![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Rumor Mill: 55nm GTX 270/290 launching at the end of August
- Thread starter Xeth
- Start date
HagardProductions
Weaksauce
- Joined
- Mar 7, 2007
- Messages
- 78
Gonna get me a GTX 290 if this is true...its well within my step-up period. Thanks ATI for making NVIDIA put some even better cards out!
AthlonXP
Fully [H]
- Joined
- Oct 14, 2001
- Messages
- 20,588
so the 280 will just be called the 290 then???
Seems so. It's actually a nice naming scheme. Much better than 9800 GTX+. Or 8800GTS 512.
AthlonXP
Fully [H]
- Joined
- Oct 14, 2001
- Messages
- 20,588
True, but besides the die shrink, will there be an huge difference between the two?? It seems not the case.
Waffletron
Limp Gawd
- Joined
- Jul 22, 2008
- Messages
- 222
That is a much better naming solution, far better than the whole disaster with the 9 series. The 9 series SHOULD have been 8900s and 8700s, not 9800s and 9600s.
vick1000
2[H]4U
- Joined
- Sep 15, 2007
- Messages
- 2,443
Yeah, kinda lie the 8800GT magically became the 9800GT.
TheBluePill
2[H]4U
- Joined
- Aug 17, 2005
- Messages
- 3,773
With a die Shrink will come a Clock Bump.
the GTX 280 will perform slightly less than the 290, and the 270 will fit in-between
That is probably the logic behind the 20-point Difference in products, to allow a half-step while
keeping the older products performance viable products.
Hence 290>280>270>260
Versus the Old 8800GTXULTRA>8800GTX>8800GTS512>8800GT>8800GTS640>8800GTS320
Damn.. Insnt that just so much simpler for the consumer!!
the GTX 280 will perform slightly less than the 290, and the 270 will fit in-between
That is probably the logic behind the 20-point Difference in products, to allow a half-step while
keeping the older products performance viable products.
Hence 290>280>270>260
Versus the Old 8800GTXULTRA>8800GTX>8800GTS512>8800GT>8800GTS640>8800GTS320
Damn.. Insnt that just so much simpler for the consumer!!
known12345
Limp Gawd
- Joined
- Nov 17, 2007
- Messages
- 311
Man so now I have to wait for these cards to release to even consider buying a video card if I want a good deal. I mean ive never seen so many new graphics cards released in 1 year...
King of the graphics card hill comes ~18 months. Midrange graphics cards come every ~9 months. Just buy somethign and be happy. If your not running 1920x1200+ most of these cards will run everything.
Cliff Couser
[H]ard|Gawd
- Joined
- Dec 22, 2003
- Messages
- 1,860
+5 for speed bumps.
+10 for die shrinks.
+100 for new generation.
IMO, that makes more sense than +, X, GS, GT, GTS, GSO, XXXFG++1ity, etc.
+10 for die shrinks.
+100 for new generation.
IMO, that makes more sense than +, X, GS, GT, GTS, GSO, XXXFG++1ity, etc.
heatlesssun
Extremely [H]
- Joined
- Nov 5, 2005
- Messages
- 44,154
So it will be intresting to see what the clocks on these are. I'd be tempted to upgrade if they had clocks in the 720/1700/1300 range, but I have XFX cards so no step up. If I could get $150 to $200 for my 280's, plus hopefully $360 in rebates for my 280's and these cards came in at like $400 each, that would only be about $400 out of pocket, not too bad.
And of course, what about the fall refresh? Hard to see that happening with 9800 GX2, then three months and GTX 280/260 and then two months and GTX 290/270 and then another release in three months. But if AMD has the lead there could be something. Four top line releases in one year. Wow. Three is already a lot.
And of course, what about the fall refresh? Hard to see that happening with 9800 GX2, then three months and GTX 280/260 and then two months and GTX 290/270 and then another release in three months. But if AMD has the lead there could be something. Four top line releases in one year. Wow. Three is already a lot.
chinesepiratefood
2[H]4U
- Joined
- Oct 8, 2005
- Messages
- 2,629
Might step up if true from my 260, depends on power consumption(as a GTX 280 would be just about as much as my Corsair HX520 could handle), and added performance.
sweet I am eligible for step up for another 90 days, so two months after the supposed launch for these cards.
wasn't that time frame suppose to be the original launch time for the gtx 280 series, because I used to read that Nvidia would release the gt200 series in September, but they launched it two months early because of Ati, so may be the gt200 was meant to be launched at 55nm but they probably knew that the process wont be ready until august or so, and they had to launch it at 65nm to keep the competition with ati, it kinda makes sense now that I think about the old launch dates that I used to read about.
wasn't that time frame suppose to be the original launch time for the gtx 280 series, because I used to read that Nvidia would release the gt200 series in September, but they launched it two months early because of Ati, so may be the gt200 was meant to be launched at 55nm but they probably knew that the process wont be ready until august or so, and they had to launch it at 65nm to keep the competition with ati, it kinda makes sense now that I think about the old launch dates that I used to read about.
Sorry Heatless, I hereby ban you from even discussing GPU upgrades until you buy a damn 30 in LCD. A man with a godly system like yours should never even dream about gaming at anything other than 2560X1600.
