NVIDIA GeForce GTX 880 and GTX 870 to Launch This Q4

[Roel]

The technology keeps getting more energy efficient so I don't know what you're complaining about. If you want low power consumption then you can go for a 750 Ti which has a TDP of 60W and performs about the same as a high-end gaming laptop and the 4-year old GTX 470 which had a TDP of 215W at that time. It can play all games if you just turn the settings down a bit. The top cards are for those who want maximum performance and don't care about the power consumption.

I am building a water cooling loop with lots of radiators so I won't have any trouble with the noise. The heat will keep me warm so the central heating system doesn't have to work as hard. I would buy a 500W card if the performance was double that of a 250W card, it would be better than SLI.

 

[Octavean]

Meh,....

I just want a cheaper GTX 760 or GTX 770,....

I'm on a GTX 670 now and I find it fine for my needs but I need to upgrade another system.

 

[debs3759]

Hmm, might be time to upgrade that Matrox Mystique...

 

[TheGuruStud]

I think the much simpler explanation is the one that Cadaveca posted at the last leak. The different SKUs are getting mixed up and 3200SP and 8GB is for a dual-GPU card, the successor to GTX 690. The single GPU part, successor to the GTX 680/GTX 770 would therefore have 4GB and 1600SP. To me, this is much more reasonable.

Remember, GTX 750 Ti outperforms the GTX 660 Ti by 20% and yet it has 20% fewer shaders, so assuming the same scaling, a 1600SP GTX 880 would have almost 50% more performance than GTX 770/680, completely in line with a generational improvement.

No. Just no.

 

[The Von Matrices]

I know, right? It made sense for GTX 560 Ti (GF114) to have 384sp, and 256-bit/32 ROP. A chip with four times the cores (1536sp) with 256-bit/32 ROP is so totally unimaginable! NVIDIA would never make such a chip.

Oh wait...it did. The GK104.

Ermagerd...3200 SP and 256-bit/32 ROP? Totally unimaginable and borderline blasphemous!

If we had no idea what the Maxwell architecture was, then I could see your point. But since we have GM107 and we know it has fewer shaders than GK106 (for more performance), then it would be a very unexpected move for GM104 to have double the shaders of GK104.

No. Just no.

Care to explain your reasoning?

 

[TheGuruStud]

If we had no idea what the Maxwell architecture was, then I could see your point. But since we have GM107 and we know it has fewer shaders than GK106 (for more performance), then it be a very unexpected move for GM104 to have double the shaders of GK104.



Care to explain your reasoning?

750 is not competition for a 660 nor faster than a 660.

 

[The Von Matrices]

750 is not competition for a 660 nor faster than a 660.

You're right in that I stated the wrong name of the card, I meant a GTX 650 Ti. However, I was still using the correct numbers when comparing it to a GTX 750 Ti, so the math is still valid.

 

[Deleted member 24505]

And you know this because you asked every person who owns such a config right?
*Removed (language)*
If you are just browsing the internet or any of those other stuff you can and will do on a pc other then gaming which barely require a gpu of any kind, you dont want your pc to use up tons of electricity.

I know this because they have 2x 780ti, not exactly low powered. I think it's pretty obvious anyone who has 2 of these is gonna have a pretty high end rig with at least a 1kw PSU. So do you really think they care about power usage? if they did they would not have such a high end rig, or maybe they just use it for browsing the internet or any of those other stuff you can and will do on a pc other then gaming. or maybe you are just a *Removed (language)*

 

[LeonVolcove]

No Upgrade for me, i am still satisfied with my current rig

 

[bloodyriders]

the spec still rumored
you guys forgot about 750ti? the rumor said if it have to be between 660 and 660ti with kepler architecture
but when 2 weeks before released (CMIIW)?
it goes all wrong
so, rumor from nvidia spec = not trusted

 

[Prima.Vera]

I don't think you'll see good 4K support until 2015

Make that 2016 and realistically speaking 2017.

Long have been the times when you could have a 75%-90% performance increase over previews generation (3870-4870-5870 anyone?).
Now we should be lucky if there is a stunning 25% increase.

 

[ZoneDymo]

Make that 2016 and realistically speaking 2017.

Long have been the times when you could have a 75%-90% performance increase over previews generation (3870-4870-5870 anyone?).
Now we should be lucky if there is a stunning 25% increase.

Seems unlikely, the R9 295x2 does a fine job at 4k, if the next AMD card (r9 380x?) single card is as fast as that one we should be well on our way

 

[FreedomEclipse]

well if this brings the price of 770s or 780s down then im all for it. I need some 3 or 4GB cards to drive my 1440p monitor

 

[vagxtr]

will be based on the brand new "GM204" silicon, which most reports suggest, is based on the existing 28 nm silicon fab process (...) When 20 nm is finally smooth, it wouldn't surprise us if NVIDIA optically shrinks these chips to the new process, like it did to the G92 (from 65 nm to 55 nm). The GM204 chip is rumored to feature 3,200 CUDA cores, 200 TMUs, 32 ROPs, and a 256-bit wide GDDR5 memory interface. It succeeds the company's current workhorse chip, the GK104.

Right. Onec again nicely rigged numbers make itself a news. How about that rumor site tells us how SweClockers would manage desing overstuffed chip with 2.1× times more shaders over GK104 on same 28nm processing node and stick only 200TMUs to accompany it with on measly 256b bus. Maxwell might be great as GM107 already show us that, but only thats done mostly by reconfiguration of avaliable resources so we can enjoy quite massive gaming improvenets over Kepler. And really, nobody even try to speculate on tha. It's always some rigged bigger numbers which arent possible when nVidia already has pretty big die to start with.

