AMD: Sea Islands R1100 (8*** series) Speculation/ Rumour Thread

Whats the Tflops on a 925mhz 7970? And what speed would the 7970's core need to be at to reach 4.5 Tflops?

3.788. 1,110.

damn, too slow.

edit2:
So, this is what Nvidia's been expecting from Tahiti in the first place. :) 2.304 shaders and just shy of 1 GHz.
 
Over 4.5 Tflops could also mean 40 CUs (10 quad clusters) at clock-rate below 1GHz. :p

This sounds like the real winner. Would love to see 40CUs and a cool backstory about redundancy. Unless it is 8x4, and they just make a new 10 x 4 die.
 
Tenerife_B3.png


http://semiaccurate.com/forums/showpost.php?p=155075&postcount=148

nice sharp .png image.. it would be more credible if it's taken through monitor by a 1MP camera

https://en.wikipedia.org/wiki/Tenerife

Hmm.., Interesting.

But this seems to be too close to the launch of 79xx. Hard to believe this is going to launch this year.
 
Rangers said:
Probably be out before the mythical GK110 at this rate.
What does that have to do with anything? Question answer fail.

AMD has hinted to expect GCN 2.0 either this winter or next spring, so if this is coming sooner than that it would be news. I don't see how my question could have been misunderstood... but just in case... At what time are we to expect this GCN 2.0 card?
 
Probably they will clock for sure at this time using 6.0 gigabits gddr5 @ 1500MHz that will give 288GB's bandwidth, I think that will be only choice for ATI/AMD to increase bandwidth if they still stick to 384bit wide bus memory.
 
Last edited by a moderator:
I can't tell if you guys are taking this slide seriously (sarcasm is tough to hear on the internets).

But for anyone who can't spot the fake: this is a total fabrication.

There's an ugly typo "enableing", and the phrasing doesn't match real marketing material. "has been calculated at over 4.50 TeraFLOPS (single-precision FP operations) raw processing power" -awkward.

This is plainly the work of a bored computer nerd sitting at home, trolling the internets.
 
Probably they will clock for sure at this time using 6.0 gigabits gddr5 @ 1500MHz that will give 288GB's bandwidth, I think that will be only choice for ATI/AMD to increase bandwidth if they still stick to 384bit wide bus memory.

If Cape Verde is any indication they will increase cache size and/or speed to compensate.
 
I can't tell if you guys are taking this slide seriously (sarcasm is tough to hear on the internets).

But for anyone who can't spot the fake: this is a total fabrication.

There's an ugly typo "enableing", and the phrasing doesn't match real marketing material. "has been calculated at over 4.50 TeraFLOPS (single-precision FP operations) raw processing power" -awkward.

This is plainly the work of a bored computer nerd sitting at home, trolling the internets.

Not that I don't agree with you in general, but the awkward TFLOPS statement is a pretty close match to this:
"* As of November 30, 2011 the AMD Radeon™ HD 7970 Series GPU has been calculated at over 3.5 Teraflops of processing power."
Which is a straight CnP from the Southern Islands' Tech Day Deck footnotes.
 
Back
Top