NVIDIA To launch GeForce 9 Series In February

hehe it`s a crazy game this pc business.

I was just thinking I got 2/3 months to enjoy 8800 goodness b4 it isn`t "current tech".

All good tho.

I guess a new mid range card isn`t comming out in the meantime, s`pose there isn`t really a need.

Crysis @ 2600x1800 on full ???!?! lol
 
Why do i not believe this? WHy would they make the midrange card 55nm and the mainstream 65? THat dosent make any sense at all. I call rubbish.
 
nice if they stick to february i reckon thats about the same time student loan goes in 9 series come to me baby lol
 
I was reading about this yesterday morning (hence the sig lol).

PP, someone on another forum offered this as an explination, sounds plausible

It's just the approach NVIDIA has been doing, I think the strategy is to "test the waters" on the mid range cards first to be sure the next highend series production will function smoothly. Kind of like Intel testing the grounds on a lower manufacturing process on same architecture before letting out new chips based on a new architecture.
 
name='PP Mguire' said:
Why do i not believe this? WHy would they make the midrange card 55nm and the mainstream 65? THat dosent make any sense at all. I call rubbish.

Nvidia have mirrored Intel with their annual changes. Every 2nd year being a large change.

The 8800GTX & Ultra came out before the later mid range cards (based on the smaller manufacturing process) ie. 8800GT.

I'm guessing this is how it's going to be for their 9 generation cards too.

The enthusiast card will be released first, based on the same process that the current G92 process or something similar, with the mainstream, mid/low range cards being released some time later on a smaller process
 
name='Yeungster' said:
The enthusiast card will be released first, based on the same process that the current G92 process or something similar, with the mainstream, mid/low range cards being released some time later on a smaller process

Think I know where u`re going with that. Looking into the future, what we`d think is that b4 a theoretical Geforce 10k card comes out, a consolidating Geforce 9800 GT may come out based on a G100 chip.
 
Yep that's the way both companies do things. Lower range chips tend to inherently have lower yields as they don't have to be blistering so the companies move to lower nm processes with these as they are easier to get R&D'd properley and in numbers. Also less risky than the expensive high profile top-end GPU's. It completely makes sense from a business perspective
 
typical .. i go and buy a new 8600GT card for my new PC and Nvidia decide to bring out a new card around feb lol ... could of saved some extra ££££ to get one :(
 
Aint no point in thinking like that m8.

By the time u w8d and bought the newer card, a press release will come out for a newer rev of something.

Best just to get what u want at the time - unless ur talking about weeks - or ur fixing a d8 to upd8 a rig as a whole.

By the time the competitor for the 8600 is due, from the report, it will be l8r than feb. Feb seems to be a pointer to the top end card so it seems.
 
i spose yea ... computer technology is moving along a lot faster than it ever has before, once you have bought something top of the range; within months something will be out to 'better' it
 
i KNOW its not like that, but he might not.. i know how all this works lol :P But thats the only way i can see him getting to 4.0Ghz
 
This is standard operating procedure. 8800GTX was 90nm. 8600/8500/8400 were 80nm. Nvidia eventually re-released their tweaked high-end GPU at 65nm, the 8800GT.

ATI does the same. The HD 2900XT was 80nm, 2400/2600 cards were 65nm. And also just like Nvidia, ATI re-released their tweaked high-end card on a 55nm process.

I may be wrong but I believe it is done for cost reasons, the lower/midrange parts have a much smaller profit margin, so the smaller the physical GPU size the higher their profit to offset this. The second, and likely just as important reason is it also will give the fab facilities and GPU companies time to work the kinks out of the smaller node process size, using the simpler lowend/midrange cores. Once they have the bugs worked out they can use the process size on much larger and more complex cores, as they have both done with G92 and RV670.
This is how somebody explained it to me. Makes perfect sense actualy.
 
Back
Top