Sorry Heatless, I hereby ban you from even discussing GPU upgrades until you buy a damn 30 in LCD. A man with a godly system like yours should never even dream about gaming at anything other than 2560X1600.
LOL that was a good one, I really agree with you on that, that is one reason I never go Sli for now because there is not a game out there that does not play at 1920x1200, that even includes crysis with everything high.
I am skeptical about this and it is possible that NVIDIA is spreading this rumor in order to detract the attention from the X2 launch.
Maybe the card will be paper launched in August but it is my opinion that volumes will come at least 4 or 6 weeks later.
Maybe the card will be paper launched in August but it is my opinion that volumes will come at least 4 or 6 weeks later.
Simpson5774
2[H]4U
- Joined
- Feb 1, 2005
- Messages
- 2,937
When was the last time nvidia paper launched a product costing over 200? if it wasn't for nvidia, ATi would still be paper launching products which don't appear for 2-3 months.
Anything is possible however, except a paper launch.
hopefully it will be w/ in my step up period, but if it's not a significant boost, I might jump on the x2 train
if u look at the 9600GT+ and 9800GTX+... u will get an idea of how a gtx290 would perform.
bout the same
Simpson5774
2[H]4U
- Joined
- Feb 1, 2005
- Messages
- 2,937
One rumor which for some reason died off the front of the video card forum said something along the lines of nvidia added transistors to this chip to aid in the clocking of the shaders aka = more cowbell.
I don't like reading babbelfish translated websites, that claim they contain the answers, because 90% of them don't, or misheard it, or plain made it up.
Looking at the gtx+ you would expect a 17% increase. I wouldn't call that about the same.
The 9800 GT is not strickly a 8800 GT. Fab process aside, the 9800 GT also has OpenGL 2.1 support. The 8800 GT only has 2.0.
If you think NVIDIA is losing money with GTX 200 cards, think again. They are however, not making as much money as I'm sure they would like to, due to competition from AMD/ATI.
alg7_munif
Supreme [H]ardness
- Joined
- Oct 9, 2006
- Messages
- 5,862
9800GTX+ was late, why would the GTX 290 be on target?
heatlesssun
Extremely [H]
- Joined
- Nov 5, 2005
- Messages
- 44,154
Since nothing has been announced yet, there is no target. I guarantee with the current market, that if and when nVidia announces a date for these things, they'll be there. They are in a hole, they don't need to make it bigger by announcing more dates and missing them. It would be better to stay quite.
GTX+ wasn't late. They announced it before they launched it. They never released a release date.
The 9800GTX+ was a paper launch, so its not like Nvidia is out of the business of doing such moves.
That said, clocks will depend on whether Nvidia fixed the biggest issue with higher clocked GT200 cards: the shader domain. If its fixed, we should expect higher clocks. If not, then don't expect too big of a change since 55nm is an optical shrink, primarily for saving money, as actual transistors aren't going to get a big boost like say 65nm to 45nm, a real node shrink.
That said, clocks will depend on whether Nvidia fixed the biggest issue with higher clocked GT200 cards: the shader domain. If its fixed, we should expect higher clocks. If not, then don't expect too big of a change since 55nm is an optical shrink, primarily for saving money, as actual transistors aren't going to get a big boost like say 65nm to 45nm, a real node shrink.
not losing money ...true... but they not making money hand over fist as they were. 55nm will help with the bottom line but thats about it
Depends on how much they changed in GT200b. Bear in mind that they went from launch to new launch in 2 months so most things were set in stnoe before GT200 even hit the streets. So unless they knew beforehand GT200 was underperforming or that ATI would be very competitive, which I doubt, then GT200b was a side project meant for saving money long ago, since at least they knew from early on that a massive die would cut into profit margins.
How many changes/performance increases we should expect, I don't know.
But for god's sake don't name it with a "+" please
How many changes/performance increases we should expect, I don't know.
But for god's sake don't name it with a "+" please
heatlesssun
Extremely [H]
- Joined
- Nov 5, 2005
- Messages
- 44,154
I have a felling that nVidia had good idea that the GTX 280/260 were not going to be performance vs. price competitive for a good while before launch. If Charlie D. at the INQ knew, then nVidia had to have some idea.
No doubt a major piece of the puzzle is price in going to 55nm, but another piece is to no doubt to have a great part. The key to this is going to be clocks. If the core and shader clocks of a standard clocked GT200b can reach 750/1700 and it come in at $400, this is going to be a very competitive card. At 65nm, nVidia couldn't reach the clocks it wanted and since some factory overclocked GTX 280's are already in this ball park I don't think that this huge stretch, we shall see.
My prediction. A $400 dollar GTX 290 that's within 10% overall of a 4870x2. If the 4870x2 comes in at $500 even $450, the GTX 290 or whatever its called could be a very compelling part.
I just don't see nVidia launching something that doesn't raise an eyebrow or two as the GTX 280/260 launch was so bad. We shall see. At any rate, I was ready to jump on a 4870x2, I'll probably wait a bit and see what the end of August looks like.
