx series is coming out by December 2011. That might get delayed to Jan or Feb. Soo nVidia and AMD are close to each other with releases. But nVidia I noticed always lets ATI come out with their power card,, then nVidia answers back.. as they will answer back with Kepler next year. gl

 

[ArchAngel777]

Why? Just because it happened once with the HD5000 series? Last I checked Nvidias GT200 was out a week before HD4000 and 8800GTX was out very early compared to HD2000. 6800 Ultra was launched a tad before X800XT PE. Rest of the launches were a bit out of sync.

I don't see anything out of the ordinary here.

He was kidding...

 

[bunnyfubbles]

Why? Just because it happened once with the HD5000 series? Last I checked Nvidias GT200 was out a week before HD4000 and 8800GTX was out very early compared to HD2000. 6800 Ultra was launched a tad before X800XT PE. Rest of the launches were a bit out of sync.

I don't see anything out of the ordinary here.

nope, ATI/AMD has almost always beaten nVidia to a new fabrication process. Yes, nVidia has beaten ATI/AMD to the punch when it comes to releasing a new generation of GPU and a higher end part, but I can't remember when they've been first to a new fabrication process.

the GT200s were 65nm when ATI had long been on 55nm with the 3000 series, and was why the 4000 series was such a huge blow to nVidia despite ATI being unable to claim the performance crown, they had such a huge thermal and price/performance advantage that it didn't matter.

Same went for the 8800 GTX (90nm) vs. the 2900 (80nm). And again for the 6800Ultra only 130nm, while ATI was first to 110nm.


So unless nVidia releases a 40nm version of Kepler (which would pretty much nullify any chance that it will be an earth shattering performance part) it would mean they'd be first to the 28nm party and first with a new fabrication process in forever.

 

[bryanW1995]

I seriously doubt they will get any high end Kepler parts out this year. The rumors have been pointing to die shrinking GF116 and GF118 for notebooks (the gtx560m and certain gt555m gpu's are powered by GF116), which make the most sense becaues GF116 and GF118 are smaller sized chips and are known quantities which would make manufacturing them on 28nm easier and quicker than new designs with big cores, providing engineers with insight and technical knowledge of the 28nm process as soon as possible.

There was the story a while back on semiaccurate (gotta take it with a grain of salt though because Charlie spins everything Nvidia in a negative light) and Fudzilla that said Nvidia had taped out certain Kepler parts. If true, I still think it will be one of the smaller derivatives - despite Nvidia's track record of launching the big chip first - just because of how many problems they had getting GF100 initially out of the door. Make the smaller part first, fix it's problems which will be fewer than the bigger part, and scale it up to size.

I'd bet anyone here that we don't see 28nm cards this year that can beat out a gtx560ti / hd6950.

Yeah, best we're gonna get is LP 28nm this year.

nope, ATI/AMD has almost always beaten nVidia to a new fabrication process. Yes, nVidia has beaten ATI/AMD to the punch when it comes to releasing a new generation of GPU and a higher end part, but I can't remember when they've been first to a new fabrication process.

the GT200s were 65nm when ATI had long been on 55nm with the 3000 series, and was why the 4000 series was such a huge blow to nVidia despite ATI being unable to claim the performance crown, they had such a huge thermal and price/performance advantage that it didn't matter.

Same went for the 8800 GTX (90nm) vs. the 2900 (80nm). And again for the 6800Ultra only 130nm, while ATI was first to 110nm.


So unless nVidia releases a 40nm version of Kepler (which would pretty much nullify any chance that it will be an earth shattering performance part) it would mean they'd be first to the 28nm party and first with a new fabrication process in forever.

I wonder how this will change going forward now that Killebrew is gone. Is AMD getting back into the "high performance, large die area" gpu arena after years of settling for Nvidia's scraps? Maybe they saw what Maxwell was going to bring to the table and realized that they needed a massive overhaul now while there is still time.

 

[tviceman]

I wonder how this will change going forward now that Killebrew is gone. Is AMD getting back into the "high performance, large die area" gpu arena after years of settling for Nvidia's scraps? Maybe they saw what Maxwell was going to bring to the table and realized that they needed a massive overhaul now while there is still time.

I find it amusing (and interesting) that while AMD has long held the performance per mm^2 crown, they are now more actively pursuing the same strategy as Nvidia - that is, the high end part is a multipurpose chip. Since AMD started actively implement this strategy with Cayman, the performance per mm^2 dropped off (vs. Cypress), and as they continue to move more toward the multipurpose chip strategy, it will be interesting to see what their die sizes and performance per mm^2 end up compared to Nvidia GPU's.

 

[s44]

I wonder how this will change going forward now that Killebrew is gone. Is AMD getting back into the "high performance, large die area" gpu arena after years of settling for Nvidia's scraps?

If they do, it won't be for us, but for the professional/HPC market.

But even the rumored new directions (esp. mobile) are trying to follow Nvidia anyway! I'm not sure how AMD can get out of the box they're in.

 

[tviceman]

There was the story a while back on semiaccurate (gotta take it with a grain of salt though because Charlie spins everything Nvidia in a negative light) and Fudzilla that said Nvidia had taped out certain Kepler parts. If true, I still think it will be one of the smaller derivatives - despite Nvidia's track record of launching the big chip first - just because of how many problems they had getting GF100 initially out of the door. Make the smaller part first, fix it's problems which will be fewer than the bigger part, and scale it up to size.

I'd bet anyone here that we don't see 28nm cards this year that can beat out a gtx560ti / hd6950.

http://www.anandtech.com/show/5064/nvidia-geforce-gtx-580m-receives-major-price-cut
Just saw this on Anandtech's front page. It doesn't confirm anything, but it leads credence to the theory of new mobile parts coming soon.

 

[sontin]

Same went for the 8800 GTX (90nm) vs. the 2900 (80nm). And again for the 6800Ultra only 130nm, while ATI was first to 110nm.

NV40 and r420 were both on 130nm.

So unless nVidia releases a 40nm version of Kepler (which would pretty much nullify any chance that it will be an earth shattering performance part) it would mean they'd be first to the 28nm party and first with a new fabrication process in forever.

nVidia changed their direction. Their were very aggressive with 40nm. AMD announced and sold the first 40nm GPU to the end user it was nVidia which archived the 10 millions mark in january 2010, while AMD sold only ~3,5 millions 40nm products. nVidia had 4 different products on 40nm: Tegra 2 (the first 40nm SoC), MCP89 (the first 40nm chipset), GT21x and Fermi in the first half of 2010.

 

[tweakboy]

Knowing nVidia and how they delay. I see kepler coming after 7xxx.

7xxx will be out in December / January. nVidia kepler will come Feb or March. gl

 

[Chinoman]

I remember reading that the initial line of 28nm Nvidia chips would be the entry-level mobile cards aimed at OEM's. Somewhat like the 300 series.

 

[SirPauly]

Nvidia launching before AMD would violate the Temporal Directive or something, wouldn't it?

I don't know if nVidia follows Star Fleet protocols and didn't take part of the Temporal Accords. I really think Jen-Hsun Haung may be Suliban.

 

[notty22]

I don't know if nVidia follows Star Fleet protocols and didn't take part of the Temporal Accords. I really think Jen-Hsun Haung may be Suliban.

Or a Changeling Founder ?

 

[MentalIlness]

7870 / 7850 specs at the bottom ? Possible ?