AMD HD8000 Series [Or: Here we go again...]

[GodisanAtheist]

Honestly a 40% performance increase doesn't sound too far fetched for the 8000 series refresh. HD7000 had a lot of stuff going against it, really. Brand new arch and new process technology, coupled with unnecessarily strict TDP targets resulting in mountains of untapped performance in the form of clock-speed increases.

This was, in essence, AMD's GTX480. AMD has some very competent engineers and their next release, if these rumors are anywhere near target, could surprise everyone for the better.

 

[bunnyfubbles]

same rumor was about gtx 680 vs radeon 7970 .... 45% faster ? no ! only 10% ...

that rumor was when people thought it was going to be an actual 680, not a 660 relabeled as the 680

an HD8000 GPU fleshed out to be as big as these specs suggest with a TDP to match could easily be 40% faster. Although an equally large boost in MSRP would follow suit as process size won't be decreasing.

 

[bunnyfubbles]

Funny watching people troll themselves.

1080p 120hz gaming is of real benefit and hardware is lacking in modern titles. You can get close with 2 GPU's, but we don't have the CPU's to push those frames anyway in some modern titles (BF3 especially). For the first time in my gaming life, I am not very interested in these new cards. I can see it now. I install two of them, get super excited, load up BF3 and watch my framerate drop by a few FPS due to slightly more CPU overhead than with the GK104 cards.
Instead of getting 60% GPU usage, i'll be getting 30%. Sounds like something to get real excited about.

Give me a better CPU please.

My favorite quote of all time. It was the CEO of Nvidia who said this several years ago regarding CPU's...."You don't need a fast one anymore". Lawlerz.

better CPU? how about better software development

there's something seriously wrong with game development when I'm getting ~20% CPU utilization on a 4.7GHz 3930K and still am ultimately CPU limited because modern games can't (or won't) properly make use of more than 1 or 2 processing threads

 

[ShintaiDK]

better CPU? how about better software development

there's something seriously wrong with game development when I'm getting ~20% CPU utilization on a 4.7GHz 3930K and still am ultimately CPU limited because modern games can't (or won't) properly make use of more than 1 or 2 processing threads

Because multithreading is such an easy word.

Some code doesnt benefit from it. Others benefit too little to be worth it.

And the development cost skyrockets too.

Huge amount of server software is essentially not multithreaded. They just live on the concurrent users concept. Even look at MS .net compile junk. Singlethreaded. And its a nightmare if for example your exchange server is a dual 2Ghz 8 core CPU vs a 3Ghz single dualcore. It really takes around 2 hours to install SP2+Rollup3 on a 2Ghz for example

 

[BallaTheFeared]

This was, in essence, AMD's GTX480. AMD has some very competent engineers and their next release, if these rumors are anywhere near target, could surprise everyone for the better.

Only if the 480 had been slower while using more power would this statement make sense.

As it were, the 480 was faster at stock while using more power, and offered a considerable performance boost through overclocking that the 5 series simply could not compete with.

The 7970 in comparison only yields a minor performance advantage via a massive perf/watt gap.

 

[Dark Shroud]

that rumor was when people thought it was going to be an actual 680, not a 660 relabeled as the 680.

Do you have anything to back this up? I see people say it but I've never seen anything that can confirm that.

 

[Homeles]

Do you have anything to back this up? I see people say it but I've never seen anything that can confirm that.

Die size.

/irrefutable argument

Is it really a "relabeled 660?" No. But as far as chip size goes, this would normally be labeled as a x60 part.

Only if the 480 had been slower while using more power would this statement make sense.

As it were, the 480 was faster at stock while using more power, and offered a considerable performance boost through overclocking that the 5 series simply could not compete with.

The 7970 in comparison only yields a minor performance advantage via a massive perf/watt gap.

It took a gargantuan die to top the 5k series. The 400 series was not Nvidia's shining moment -- the top part wasn't even fully functional.

The actual perf/watt gap between GCN and Kepler isn't very large, with the exception of the Giggaderps Edition. Still, Nvidia really hit a home run with Kepler. Hell, they landed themselves the Apple contract.

What's going to be the hot topic this year is compute performance. I'm really interested to see who comes out on top -- Intel (no way, lol), Nvidia or AMD. GK110 just looks stupid for compute... I can't imagine how well that thing will fold...

 

[Dark Shroud]

Die size.

