Will there be a new generation/refresh this year?

Joined
Jul 31, 2005
Messages
997
Will there be a $199 8800GS that performs as well as the 8800GTX? Will there be an 8900GTX? Will there be a 9800GTX(next-gen)? I am talking about this year of course.

Did the shittyness of ATi/AMD fuck up the cycle and screw the customers over?
 
yes!

eventually i'll put the side panel back on my case too!

damn summer!

*shakes fist*

perhaps... on october 12th!
 
Not to offend, but this thread is pointless. You will /know/ when anyone here even thinks they know about anything you're asking about. Probably a month+ before there's any real data. This forum will /flood/ with rumor threads and projected specs. 3/4+ will be flat out wrong. Some will get parts right.

Until performance numbers are in we won't /know/ anything. The 2900XT/XTX was supposed to be great, and the spec sheet supported that. The performance numbers are in, and it flat out cannot keep up with what is now almost year old hardware.

Don't waste your time theorizing or listening to the rumor mill. Make up your mind once you see a bunch of benchmarks and a direct comparison with the previous generation. Anything less is worthless.
 
There will be a 65nm part from Nvidia before the end of the year. There's no point in waiting, now that ATI has shown TSMC's 65nm yields are decent.

If the rumors are correct, the card will be an 8800 GS and have performance bridging the huge gap between the 8600 and 8800 GTS. This is the most logical refresh move Nvidia could make.

Beyond that, we know nothing.
 
Kyle dropped a hint recently, as is somewhat customary for him to do.

The bottom line? I believe the answer is "yes".
 
As far as i know the next (upgrading) refresh for the 8800 Series is the g92 in q1 08.
 
http://www.beyond3d.com/content/news/230

NVIDIA confirms Next-Gen close to 1TFlop in 4Q07
Wednesday 23rd May 2007, 12:12:43 PM, written by Arun


In recent analyst conferences that were publicly webcast on NVIDIA's website, Michael Hara (VP of Investor Relations) has claimed that their next-generation chip, also known as G92 in the rumour mill, will deliver close to one teraflop of performance.
In a separate answer to an analyst's question, he also noted that they have no intention from diverging from the cycle they have adopted with the G80, which is to have the high-end part ready at the end of the year and release the lower-end derivatives in the spring.

Assuming that NVIDIA manages to hit these aggressive release schedules, it implies that the chip will compete with any potential R6xx refresh at the beginning of its lifetime, but also eventually with R700 as it seems unlikely NVIDIA will refresh again before the second half of 2008,
unless they go for an optical shrink from 65nm to 55nm. It also remains to be seen how aggressive ATI will be on the process front this time around.

There also were a number of other highlights during the conference, including a major emphasis on GPGPU (aka 'GPU Computing') and a short mention of Intel's upcoming GPU efforts through their Larrabee project. Micahel Hara seemed far from certain about Intel's exact strategy there,
although he did mention that it was possible Intel was more interested in the GPGPU market than the gaming one. This is something we have already said in the past.

And finally, he mentioned that although he does not believe R600 will have any impact on their G80 sales, RV610 and RV630 are much more competitive parts that are likely to gain traction in the marketplace. He argued that he was not convinced 65nm gave AMD a real advantage in terms of costs
because of the yield curve, and seemed confident that their own 65nm mainstream parts will be superior. We can't help but wonder how much that matters when you release them 9 months later, though? It will also be interesting to see who's first to 55nm, and how good of a half-node it will be.
 
Not to offend, but this thread is pointless. You will /know/ when anyone here even thinks they know about anything you're asking about. Probably a month+ before there's any real data. This forum will /flood/ with rumor threads and projected specs. 3/4+ will be flat out wrong. Some will get parts right.

Until performance numbers are in we won't /know/ anything. The 2900XT/XTX was supposed to be great, and the spec sheet supported that. The performance numbers are in, and it flat out cannot keep up with what is now almost year old hardware.

Don't waste your time theorizing or listening to the rumor mill. Make up your mind once you see a bunch of benchmarks and a direct comparison with the previous generation. Anything less is worthless.

