• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

[wccftech] rumor: amd hawaii benchmarked in 3dmark11 firestrike AMD attacks titan

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
The Titan shouldn't even be discussed anymore. Galaxy GTX780 HOF is $690 and it will beat the Titan without any problems. It even has the world's record for the fastest single GPU on LN2. At this point 9970 needs to offer better value or performance than after-market GTX780s since many after-market 780s are now 15-17% faster than a stock 780.

Agreed! Only nVidia fans were excited about Titan, even on launch day. The 780 was a kick in the nuts to everyone who dropped $1k on a Titan.

AMD definitely needs to look at the 780 as the target card. Of course, they should shout about Titan at a grand from the rooftops, but realize it's really the 780 that they are competing against for any non biased consumer.
 
single large die at 28nm. probably upto 480 sq mm. 5.5+ billion transistors. GK110 competitor.

Definitely a possibility, I'm guessing it's somewhere in the ballpark of 440-500 sq mm. Probably towards the large end of the spectrum because AMD will want to make another push into HPC. They really need to increase their presence in that market now and encourage the use of OpenCL before Intel's next generation of Phi shows up.
 
Last edited:
Wasn't there this comment from someone at AMD about the time when all of GCN had been launched*, about how they had learned a lesson from Pitcairn?

I always interpreted that to mean that their next high-end card would drop a bit of DP performance. So maybe 9970 will be something along the lines of a doubled-up Pitcairn with whatever the learned with Bonaire plus some tweaks and a better balance of ROPs, TUs and so on. By all means a bigger die but optimised for gaming with some DP performance taken away - a bit like the idea of GK104 but not overly crippled/segmented since AMD don't have the Tesla market.

*I know rather vague - can't find the quote atm though.
 
Why would they go so big with such a weak pro market?

Not sure why you say this because the PC gaming market is growing year by year. It used to be mid-range was $200, now that has moved towards $300 or above. Gamers dont seem to care about global financial downturns or tablets/smartphones taking over most PC related tasks.
 
single large die at 28nm. probably upto 480 sq mm. 5.5+ billion transistors. GK110 competitor.

I don't think so. Only because right now, I think AMD needs to try to be nimble and light. I think Rory Read knows this. I don't think AMD's engineers are capable of making such a large die. After all, most of them were let go.
Sure, I could be wrong, but I haven't seen anything at all indicating that AMD is going to be pulling out a 450+mm2 28nm die to compete with Titan.

I absolutely COULD see them taking on Titan with an X2. Basically take Bonaire arch. and add in hardware based frame metering solution. Then put two of them on a single card. That seems to be far more plausible than a single large die from AMD.
 
They've been "nimble and light" for many generations already, not working out well for marketshare, mindshare or profits. Go big or go home for the top GPU.
 
Not sure why you say this because the PC gaming market is growing year by year. It used to be mid-range was $200, now that has moved towards $300 or above. Gamers dont seem to care about global financial downturns or tablets/smartphones taking over most PC related tasks.

Pro = Professional
 
They've been "nimble and light" for many generations already, not working out well for marketshare, mindshare or profits. Go big or go home for the top GPU.

No. They've been mismanaged for many generations. If they had been on their toes, they would never have sold their mobile division off during the biggest growth in that segment. Wow what a bad move. Anyway....
Go big or go home only works when you have the resources to go big. AMD is actually going small as is evident with their APU's and ending their FX desktop line announcement.
So, rather than go home, AMD does have X2 options with smaller dies AS LONG as they keep entirely focused on smooth crossfire performance across the board. If they do that, they can stay alive for a long while competing with X2 cards.
 
Last edited:
If they were just going to do an x2 than why have a 28nm refresh at all? They could just let the 7990 hang out at $699.

They have experience with using a larger die budget on the same node from when it became clear node durations were increasing at 40nm and they launched the 6900 series.
 
If they were just going to do an x2 than why have a 28nm refresh at all? They could just let the 7990 hang out at $699.

They have experience with using a larger die budget for when it became clear node durations were increasing at 40nm with the 6900 series.

Because 7990 doesn't have hardware "frame metering" and the 7xxx core is aging and inferior to Bonaire, even inferior to Pitcairn. I have a hunch it will be a Bonaire based x2 with a hardware based frame metering solution and called the 9xxx series.
 
I didn't think you were basing it on anything valid. Just thought that maybe you actually knew something that I hadn't seen. Thanks for the confirmation. :thumbsup:

Confirmation of what? Ah, so for you to think that I had an inside source means that you truly don't know if I was right, or if AMD has a big die to offer either. Thanks for that confirmation.

At any rate, in AMD's current state of financial and engineering talent, I think multiples of smaller better dies are far more believable than an Nvidia large die competitor. Don't you? I mean how could you not?

I just don't want to see another swarm of PR BS descend onto this forum.
 
Last edited:
No. They've been mismanaged for many generations. If they had been on their toes, they would never have sold their mobile division off during the biggest growth in that segment. Wow what a bad move. Anyway....
Go big or go home only works when you have the resources to go big. AMD is actually going small as is evident with their APU's and ending their FX desktop line announcement.
So, rather than go home, AMD does have X2 options with smaller dies AS LONG as they keep entirely focused on smooth crossfire performance across the board. If they do that, they can stay alive for a long while competing with X2 cards.

