ATI to delay 80nm GPU migration?

kyetech said:
(Excuse the double post in a row)

So basically I speculate the following :

g80 (nv50) Aug. 2006 80nm.
g90 (nv55) march 2007 65nm. ( I HOPE)
g100 (nv60) fall 2007 65nm (mature)

its the NV60 im most looking forward. And does any body have any idea when we are going to get 512bit memory bus. I mean, memory just seems to be crawling along in progress compared to the GPUs at the mo.

Just for fun does anybody wanna guess at transistor counts for these things !
Click to expand...

I think it's going to be more like

G80 (NV50) on 80nm - Fall 2006

"G81" ("NV51") speedbump on 65nm - Q1 or Q2, 2007

G90 (NV55) a major refresh - on 65nm - H2, 2007

"G91" ("NV58") or whatever, a speed bump of the refresh (ala G71) on 65nm or 45nm - H1, 2008

then finally, G100 (NV60) on 45nm - H2, 2008

we're not going to see NV60 / NV6X until 2008, with NV50 arriving late 2006, we will see all the usual speedbumps, refreshes and speedbumped refreshes, as we have the last 3-4 years.
 
chavvdarrr said:
7600 is smaller than X1600

Also 0.11 had no low-k, both 0.09 and 0.08 are low-k with the latter only optical shrink
Click to expand...
optical shrinks aren't done anymore, most propperties change significantly enaugh to redo the chip almost entirely but obviously there are tools to make the transition betwean 90nm & 80nm very smooth, but it takes a lot more time than a optical die shrink would
 
maaoouud said:
Are you suggesting that AMD would take a role as a foundry for Nvidia ? I think that very unlikely. By the time that AMD would be able to put capacity aside for foundry services the manufacturing process at fab 30 would probably not be attractive for Nvidia.
Click to expand...


In my opinion this is very likely. My bet is that in 2007 AMD will offer foundry service to Nvidia. My bet would be less complicated designs for instance chipsets and IGPs and i still think that once FAB36 is running quite well AMD will start to deploy new tools/machines to FAB30 for smaller process nodes. But they will do that not in a hurry.
FAB30 is still to tempting for AMD to drop it. The location is great because distance to FAB36 is - well rather small :) -.
I also can imagine that they are trying to experiment with Nvidia on a combined CPU/GPU part in the long.

I think they will do something together in 2007 and i think it is kind of a foundry agreement.
 
I'm not sure how to understand it... RV560/570 will be delayed?/cancelled?/produced on 90nm?/or only UMC vesion will be available?/...
icon_upset.gif
 
I first thought that RV560 was the 8 pipe 128bit version of RV530/R580 in 90nm, and RV570 was the shrink to 80nm.

now there's this rumors of 8 pipe 128bit RV560 on 80nm, and 12 pipe 256bits RV570 on 80nm but I find it hard to believe; why make a X1900 without one of its quad. 256bits bus would make the cards expensive, so why not just use a R580. Or a R590. What would be the use of such a RV570 when you're already making the R590 to provide cheaper X1900's?

I sure don't think ATI has produced enough R520 for X1800GTO to compete against the 7600GT for another 8 monthes or so; 90nm RV560 seems logic to me (as a scaling of RV530/R580) and would be much needed by ATI imo.
 
My presumption :

RV560/RV570 will be a refined R580 architecture which have 3 Quad per dispatch processor compared to R580/RV530 specs with 4 Quad per dispatch processor will have better ALU utilization . However , the rumor previously had that RV570 is a 12 Quad PS seems unlogicial to me from primary cost issue point of view.
 
While waiting for 80nm to gets its kinks fixed, anyone care to speculate on what 90nm chip(s) UMC will be producing for ATI? RV515 would be an obvious guess, but any reason to think they might want to fab a more complex ASIC on UMC's 90nm process?
 
Seems like a rather uninteresting trail no matter where it leads. He should be barking up the RV560/570 tree instead.
 
Near future ati products won't have any gddr4. R600 probably being the first one

RickBergman: Roadmaps, roadmaps, that is what everyone wants from us! All we can say is that it’s coming though. We have no products to announce that will use it in the near future. I think the reason for this is because GDDR3 went so much higher than anyone expected and it means there has been no real urgency for GDDR4. It is true though that GDDR4 will bring with it some performance improvements when used in conjunction with our memory controller.
 
You must log in or register to reply here.

Similar threads

E
Microsoft designed the 360 GPU?
Replies
10
Views
2K
davis.anthony
D
J
Graphic card/CPU configuration for best memory graphics in Flight Simulator 98?
2
Replies
37
Views
3K
homerdog
homerdog
D
Hardware implementation of threading models in contemporary GPUs
Replies
16
Views
3K
Lurkmass
L
Rys
Reopening Architecture and Products + rules and guidelines changes
3 4 5
Replies
90
Views
13K
Shifty Geezer
Shifty Geezer
D
Video Game Console Cooling - A Hot Topic (pun kind of intended).
Replies
12
Views
2K
arandomguy
A
Back
Top