Aside from what Kowan posted as "all that has been said", I feel the need to quote the above wise-man tbh. QFT, and it's likely it will be a true mid-range card anyway...so you might as well buy now if you're ready. All the great games are coming out soon so you've got little time ;), just hope prices don't go up for other parts.
 
i just hope that, if i pull the trigger on a GTS 640, i can take advantage of EVGA's step up program...should purported cards deliver better performance. if it's a mid range...then...bleh.
 
There will be a 65nm part from Nvidia before the end of the year. There's no point in waiting, now that ATI has shown TSMC's 65nm yields are decent.

If the rumors are correct, the card will be an 8800 GS and have performance bridging the huge gap between the 8600 and 8800 GTS. This is the most logical refresh move Nvidia could make.

Beyond that, we know nothing.

Nah, Nvidia won't need to bridge gap between 8600 and 8800. After they release g92, it will be new king of the hill. 8800 GTX/ultra will fill the upper-midrange segment, and 8800 GTS will become the best midrange/budget card we've seen in years. The 8600 will become the new ultra-budget card, hopefully sporting a new design that has higher memory bandwidth. Next spring, the 8800 GS (or whatever they will call it) will likely fall somewhere between 8800 GTS and the 8800 GTX, and compliment that already amazing midrange lineup.
 
Nah, Nvidia won't need to bridge gap between 8600 and 8800. After they release g92, it will be new king of the hill. 8800 GTX/ultra will fill the upper-midrange segment, and 8800 GTS will become the best midrange/budget card we've seen in years. The 8600 will become the new ultra-budget card, hopefully sporting a new design that has higher memory bandwidth. Next spring, the 8800 GS (or whatever they will call it) will likely fall somewhere between 8800 GTS and the 8800 GTX, and compliment that already amazing midrange lineup.

I don't think a "GS" fits between those two. Besides, this is just what the rumors are, that'd it'll be midrange. A G92 is a mid-range class card, hence the "2", where it will fit with the current 8800 line is anyone's guess really. Having the 8800GTS moved down to super cheap levels (assuming this G92 trumps all 8800 hardware) would be too good to be true imo.
 
Arun said:
We can't help but wonder how much [65nm parts] matter when you release them 9 months later, though? It will also be interesting to see who's first to 55nm, and how good of a half-node it will be.
That's a damn weird thing to say.
 
I still seriously don't believe that G92 = next high end, esp if 8800s came out in nov of last year, it would just be retarded of nvidia to put something above their high end when their skill king of the hill in that department.

If g92 were to be 65nm, its not going to be the next high end, nvidia, ati, amd, intel all move down the ladder on nanometers in the middle of the generation, so if g92 were to be 65nm, it would most likely be a refresh, or what has been speculated before, something between 8600gts and 8800gts.
 
Not retarded at all, in my opinion. Consider the kind of people that make up the high-end graphics card market. It isn't unreasonable to assume that many high-end buyers upgrade more frequently than mid-range and low-end buyers, so supplying new high-end products every six to eight months isn't unreasonable, even if competition is somewhat lacking. Besides, if G80 is any indication, G9x has been in the works for a fair amount of time already, and may already be ready for silicon, especially if it's just a hot-rotted, 65nm G80. There are many costs involved in releasing a product to market, but in this arena, R&D is a pretty significant factor, perhaps even more expensive than the almighty marketing costs. You could sit on that investment for a while before bringing it to market, or you could release it when it's ready and take the market by the balls. The early adoption rate will be stereotypically strong, and will then slow down fairly significantly because of over-saturation, but that's still probably a profitable scenario. It makes more sense to assume that NVIDIA does things to increase profits, rather than to decrease profits, don't you think?

If NVIDIA brings it to market in Q4, it's because they made the determination that it's the right move for the bottom line. To call the move retarded without knowing the full details on NVIDIA's operations doesn't make much sense to me.
 
True, we just have to wait see, that's the real answer. Either way I don't care, mid or new high-end; I don't have the $$$ to buy another high-end card. For me it is becoming a trend of shooting for certain PC games as all I play on the PC are just FPS'. Everything else I do on consoles. After Crysis comes out, it will be a while before I considering upgrading again (especially if nothing good comes out for a long time). It's safe to say that I built this rig for a few current games and mainly for others like Crysis, UT3, Starcraft 2 and maybe 1 or 2 more ;) who knows.
 
