AMD: Southern Islands (7*** series) Speculation/ Rumour Thread

IMHO,

not necessarily.
Let’s just pretend that they can reach 1400MHz with a new PCB layout and 2x8. Therefore such a card would be up to 50% faster than a HD7970. What would you prefer then, such a single-GPU card or a slightly (up to ~20% with perfect scaling) faster dual-GPU card with microstuttering and driver quirks?

IMHO AMD could therefore also bring a card with binned Thaiti GPUs @ up to 1400MHz and name it HD7990.

Well, 7990 was reserve for dual GPU solution or we would think that.... Since we had 6990 Dual Cayman.
 
Well, 7990 was reserve for dual GPU solution or we would think that.... Since we had 6990 Dual Cayman.
But then again we had 5970 as a dual chip card but 6970 was a single-chip card.

I wouldn't be too surprised (but I'm not expecting it either) if they released a 7990 single-chip at well over 1 GHz and called the dual chip 7995.
 
OTOH, I still think it'd be perfectly fine if AMD just relaunched all the HD7000 series as HD8000 with 50% higher clocks across the board, given they could manage to keep the power consumption in check (through voltage regulation and better 28nm maturity).

Heck, nVidia just got away with G92 spread across some 5 generations...
 
I think you guys are being wildly optimistic regarding the clock speeds AMD may be able to achieve on this supposed "7980". I do not expect them to reach anywhere close to 50%, not while managing to keep power consumption at a reasonable level at least. And not to mention they'd have to be able to mass produce it so if they wanted serious quantities they'd have to bin accordingly.

Take the 4890 v/s 4870 for example; even with a new spin, chip timing adjustments, cap ring, and almost another year of process maturity, they were able to move from 750 to 850 mhz. So for a hypothetical "7980", my guess is that at most they'd be able to reach somewhere around the current overdrive limits of Tahiti, i.e. 1.1 Ghz core/1.5 ghz memory.

Oh and im just gonna quote a post of mine from a while back - http://forum.beyond3d.com/showpost.php?p=1569248&postcount=555
With regards to die size, i dont think we're going to go back to RV770 levels of die size. I expect something close to Cypress. With APU's set to capture the entire low end and lower mid-range market, the performance gap between APU's and mid-range discrete will have to go up. Otherwise there would be no reason to purchase mid range discrete GPU's. I think they'll have to have a three chip lineup going forward, similar to the current gen where they have Cayman, Barts and Turks. A chip like Caicos will not be required in 2012 when we have Ivy Bridge and Trinity. So for the three chip lineup i expect something like a ~350 mm2 GPU, a ~250 mm2 GPU and ~150 mm2 GPU

I guess i was right about the three chip lineup, and i think the die sizes should be about right too :D
 
Last edited by a moderator:
I think you guys are being wildly optimistic regarding the clock speeds AMD may be able to achieve on this supposed "7980". I do not expect them to reach anywhere close to 50%, not while managing to keep power consumption at a reasonable level at least. And not to mention they'd have to be able to mass produce it so if they wanted serious quantities they'd have to bin accordingly.
So far for what I've seen, people have both OC'd to CCC limits and at the same time undervolted the cards without problems, so getting quite high clocks while staying under 300W, or even 250W, shouldn't be that hard to do
 
I guess they are binning chips for dual Tahiti already and we wont see any XTX with present stepping.. according to Charlie's calculations small stepping should take ~3 months and big one ~6 months.. since Tahiti is working perfectly, 6 months doesnt make any sense.. considering AMD finishes all 7000 lineup in March, They may release XTX in Summer..
 
So far for what I've seen, people have both OC'd to CCC limits and at the same time undervolted the cards without problems, so getting quite high clocks while staying under 300W, or even 250W, shouldn't be that hard to do

Yes but as soon as volts jump, power consumption soars. I don't expect anything beyond 1.2Ghz and even that would probably be an OEM OC. 25% is plenty to distinguish it from the 7970 anything beyond that is likely to encroach on refresh parts. And I wouldn't expect to see it at all until competition launches something faster because they will need to compress their price structure for it to make sense.
 
Kaotik said:
So far for what I've seen, people have both OC'd to CCC limits and at the same time undervolted the cards without problems, so getting quite high clocks while staying under 300W, or even 250W, shouldn't be that hard to do
Did those people submit those cards to a full system qualification? Check all possible load scenarios? Run bit error tests? Check with chips from various corner lots? All at different temperatures, from 0C to, say, 35C?

There's a reason AMD and Nvidia set certain (low) clocks and then encourage overclocking. Leaving performance on the table so the competitor can take the top spot is not one of them.

Not saying there won't be an official higher clocked version, but let's not go overboard with specs based on some dude in his basement not seeing visual corruption.
 
Last edited by a moderator:
From 50 to 70:
HD 5k8: +30%
HD 6k9: +20%
HD 7k9: +32% (in need for DOUBLE CONFIRMATION!)
-----------------
Nothing out of the ordinary in my books
 
If it overclocks to 1200mhz it be likely the price/performance deal of a long time.
It won't oc to 1200mhz unless you add volts at least to where the 7970 is, disable power tune and the power delivery doesn't blow up on you.

That being said, wasn't there supposed to be no reference 7950 designs but only custom ones?
 
Back
Top