Impact of Early Vista on GPU Release Schedules?

[Jawed]

The other side of the ball is even more interesting, I think. The Green Team vibes (6800GS, 7800GS) seem to be towards pushing back the next gen, not pulling it forward. Obviously they must be closely in touch with MS on Vista timing as well, so what does this mean if we aren't going to see a 90nm high-end part from them until, say, March? Could it be later? Might they be going straight to their Vista part on the next one?

Both 6800GS and 7800GS look like "fill-in" because the 7600, 90nm, part isn't ready.

I don't think things are rosy in NVidia's 90nm garden.

7300Go is practically out there as a 90nm part, but...

Jawed

 

[Fodder]

I don't think the 7800GS is that much of a 'fill-in', given that there surely would be more than a few cores that didn't make GT spec. The 6600GT arrived fairly promptly but we still got vanilla and LE 6800s, so I suspect that even if the 7600 were out by now (as was expected) we'd still see GS parts. The 6800GS smelt like a stock-clearance at first, but with AGP models on the way I'm no longer so sure about that.

 

[BrynS]

Both 6800GS and 7800GS look like "fill-in" because the 7600, 90nm, part isn't ready.

I don't think things are rosy in NVidia's 90nm garden.

7300Go is practically out there as a 90nm part, but...

Jawed

Why hasn't any info on nV's larger 90nm ASICs leaked out?

Apart from the mention in the conference calls relating to C51 and what is now the 7300Go as well as the AnandTech piece on 7600 launch probability in Feb-March 2006, I don't recall any tapeout/proposed ramp details or discussion for 7200, 7600 and 90nm G70.

Anyone think the 7200 desktop variants will be 7300Go derivatives with slightly higher clocks and TurboCache as required or a slightly beefier configuration with an addtional fragment quad? What's the transistor count on 7300Go?

 

[Sazar]

I read the title and laughed aloud

The impact of EARLY vista

It would have been early LAST year.

 

[Geo]

I read the title and laughed aloud

The impact of EARLY vista

It would have been early LAST year.

It all depends on where you are in the story. . . The twists, the turns. . .the chills and spills.

 

[Geo]

Considering what a mess ATI made of pretty much everything a year ago, basing R580's timetable on what happened to R480 is nonsense.

Why refresh a product intentionally to miss the biggest sales-volume period of the year? Especially when the refresh was supposed to be an "availability refresh".

I still think R580 was meant to be a November product, according to the original schedule of R520 in May, RV515 in August and RV530 in September/October.

Jawed

Reasonable point. If we go back a year further when "ATI could do no wrong", we find 9800Pro on March 6 and 9800XT on Sep 30. That was back in the "one month to availability" days, so roughly Nov 1 to buy 9800XT. Which would suggest that an R520 at Computex (Jun 1, roughly) was already at the outside boundaries of the planned schedule originally.

 

[Tim Murray]

Reasonable point. If we go back a year further when "ATI could do no wrong", we find 9800Pro on March 6 and 9800XT on Sep 30. That was back in the "one month to availability" days, so roughly Nov 1 to buy 9800XT. Which would suggest that an R520 at Computex (Jun 1, roughly) was already at the outside boundaries of the planned schedule originally.

I don't know how true that is. Remember, this was about the time when it was widely decreed that ATI and NV were slowing release schedules, and since we went from 9700 Pro to 9800 Pro in the same amount of time as 9800 Pro to 9800XT, a new architecture could easily justify an extra month or so.

 

[Geo]

I don't know how true that is. Remember, this was about the time when it was widely decreed that ATI and NV were slowing release schedules, and since we went from 9700 Pro to 9800 Pro in the same amount of time as 9800 Pro to 9800XT, a new architecture could easily justify an extra month or so.

Sure. But I said "at the outside boundaries", not "outside the boundaries". I think of these things as being targeted at an envelope rather than a specific date.

 

[ninelven]

So you expect them to ride g70 for another 8mos+?

Sorry, that should have been 2Q instead of 2H.

 

[Sazar]

It all depends on where you are in the story. . . The twists, the turns. . .the chills and spills.

It's a thrilla I tells ya

 

[Geo]

I don't think the 7800GS is that much of a 'fill-in', given that there surely would be more than a few cores that didn't make GT spec.

I don't think I can agree with that just yet --remember that GT is already GTX rejects, and as you move down the price range volume demand goes UP. But I guess if we see a 7800GS we'll find out how many of them unlock to GT reliably.

 

[dizietsma]

Can anyone provide links to when an operating system launched early ?

 

[Geo]

Can anyone provide links to when an operating system launched early ?

For the purpose of what?

 

[Pete]

Odds are, of mocking MS.

 

[Fodder]

as you move down the price range volume demand goes UP.

Unless there's an equivalent part out to soak up that demand. 6800LEs were relatively scarce compared to 6600GTs, and if the 7600 had arrived I expect the 7800GS would find itself in the same situation. With the 7600 AWOL I wouldn't be surprised to see G70 cores artificially crippled, but I think the GS would have existed regardless.

I guess what I'm getting at is that I think the GS models were always on the cards, they've just suddenly had more important roles thrust upon them.

 

[DegustatoR]

Sorry, that should have been 2Q instead of 2H.

That's optimistic. I think they'll 'ride the g70' till 2h2006. And i see no reason why they shouldn't.

 

[ninelven]

And i see no reason why they shouldn't.

Does getting beat in every benchmark known to man by the R580 count as a reason?