 

[vagxtr]

I have to disagree with you here. 20nm isn't going to be less expensive than 28nm per transistor, so there's no financial incentive for a die shrink and thus it won't be done. It makes more financial sense to sell a large 28nm chip than a smaller 20nm chip.

20nm will only be for the extreme high end this generation and will only be used in cases where it's impossible to manufacture a larger 28nm chip (e.g. you can't make a 28nm, 15 billion transistor, 1100mm^2 GM100). 20nm won't become mainstream until NVidia (or anyone else) can't achieve their performance targets on 28nm, which likely will not happen until the generation after this.

Actually 20nm according to same jumping jack flash tsmc rules would never become "mainstream" at all, it's just available node for a year or so , and we all expect die shrinks from eagerish AMD to implement it with real GPUs already fabbed on it. nVidia bets all their cards on overpromised but probably quite ready 16nm/14nm FinFet HKMG TSMC node which is ready for test since early this year. So if they really had Maxwell done and working they might just delay it and jump on more promising riskier 16/14nm node. Din't anybod ask themselves why rebrand of HD7000 series last year when they had available that obviously buggerishly plagued 20nm node? Which is "production ready" since April 2013 or so!

 

[vagxtr]

I know, right? It made sense for GTX 560 Ti (GF114) to have 384sp, and 256-bit/32 ROP. A chip with four times the cores (1536sp) with 256-bit/32 ROP is so totally unimaginable! NVIDIA would never make such a chip.

Oh wait...it did. The GK104.

Ermagerd...3200 SP and 256-bit/32 ROP? Totally unimaginable and borderline blasphemous!

You should learn more about those two totally different chip design approaches. This thing is pretty much whatCUcallit "blasphemous" as we already previewed Maxwell architecture in GM107 and it really didn't BOOST NUMBERS OF SPs so we can really "wildly guess" that this ain't approach nVidia position themselves with whole Maxwell lineup as long as we talk SAME NODE HERE (hint: 28nm)

 

[AnticiudadanoNumer1]

I dont know what to do, If to buy a gtx 780 GHz Edition or wait to GTX 800 series, and meanwhile buy a 750 Ti, I dont dont know what to do, I confused, if someone can help me with this existential question.

 

[Prima.Vera]

Seems unlikely, the R9 295x2 does a fine job at 4k, if the next AMD card (r9 380x?) single card is as fast as that one we should be well on our way

Read again what I've wrote. Your logic applies only if the next generation cards are 100% faster than the previews one (295x2 is a dual GPU btw...)

 

[RejZoR]

They usually are, though not always. The HD4800 to HD5800 transition was awesome, because the boost was actually 100%. I wish every gfx iteration would be in such a way. Otherwise we pay twice the price for 30% bumps that are almost an insult to gamers.

 

[GhostRyder]

They usually are, though not always. The HD4800 to HD5800 transition was awesome, because the boost was actually 100%. I wish every gfx iteration would be in such a way. Otherwise we pay twice the price for 30% bumps that are almost an insult to gamers.

That's just probably how its going to remain for the foreseeable future sadly. On all fronts the walls are starting to slow us down to crawls when it comes to making things bigger, faster, stronger, (etc joke). CPU's barely get faster, GPU's jump by small increments, its just the way things are going to keep going. Likely things will only change once they start working on power consumption and getting it down then eventually throwing that out the window and turning the cards loose. Or maybe there is something coming that will cause a huge jump, who knows in reality...

 

[Prima.Vera]

That's just probably how its going to remain for the foreseeable future sadly. On all fronts the walls are starting to slow us down to crawls when it comes to making things bigger, faster, stronger, (etc joke). CPU's barely get faster, GPU's jump by small increments, its just the way things are going to keep going. Likely things will only change once they start working on power consumption and getting it down then eventually throwing that out the window and turning the cards loose. Or maybe there is something coming that will cause a huge jump, who knows in reality...

Is the limit of technology really. They are almost at atomic level and there is no technology yet to further decrease the printing of circuits components on/for CPUs/GPUs. Adding extra transistors is not going to work good either, because the good yields percentage will decrease even more, meaning even higher price for a marginal only better product.

Let's face it, the semiconductor industry is getting stuck and no good things on the horizon unfortunately.

 

[ZoneDymo]

Read again what I've wrote. Your logic applies only if the next generation cards are 100% faster than the previews one (295x2 is a dual GPU btw...)

I know what you wrote and I know the 295x2 is a dual gpu, I did not say R9 380x for nothing, that would be the logical name for their new single gpu card following the current R9 280x, the next dual gpu should be called R9 395x2.

Anywho for a long time the next gen single gpu high end card from AMD was as fast as the previous gen dual gpu card.
If that streak is continued now, with the current R9 295x2 doing quite well at 4k, the next gen should be quite well on its way to be 4k ready.

 

[heydan83]

Alright, I don't expect any miracles then. Same process, but more cores? It's just Kepler with 400 more cores on a slightly more energy efficient architecture. So they might deal with the heat increase by adding more cores by using the slightly more efficient archi, and in turn gain a small performance increase from 2880 cores to 3200. I'm assuming the 870 will have ~3000 cores to hit a price point between the two.

Call me cynical, but I don't see the 780ti lowering in price and the 880 taking its place. The 880 is going to hit a higher price point. Then there's the simple fact that the 860 is probably going to just be a rebranded 780ti and everything else below will likely be a rebrand too. Ugh... new GPU releases are so disappointing these days... nothing to get excited about, especially when you know the price gouging is imminent.

I think is too optimist think that the 860 will be a re-branded 780ti, I think maybe if thinks go well it would be the 870 the re-branded 780ti and we only will have the 880 and 880ti as new chips, hopefully not, but I understand that neither Nvidia or AMD trust TSMC to deliver a large IC in commercial quantity so this generation would be like "in the meantime while we wait".