It's all about market share. When AMD had Intel by the balls with K8, instead of constantly pushing for new releases, lowering prices, and pulling out the next generation architecture, they sat around doing nothing. Now, Intel, despite dominating AMD in CPUs, is on a 2 year cycle to new architecture with a transistor shrink in between and AMD is struggling to fight back.

If NVIDIA sits around and ATI pulls a miracle and comes out with a great card, NVIDIA would only shoot itself in the foot when it could have sold to consumers a new card that would keep people a nvidia customer before they could switch to a new ATI product.

Anyways, hopefully the next gen / refresh comes out sometime in November as the G80 did, so I can buy a new card in a month and step up to it.
 
Hasn't nvidia released a new highend card ever year for the past 5 years or so? Haven't they made statements that they have no plan to change their development cycle? Everything seems to point to a new high end at the end of this year and a new midrange the following spring.
 
Hasn't nvidia released a no highend card ever year for the past 5 years or so? Haven't they made changes that they have no plan to change their development cycle? Everything seems to point to a new high end at the end of this year and a new midrange the following spring.
no it used to be 18 months between high end
 
6800 = April 2004
7800 = June 2005
8800 = Nov 2006

about 15-16 months. Therefore...maybe...end of this year or early next year. Any earlier is "wishful thinking" for a new high end IMO. Still a good shot for a new midrange or a slight tweak to the top end.
 
Well previous dates don't mean a whole lot if nvidia did indeed move to the new Q4 high end release, Q2/3 refresh type of schedule... And keep in mind the 6800 didn't have a refresh and was out in 14 months whereas the 7800 had a refresh and the 7 series lasted 17 months.... so it's really moot to compare history.
 
Well previous dates don't mean a whole lot if nvidia did indeed move to the new Q4 high end release, Q2/3 refresh type of schedule... And keep in mind the 6800 didn't have a refresh and was out in 14 months whereas the 7800 had a refresh and the 7 series lasted 17 months.... so it's really moot to compare history.

But you don't know that they will be releasing a new highend in Q4. Therefore calling my dates moot with no hard data on your point, well makes your point moot. :p
 
There's no guarantee it is high end but they strongly hinted at it in this article: http://www.bit-tech.net/news/2007/05/25/nvidia_says_g92_to_near_1tf/1

According to Nvidia's Michael Hara, Vice President, Investor Relations and Communications, the new chip will be ready at Christmas. This is actually going to be Nvidia's new launch plan after its wonderful success with the G80 - the high-end parts will be released shortly before the Holidays, and then the mid- to low-range parts will start filtering out in the spring. This gives the high-end early adopters a bit of extra time to cradle their expensive toys.

/shrug
 
"G92" is supposedly the new high-end part, with over double the computational power. If anything I think this is more than likely going to be a G80 refresh, with higher clocks and possibly more shaders, TMUs, and ROPs packed into a much smaller and energy efficient 65nm die, following suit with the tweaks applied to the GeForce 6 architecture to the 7xxx series. Then again I could always be wrong. IMHO nVIDIA has little need at present to reinvent the wheel beyond efficency tweaks when it is already king of the hill, unless ATi can pull a rabbit out of their collective ass.

Then again just a refresh isn't always the smartest strategy, since that's what AMD did against Intel with K8 and now they're on the virge of going out of buisness if they can't find more cash...

Resting on your laurels is not good buisness.
 
Well it might not be called a refresh but from what's known is that it will be heavily based on the current G80 architecture since its well done. Obviously, they will have to fix and tune problems. A move to 65nm will mean lower power consumption and heat output also for the same amount of power. Of course, if they increase overall transistor numbers and so on, it will still put out a decent amount as any high end would.
 
One more thing I would like to add. NV40 = 130nm, G70/71 = 110nm, G80 = 90nm. Now nvidia has 80nm parts (G84/86) but their not high end. The last time nvidia tested the water with a new process on the high end, it failed (5800ultra,130nm). G80s 90nm silicon is the bomb, no doubt about that, it kicks the daylights out of the R600 while producing less heat, using less power, and essentially being slower in terms of throughput. Nvidia is a company that is always improving on the efficiency of its chips (look at G71), it was better than G70 with less transistors.

