• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

G71 taped out.

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Originally posted by: Janooo
If NV is sliding what happens to G80? What would be the time for G80?

Still think about a 6 month release cycle for the high end cards. If the G71 gets in for March, that's actually a slightly shorter cycle (4 months, rather than 6).
 
Originally posted by: Ronin
Originally posted by: Janooo
If NV is sliding what happens to G80? What would be the time for G80?

Still think about a 6 month release cycle for the high end cards. If the G71 gets in for March, that's actually a slightly shorter cycle (4 months, rather than 6).

I would say its been more like 9 month cycle, 7800GTX 512MB was as far as I know the same core with clock speeds bumped way up and more memory
 
And you don't think that required some development?

It's not the same memory. The 512MB GTX used 1.1ns Samsung chips.
 
I dont think of the 512gtx as a refresh, it's just the "ultra" model that might have been released back in July had the r520 come out when it was originally scheduled. It's higher clocked, more mem, faster mem, but the underlying architecture is basically unchanged. How is that any different from a 6800u, compared to the 6800gt, or even the 6800ultra extreme, which also came out slighly later, featuring similar clockspeed and memory bumps (not to mention price and availability 😉). Actually, given how much supply problems the 512gtx had in November and December, there's no way it could have been released in any significant quantity last July, but maybe a 512mb 500mhz part would have been possible, using 1.26ns mem.
 
Originally posted by: Ronin
And you don't think that required some development?

It's not the same memory. The 512MB GTX used 1.1ns Samsung chips.

a tiny bit but that is hardly the same thing as going from G71 to G80.
 
Originally posted by: Cookie Monster
I heard that the G71 will be using 1ns ram. NOT 1.1ns. I could be wrong here.


I hope not or rumors of a very expensive card could be true.
 
Originally posted by: ElFenix
Originally posted by: Ronin
And you don't think that required some development?

It's not the same memory. The 512MB GTX used 1.1ns Samsung chips.

a tiny bit but that is hardly the same thing as going from G71 to G80.

Well, you can't just throw faster RAM on a PCB and expect it to work like the previous did. There are a lot of variables to consider. Same holds true for increasing the clock speed of the core (and the voltage, for that matter).
 
Originally posted by: Ronin
Originally posted by: ElFenix
Originally posted by: Ronin
And you don't think that required some development?

It's not the same memory. The 512MB GTX used 1.1ns Samsung chips.

a tiny bit but that is hardly the same thing as going from G71 to G80.

Well, you can't just throw faster RAM on a PCB and expect it to work like the previous did. There are a lot of variables to consider. Same holds true for increasing the clock speed of the core (and the voltage, for that matter).

right, but nvidia had this card designed from the get go. it probably uses the same pcb as the quadro cards do. they just sat on it to spoil ati's fun. (well, that and the ram they designed it for wasn't out). regardless of what it is, what it isn't is a genuine refresh or generational release.
 
Well regardless when it comes out Nvidia was smart to stop the paper launches with the 6 series. IF they were smart again they would just delay it a month rather than paperlaunch.
 
Originally posted by: ElFenix

right, but nvidia had this card designed from the get go. it probably uses the same pcb as the quadro cards do. they just sat on it to spoil ati's fun. (well, that and the ram they designed it for wasn't out). regardless of what it is, what it isn't is a genuine refresh or generational release.

AFAIK, the GTX512's PCB had to be totally redone to accomodate the double density 64-bit 1.1ns memory modules Samsung had, and for different voltages.If you notice, the memory is all on one side, and thats 512MB worth. The regular GT/GTX has 1/2 that memory but is on both sides. This is presumably the reason for the initial high cost of the card. That, and the memory itself. It's availability is another story. So, more than a tiny bit of R&D had to go into the GTX512.

 
Originally posted by: Ronin
Why? Even if they don't sell all that money, they made plenty in the gap between June and November when ATi had nothing to offer up as competition.

They're hardly breathing heavy about this, even if ATi has a window.

I'm sure they did make good money, but right now their competition has a better product in mass quantity (ok, single card product) and their next big launch is realistically 2months away, and that's if everything goes right. Corporations never want their competition on top, ever, no matter for how short a time it may be. Probably a minor setback, yes.
 
Originally posted by: Ronin
Why? Even if they don't sell all that money, they made plenty in the gap between June and November when ATi had nothing to offer up as competition.

They're hardly breathing heavy about this, even if ATi has a window.



