Radeon HD 6900 Series Products Listed on European Store

[Tatty_Two]

so its basically double bart, so i expect it will have double the performance too

Possibly but not necessarily, two HD5770's even with scaling below 70% are faster than an HD 5870

 

[bear jesus]

so its basically double bart, so i expect it will have double the performance too

the problem is with the new architecture it means any form of direct comparison will be wrong, i was mainly curious about the supposed changes in transistors per square millimeter

Game devs won't stay that long. They will make better graphic games for PC and the console games will look shit. It's just like the time when EA released NFS Most Wanted for PS2. Had the shittiest graphics and the PC one was fabulous at that time. And that was back in 2005 when the PS2 was like 3yrs old.

I really want the next gen consoles to come out as it always takes a while for them to be pushed as far as they can go meaning for a little while even console ports come with better an better graphics.... that's are already dated on the pc

 

[MxPhenom 216]

Facts are, 6970 will be faster than the 580 by 10-20%, 6950 will be on par or slightly faster than the 570.

To bad its not. AMD first intented for the 6970 to be 20 to 30 percent faster then a 480 and now that the 580 its basically already covered that. so now im thinking the 6970 = the 580. give or take a few

 

[mdsx1950]

To bad its not. AMD first intented for the 6970 to be 20 to 30 percent faster then a 480 and now that the 580 its basically already covered that. so now im thinking the 6970 = the 580. give or take a few

I wouldn't be surprised if the 6950=GTX580.

 

[Tatty_Two]

I wouldn't be surprised if the 6950=GTX580.

I would be!

 

[Over_Lord]

so its basically double bart, so i expect it will have double the performance too

umm math???

double bart = 3.4Billion transistors... seriously!

 

[Xaser04]

There was plenty of stock in the UK, 6870 never sold out at all at OCUK, it was not in demand. 580 in the UK is currently extinct lol, Nvidia can't produce them in any substantial capacity, it was said a 512 core fermi was unmanufacturable, it just about is. Stock increased after a few week for the 68** cards, the opposite is true of the 580, stock has gone and it's not being replaced.



The 6950 is 1GB, it will be well less expensive than the 6970, easily under £300. And I'm talking prices without etailer gouging.

The 6870 never sold out on OCUK because they were ovepriced compared to other websites. The same could be said of the HD6850 at launch.

Of course the Don's didn't like everyone pointing this fact out on the forums...... D

I do hope you (Gibbo) are right on pricing though, £350 for a HD6970 sounds tasty although I will be waiting for the HD6990 myself ( I want an upgrade not a sidestep).

Oh and hi Raven......

 

[PopcornMachine]

I have no basis for any conclusion at this point...therefore I conclude (based on preconceived bias)...

 

[Benetanegia]

so its basically double bart, so i expect it will have double the performance too

Aside from the fact that double Barts would be 3.4 billion and 510 mm^2, you are forgeting the many things that AMD had to cut down in order to achieve that die size:

http://www.anandtech.com/show/3987/...renewing-competition-in-the-midrange-market/2

1- One of them was the slower memory controler which they borrowed from Redwood and is half the size of the one in Cypress. The high speed memory in Cayman will need a controler as big or bigger than Cypress, since memory is much higher clocked.

2- Lack of double precision math on Barts.

The effect of adding double precision can be seen (or sensed) in Redwood vs Juniper vs Cypress. Each of them is half the next one when it comes to SP/TMU/ROP/MC, but let's look at transistor count:

Redwood = 627 million
Juniper = 1040 million (+66%)
Cypress = 2154 million (+107%)

As you can see the transistor count jump is much bigger in Juniper->Cypress, and the main reason as explained by AMD themselves is that both Redwood and Juniper lack double precision, just like Barts lacks it too as explained in the Anand article.

 

[WarEagleAU]

Hmm, if prices are right, this will be good.

 

[Benetanegia]

Another reason that I think that Cayman will not be as fast as most people think it will be, namely twice Barts, is that if Cayman (and VLIW 4) was trully so much better than Cypress (VLIW 5), why is Barts VLIW 5? Why didn't AMD release a Barts which was VLIW 4? IMO in general and as long as it remains efficient VLIW5 is faster per mm^2 than VLIW4 and the only reason that AMD went VLIW4 is that VLIW5 didn't scale well past 14-16 SIMDs (1120-1280 SP). Basically Barts being VLIW5 as it is has better perf/area than a similar performing VLIW4, but at similar or higher perf levels than Cypress, an VLIW4 SIMD is as fast as VLIW5 while being 10% smaller (AMD's own claim).

I know the above it's messy and hard to understand so here's a summary:

- Entire chip has less than 16 SIMDs: VLIW5 >> VLIW4, because it can do 1 more op/second and with this "low" ammount of SPs the setup engine + dispatchers can keep up. aka the chip remains efficient.

- Entire chip has more than 16 SIMD: VLIW4 >> VLIW5, because VLIW5 has demostrated that it cannot scale performance with so many SIMDs, reason for which Cypress is nowhere near 2x as fast as Juniper despite being exactly 2xJuniper. Or how the HD5850 is exactly as fast as HD5870 when both are running at the same clocks.