Pretty sure YOU of all people understand the merits of a single GPU dominating the top spot for the enthusiast market.

It simply allows AMD to make such statements such as "THE BEST" or "FASTEST" or "SUPREME PWNAGE" etc etc, which for mindshare, is worth so much more for their entire stack, including "PREMIUM RADEON" graphics in the mobile space and APUs.

They have tried to combat NV's top GPU with x2 before, it failed. The worse is this gen, the 7990 died before delivery due to all the recent focus on smooth frames. There are many enthusiasts who want a massive GPU, because they can then grab 2 and have an awesome CF/SLI rig that dont have issues of quad CF/SLI scaling or just plain broken quad support.
 
Also a small die x2 strategy would be harder to push in the HPC market. AMD clearly wants a bigger piece of that pie.
 
And neither do you for wondering if I did.
Yes. When people don't know things, they ask questions. Like I did.
First time anyone has asked me to source a question. Something new everyday.

Call it what you want. It was ridiculous with no basis to even ponder.
 
Call it what you want. It was ridiculous with no basis to even ponder.

It was an asinine question. AMD has the 7990 to do that already. $699 and much faster than a Titan. Obviously if this card exists, it is a single GPU. Someone just needs a pillow to sit on is all. 😛
 
Pretty sure YOU of all people understand the merits of a single GPU dominating the top spot for the enthusiast market.

It simply allows AMD to make such statements such as "THE BEST" or "FASTEST" or "SUPREME PWNAGE" etc etc, which for mindshare, is worth so much more for their entire stack, including "PREMIUM RADEON" graphics in the mobile space and APUs.

They have tried to combat NV's top GPU with x2 before, it failed. The worse is this gen, the 7990 died before delivery due to all the recent focus on smooth frames. There are many enthusiasts who want a massive GPU, because they can then grab 2 and have an awesome CF/SLI rig that dont have issues of quad CF/SLI scaling or just plain broken quad support.

Precisely why I believe AMD is now focused on frame metering. To introduce a 9xxx series with a Bonaire type core range with an X2 at the top few tiers that have hardware based frame metering in place for smooth performance across the board. As for the "FASTEST" or "SUPREME PWNAGE" mindshare, you can easily see from various interviews of AMD PR or even their engineers that they believe their X2 strategy is a winning one. Even tout them as the fastest cards on the planet and they truly believe that.
AMD may have learned their lesson the hard way via Tech Report's Scott Wasson and other sites that followed up with their own testing.
If you learn from your mistakes, you have a chance to better yourself. If you don't learn from them and repeat those mistakes, that is just the definition of crazy.
 
Last edited:
It was an asinine question. AMD has the 7990 to do that already. $699 and much faster than a Titan. Obviously if this card exists, it is a single GPU. Someone just needs a pillow to sit on is all. 😛

Very, very interesting. There isn't anything "obvious" about any of this.
It is all speculation and you know it is. I for one would be very surprised if AMD introduced a very large die for the 9xxx series.
I'm not writing it off either as that would be unintelligent.
 
Last edited:
Precisely why I believe AMD is now focused on frame metering. To introduce a 9xxx series with a Bonaire type core range with an X2 at the top few tiers that have hardware based frame metering in place for smooth performance across the board. As for the "FASTEST" or "SUPREME PWNAGE" mindshare, you can easily see from various interviews of AMD PR or even their engineers that they believe their X2 strategy is a winning one. Even tout them as the fastest cards on the planet and they truly believe that.
AMD may have learned their lesson the hard way via Tech Report's Scott Wasson and other sites that followed up with their own testing.
If you learn from your mistakes, you have a chance to better yourself. If you don't learn from them and repeat those mistakes, that is just the definition of crazy.

And you were just asking a question? Sure you were. Here you are saying, "I believe AMD is now focused on frame metering. To introduce a 9xxx series with a Bonaire type core range with an X2 at the top few tiers that have hardware based frame metering in place for smooth performance across the board." 🙄
 
And you were just asking a question? Sure you were. Here you are saying, "I believe AMD is now focused on frame metering. To introduce a 9xxx series with a Bonaire type core range with an X2 at the top few tiers that have hardware based frame metering in place for smooth performance across the board." 🙄

Here you go bud. My first post: These are questions.

"So is Hawaii actually AMD's x2 strategy? Two small dies to combat Nvidia's single large dies?
Like maybe two "Bonaire-Like" 7970 cores on a single card and call it the 9xxx series."

Then you asked in post 59:
"Any source for this speculation?"

Then I pondered things: These began my speculations:

"I believe AMD is now focused on frame metering. To introduce a 9xxx series with a Bonaire type core range with an X2 at the top few tiers that have hardware based frame metering in place for smooth performance across the board."

Without the rolling eyes of course.
Do yourself a real favor and try not to engage me. I keep things real, you do not.
 
Last edited:
Back
Top