Well strangely enough there aren't any codenames out there for the cut-down GT200 parts unlike G80 which everyone knew had a G84/G86 to follow, so I don't know what derivatives are down the line. Certainly a GT300 seems down the road though.
Do keep in mind that a week before the GT200 launch, Nvidia was quoted as to saying that they expected G92 architecture to last another year. Turns out that even a 55nm refresh isn't quite enough without following the 3800 model road and competing by pricing the cards well under initial prices.
Interestingly enough, the next generation will put both sides at even footing in terms of process since neither side has a fab for video cards, so they are at the mercy of TSMC ramping up its own processes. Previously, ATI had the node process advantage as they were always one step ahead of Nvidia in terms of process (80nm vs. 90nm, 55nm vs.65nm etc.) but now that die processes are at their limits, Nvidia can catch up.
Methinks that architectural efficiency is now on the priority list of both sides given that:
a) Die processes are equal on both sides now, so one cannot count on a process advantage for performance
b) Whoever gets the most out of the same process is likely to win
So the entire ATI statement that monolithic GPUs are dead is not that far out after all... and I'm sure after the pricing debacle with GT200 and the relatively poor performance increase vs. actual size increase, Nvidia is paying attention as well.
It depends on how much Nvidia knew about ATI cards and whether their cards lived up to their own expectations though. Bear in mind Nvidia could've seen Charlies article and thought it was the same spiel he said about R600 vs. G80 and turned a blind eye, or Nvidia had other priorities. Having a 55nm project in itself is no sure thing that they thought it was a failure - after all, 55nm G92 wasn't supposed to be released unless Nvidia thought it was necessary. AFAIK in terms of how Nvidia is structured, they have multiple teams on multiple projects, not all of which see the light. A die-shrink team assigned to each card release is certainly a likely part of how Nvidia works, and because the next process of 45nm/40nm isn't ready until the beginning of next year, and because an optical shrink from 65nm to 55nm is easy compared to a real die shrink (65nm to 45 or 40 nm), I can see work on GT200b as having been in the works early on.
Do keep in mind that a week before the GT200 launch, Nvidia was quoted as to saying that they expected G92 architecture to last another year. Turns out that even a 55nm refresh isn't quite enough without following the 3800 model road and competing by pricing the cards well under initial prices.
Interestingly enough, the next generation will put both sides at even footing in terms of process since neither side has a fab for video cards, so they are at the mercy of TSMC ramping up its own processes. Previously, ATI had the node process advantage as they were always one step ahead of Nvidia in terms of process (80nm vs. 90nm, 55nm vs.65nm etc.) but now that die processes are at their limits, Nvidia can catch up.
Methinks that architectural efficiency is now on the priority list of both sides given that:
a) Die processes are equal on both sides now, so one cannot count on a process advantage for performance
b) Whoever gets the most out of the same process is likely to win
So the entire ATI statement that monolithic GPUs are dead is not that far out after all... and I'm sure after the pricing debacle with GT200 and the relatively poor performance increase vs. actual size increase, Nvidia is paying attention as well.
It depends on how much Nvidia knew about ATI cards and whether their cards lived up to their own expectations though. Bear in mind Nvidia could've seen Charlies article and thought it was the same spiel he said about R600 vs. G80 and turned a blind eye, or Nvidia had other priorities. Having a 55nm project in itself is no sure thing that they thought it was a failure - after all, 55nm G92 wasn't supposed to be released unless Nvidia thought it was necessary. AFAIK in terms of how Nvidia is structured, they have multiple teams on multiple projects, not all of which see the light. A die-shrink team assigned to each card release is certainly a likely part of how Nvidia works, and because the next process of 45nm/40nm isn't ready until the beginning of next year, and because an optical shrink from 65nm to 55nm is easy compared to a real die shrink (65nm to 45 or 40 nm), I can see work on GT200b as having been in the works early on.
IMO you're crazy to think that this has anything to do with ATI. If NV was taken by surprise at the performance of ATI's 4870 release, they will not have progressed through research, engineering, development, testing, driver development, and tooling up their manufacturing partners in this little time.
If true, the 270/290 (under some name) were already on the roadmap regardless of ATI's 4870.
If not on the roadmap, certainly in the contingencies of cards that might get released if sh!t hits the fan.
That said, 2 months is a very short time for any massive redesign unless they detected failure ahead of time, which seems unlikely given how much they hyped the can of whoop ass they had that turned out to be tonic water.
heatlesssun
Extremely [H]
- Joined
- Nov 5, 2005
- Messages
- 44,154
Remember that this card was designed long ago, so nVidia is just pushing the schedule, this probably wouldn't have come out until November had the GTX 280/260 been more competitive.
At any rate the key to how well this goes down is all about the clocks. If nVidia can get them high enough at a $400 price point, this could be a damper on the 4870x2.
You know this how? All I've seen is rumours about how the gtx 2xx are selling bad and that nvidias partners are pissed of about forced price cuts.
Links? From what i've seen a 9800gtx+ performs about the same as a overclocked 9800gtxLooking at the gtx+ you would expect a 17% increase. I wouldn't call that about the same.
x2!!!!!