Well this is a first, I agree with Ronin.
 
Originally posted by: keysplayr2003
AFAIK, the GTX512's PCB had to be totally redone to accomodate the double density 64-bit 1.1ns memory modules Samsung had, and for different voltages.If you notice, the memory is all on one side, and thats 512MB worth. The regular GT/GTX has 1/2 that memory but is on both sides. This is presumably the reason for the initial high cost of the card. That, and the memory itself. It's availability is another story. So, more than a tiny bit of R&D had to go into the GTX512.

if you're arguing that it is a refresh, then by that logic then the redesigned PCB for the x1800 counts as a refresh.

again, the whole point is that the GTX 512 is not a refresh of the 7800 chip. nvidia is not on a 6 month product cycle, and has not been in years.

and even if the GTX 512 is a refresh, which it isn't, that still doesn't shed light on when to expect G80. modifying a PCB is a far easier task than designing a completely new GPU core.
 
Originally posted by: ElFenix
Originally posted by: keysplayr2003
AFAIK, the GTX512's PCB had to be totally redone to accomodate the double density 64-bit 1.1ns memory modules Samsung had, and for different voltages.If you notice, the memory is all on one side, and thats 512MB worth. The regular GT/GTX has 1/2 that memory but is on both sides. This is presumably the reason for the initial high cost of the card. That, and the memory itself. It's availability is another story. So, more than a tiny bit of R&D had to go into the GTX512.

if you're arguing that it is a refresh, then by that logic then the redesigned PCB for the x1800 counts as a refresh.

again, the whole point is that the GTX 512 is not a refresh of the 7800 chip. nvidia is not on a 6 month product cycle, and has not been in years.

and even if the GTX 512 is a refresh, which it isn't, that still doesn't shed light on when to expect G80. modifying a PCB is a far easier task than designing a completely new GPU core.

Not everything has to be an argument. I am merely discussing this. My comment was to address your comment on the amount of development that went into the PCB. I have no doubt that a core is exponentially more difficult to design and build than a PCB. That doesn't necessarily make designing and building a PCB to handle the components,voltages,clocks of a card any easier.

 
You're still ignoring what had to go into even making the 512MB GTX possible. If it was simply upping the core speed, smacking some RAM on the card, and upping the RAM speed, sure, I'd buy that.

What redesigned PCB for the X1800 are you talking about, since November?
 
This isn't new news. it has long been thought that the g71 would be out in March and at the earliest late Feb. It is funny how the ati fanboys say this is such a big deal. ATI enjoy your lead (as small as it might be) as it will be very shortlived. G71 appears to be on the expected timeline.
 
nVidia isnt going to let ATI have the fastest card for long. Rumors are flying that its early March so we'll see what happens then. I will say that whenever nVidia does get it out they better get it correct from the git go or they'll get heckled alot by the ATI fans for sure. I'm sure ATI is working on getting something ready to combat the G71 as well. The video card war is about to heat up some more which is good for us consumers. ;-)
 
Ronin, are you permitted to say anything about the G71? Expected release/introduction date? I heard March at Cebit. If you can't, thats cool.
 
This gives time for the r590 to come out soon after that i think they have started tapping them out as well.
This is what i think will happen

G71beats x1900xtx
R590 beats G71

(G71 Might not be widely available because of extreme clock & memory speeds Same with r590)

Video Ultra high end has become a war of one up every six months with hidden high end cards and refreshes

Think about it r590 wil be a refresh of a refresh.
 
Originally posted by: keysplayr2003
Originally posted by: ElFenix

right, but nvidia had this card designed from the get go. it probably uses the same pcb as the quadro cards do. they just sat on it to spoil ati's fun. (well, that and the ram they designed it for wasn't out). regardless of what it is, what it isn't is a genuine refresh or generational release.

AFAIK, the GTX512's PCB had to be totally redone to accomodate the double density 64-bit 1.1ns memory modules Samsung had, and for different voltages.If you notice, the memory is all on one side, and thats 512MB worth. The regular GT/GTX has 1/2 that memory but is on both sides. This is presumably the reason for the initial high cost of the card. That, and the memory itself. It's availability is another story. So, more than a tiny bit of R&D had to go into the GTX512.

Not to get picky but my 7800GT has ram on ONE side....

Either way what I'm figuring is I'll wait for the end of the summer to buy a new card to replace my 7800GT, I figure around then all the product lines will be mature and I could probably get a real good deal on one of these refresh cards.
 
Back
Top