 

[dir_d]

Another reason that I think that Cayman will not be as fast as most people think it will be, namely twice Barts, is that if Cayman (and VLIW 4) was trully so much better than Cypress (VLIW 5), why is Barts VLIW 5? Why didn't AMD release a Barts which was VLIW 4? IMO in general and as long as it remains efficient VLIW5 is faster per mm^2 than VLIW4 and the only reason that AMD went VLIW4 is that VLIW5 didn't scale well past 14-16 SIMDs (1120-1280 SP). Basically Barts being VLIW5 as it is has better perf/area than a similar performing VLIW4, but at similar or higher perf levels than Cypress, an VLIW4 SIMD is as fast as VLIW5 while being 10% smaller (AMD's own claim).

I know the above it's messy and hard to understand so here's a summary:

- Entire chip has less than 16 SIMDs: VLIW5 >> VLIW4, because it can do 1 more op/second and with this "low" ammount of SPs the setup engine + dispatchers can keep up. aka the chip remains efficient.

- Entire chip has more than 16 SIMD: VLIW4 >> VLIW5, because VLIW5 has demostrated that it cannot scale performance with so many SIMDs, reason for which Cypress is nowhere near 2x as fast as Juniper despite being exactly 2xJuniper. Or how the HD5850 is exactly as fast as HD5870 when both are running at the same clocks.

I think one of the reason very may well be time....They (AMD) only had the time to make Cayman VLIW4 since 32nm node was cancelled. They decided to refine the VILW5 on the other chips instead. This might not be accurate and you could be right but this is my educated guess. I really believe they didnt have the time to make VLIW4 work the way they wanted it to except for on one chip.

 

[Tatty_Two]

Another reason that I think that Cayman will not be as fast as most people think it will be, namely twice Barts, is that if Cayman (and VLIW 4) was trully so much better than Cypress (VLIW 5), why is Barts VLIW 5? Why didn't AMD release a Barts which was VLIW 4? IMO in general and as long as it remains efficient VLIW5 is faster per mm^2 than VLIW4 and the only reason that AMD went VLIW4 is that VLIW5 didn't scale well past 14-16 SIMDs (1120-1280 SP). Basically Barts being VLIW5 as it is has better perf/area than a similar performing VLIW4, but at similar or higher perf levels than Cypress, an VLIW4 SIMD is as fast as VLIW5 while being 10% smaller (AMD's own claim).

I know the above it's messy and hard to understand so here's a summary:

- Entire chip has less than 16 SIMDs: VLIW5 >> VLIW4, because it can do 1 more op/second and with this "low" ammount of SPs the setup engine + dispatchers can keep up. aka the chip remains efficient.

- Entire chip has more than 16 SIMD: VLIW4 >> VLIW5, because VLIW5 has demostrated that it cannot scale performance with so many SIMDs, reason for which Cypress is nowhere near 2x as fast as Juniper despite being exactly 2xJuniper. Or how the HD5850 is exactly as fast as HD5870 when both are running at the same clocks.

I seem to recall an AMD chart (false or not) that showed that the existing 5970 was still the competitor for the 580, the 6970 competing against the 570 etc etc "if" that chart was accurate, seems AMD themelves (unlike every speculative person around) is saying that the 6970 won't be as fast, but if that is the case, prices will be lower of course.

 

[erocker]

I seem to recall an AMD chart (false or not) that showed that the existing 5970 was still the competitor for the 580, the 6970 competing against the 570 etc etc "if" that chart was accurate, seems AMD themelves (unlike every speculative person around) is saying that the 6970 won't be as fast, but if that is the case, prices will be lower of course.

I'm pretty sure all of those slides with comparison performance are fake. The only recent news stories posted by sites come from the same source. Fudzilla.

 

[LAN_deRf_HA]

AMD first intented for the 6970 to be 20 to 30 percent faster then a 480 and now that the 580 its basically already covered that.

I'd take another crack at those numbers. For one the 580 doesn't come close to a 20-30% advance over the 480. Try 10-15%, and I doubt AMD is deluded enough to think they'd beat a 480 by anywhere near 30% faster on the same manufacturing process.

 

[bear jesus]

I'd take another crack at those numbers. For one the 580 doesn't come close to a 20-30% advance over the 480. Try 10-15%, and I doubt AMD is deluded enough to think they'd beat a 480 by anywhere near 30% faster on the same manufacturing process.

Umm is that not like saying intel would be deluded to think they could beat amd by anywhere near 30% on the same manufacturing process?

Is not cayman meant to be all about the new architecture? thus we have no clue what is possible, i mean look at intel's and amd's 65nm cpu's was there not a big gap there and all because of architecture?

I coudl be totally wrong in my thought pattern here as i'm rly dam drunk

I'm just trying to say what many others have said really, we don't have a clue what is possible until they come out due to changes in architecture, for all we know it could be 600% faster yea ok thats a little into the range of impossible but hopefully there was some point in my rambling

ok that's enough drunken posting

 

[nemesis.ie]

6970 @ OCUK will be under £350.

They are usually on the expensive side (especially at launch) so with luck prices will be lower when there is better availability or after the hols. We can hope.