Nvidia burns thru millions in R&D, thats how they make money, thats the case everywhere, but it would be sort of stupid to put out a brand NEW chip <= 1 year and rush a new chip out that only is going to account for <5% of there total revenue stream, esp when there is no reason to rush it. It would make sense that "G92" in whatever form, even if its not called G92, a chip they release around a year after there last release wouldn't be a new high end, but rather something else, possibly to fill a gap somewhere.

IMHO I believe the rumor (and it is 100% a rumor) that G92 will be high end and will be around Q4 07 is false. But then again they are put in the unique position where R600 was a dud, R650 which was supposed to be here between may and july was put off / respun, and is still out there to be launched by ati or to be scraped all together, and where Nvidia COULD jump the gun if it felt the need to, but the more logical position would be to wait intill month 15-18 (after G80) with a refined next gen chip.

Besides the G92 rumor came out two months ago, what have we heard since then, anything? nothing? Hell, midway thru sept 06 half the people on these forums believed that G80 = non unified architecture, rumors are just that, baseless.

Logic > Rumor, but one thing is for sure, nvidia is definitely working on their 80 and 65nm designs, theirs no reason for them not to.
 
If the quoted statements from Michael Hara are more than just PR, the G92 based card will arrive after some of the major titles the gamers have been waiting for.

That's great news for us all since the question of how will it handle those games will be posted by reviewers.
 
I really do hope so.

No offense, but it's been a lackluster year in terms of video cards instead of the usual competition :D .

That 8800GTX is currently enjoying a very looong reigning period. ( Ultra, I know, but it's still just a hyper clocked GTX )

I'm hoping for a new card in Q4, but not holding my breathe.
 
Well, rumor is that ATI will refresh their R600 later this year to compete with g80 ultra/GTX cards, so if you you were wondering what the incentive was for nvidia to release a new high end card, that would probably be it.
 
Simpson5774 said:
G70/71 = 110nm

G70 was a 110nm refresh of 130nm NV40; you were correct about that, but G71 was the 90nm die shink/tweak of G70...
 
etherealtml said:
Well, rumor is that ATI will refresh their R600 later this year to compete with g80 ultra/GTX cards...

Actually, word around the campfire is that there will be no refresh of R600. It will continue to be manufactured on 80nm. The only new R6xx generation part in the pipe is supposedly the RV670, which will be the replacement for RV570, better known as the X1950 Pro. More than likely, AMD has all their resources dedicated to getting the multi-chip R700 out the door to attempt to regain the high-end crown.

Anyways, I'm really interested on how this multi-chip solution will look like and perform...
 
Actually, word around the campfire is that there will be no refresh of R600. It will continue to be manufactured on 80nm. The only new R6xx generation part in the pipe is supposedly the RV670, which will be the replacement for RV570, better known as the X1950 Pro. More than likely, AMD has all their resources dedicated to getting the multi-chip R700 out the door to attempt to regain the high-end crown.

Anyways, I'm really interested on how this multi-chip solution will look like and perform...

Something like crossfire but potentially a bit slower I expect, it's a good solution if you can get it onto 1 card for relatively cheap, the 7950GX2 was an awesome piece of hardware.

However, I doubt ATI putting 2x of any of their current GPU's onto 1 card is going to get them anywhere, not with performance figures like they have and crossfire seemingly performing worse than SLI.
 
Actually, word around the campfire is that there will be no refresh of R600. It will continue to be manufactured on 80nm. The only new R6xx generation part in the pipe is supposedly the RV670, which will be the replacement for RV570, better known as the X1950 Pro. More than likely, AMD has all their resources dedicated to getting the multi-chip R700 out the door to attempt to regain the high-end crown.

Anyways, I'm really interested on how this multi-chip solution will look like and perform...

So, did they scrap the r650? I haven't heard anything recently, but rumor was it is 65nm version of r600 and scheduled for Q3 2007. Wouldn't be surprised to see a dual core ATI early next year to compete with g92.
 
So, did they scrap the r650? I haven't heard anything recently, but rumor was it is 65nm version of r600 and scheduled for Q3 2007. Wouldn't be surprised to see a dual core ATI early next year to compete with g92.

According to the Inq, Fudzilla, and VR Zone among others it never existed. And this came straight from AMD. Then again ATi could just be denying any truth to those rumors to lull nVIDIA into a false sense of security. If DAAMiT was smart they would pop out a die shrink before R700.
 
Back
Top