/irrefutable argument

Is it really a "relabeled 660?" No. But as far as chip size goes, this would normally be labeled as a x60 part.

That is not irrefutable especially considering Nvidia used a different core style more akin to AMD's stream processors this time around. And they dropped GPGPU muscle as well.

So no you can't really back that statement up. It doesn't matter how many times people repeat that rumor as fact it's still a rumor.

 

[Homeles]

That is not irrefutable especially considering Nvidia used a different core style more akin to AMD's stream processors this time around. And they dropped GPGPU muscle as well.

So no you can't really back that statement up. It doesn't matter how many times people repeat that rumor as fact it's still a rumor.

...what about die size do you not understand?

 

[GodisanAtheist]

Only if the 480 had been slower while using more power would this statement make sense.

As it were, the 480 was faster at stock while using more power, and offered a considerable performance boost through overclocking that the 5 series simply could not compete with.

The 7970 in comparison only yields a minor performance advantage via a massive perf/watt gap.

-I wasn't perfectly clear, but I meant in a new arch on a new process that will potentially be refined into a killer product, much like the GTX580 was to the 480, not only bumping performance a fair amount but bringing power/heat/noise down into the realm of the sane.

If we're comparing competitors here, the GTX 680 holds a somewhat ignoble position: its NV's first top of the line product in six generations that loses the performance crown to its red competitor.

...what about die size do you not understand?

- The GTX680 has the largest die size out of the 6xx generation. What about die size do you not understand? The top end card of the generation is the one either company releases, not the one they wished they had released. Never heard this argument when the 2900XTX was canned "The 8800GTX cleans house, but AMD only released its MID RANGE CARD"

 

[Atreidin]

It makes some people feel better to believe that a company, who doesn't care about whether they are alive or dead, only released a "mid-range" part, even though there is no other part that is higher performing. You can say whatever you want if you just redefine all the words! I guess it gives them some sad assurance in their pathetic lives to believe that their surrogate father-figure was just pulling punches against the other kid's dad.

 

[boxleitnerb]

GK104 is for a supposed highend part akin to G80/GT200(b)/GF100/GF110 uncommonly small and efficient compared to its direct predecessors. To consider it highend is a bit far fetched. I would call it beefed up midrange.

 

[Sable]

AMD were gonna release the 7971 as their highend part but they didn't because there was no competition from Nvidia.
There's no actual evidence for this but it's scientific fact.

 

[sandorski]

Bit Boys are still waiting for Nvidia/AMD to release something competitive, in order to avoid Uncompetitive Practices legal actions.

 

[BallaTheFeared]

-I wasn't perfectly clear, but I meant in a new arch on a new process that will potentially be refined into a killer product, much like the GTX580 was to the 480, not only bumping performance a fair amount but bringing power/heat/noise down into the realm of the sane.

If we're comparing competitors here, the GTX 680 holds a somewhat ignoble position: its NV's first top of the line product in six generations that loses the performance crown to its red competitor.

There isn't much difference between GF100 and GF110, FP16 is faster but has limited effect in actual results since it's use is limited.

Other than that they fixed some leaky transistors by cutting them out and slapping a higher tech cooler on the reference card instead of an old 427 with header pipes coming out the hood.

Only if you count a PR stunt that sounds like a jet on take off. If that was the case Nvidia could just release a couple cherry picked 1350 cards to reviewers and call it the "1.21Gw" edition. AMD flip flops so hard is like watching a politician, one gen it's dual gpu cards that count, the next since they can't feasibly produce a decent dual card due to high power they switch over to single with a PR stunt card that will never go on Newegg.

 

[AtenRa]

Im expecting ~20% better performance than 7970 GHz edition at almost the same wattage.

 

[Magic Carpet]

Im expecting ~20% better performance than 7970 GHz edition at almost the same wattage.

Only 20% ?, dang

 

[AtenRa]

For the specs of the OP i dont expect more.

And im sure AMD will not want to raise power consumption more than current 7970/GHz. So, at almost the same wattage a 20% average more performance at the same litho process is fine.

 

[moonbogg]

Seriously though, it should be a while before these come out from either camp right? I mean damn, talk about tech moving fast? This is ridiculous if these come out any time soon TBH.

 

[BallaTheFeared]

Seriously though, it should be a while before these come out from either camp right? I mean damn, talk about tech moving fast? This is ridiculous if these come out any time soon TBH.

I'd expect it to come out around a year after 7 series released (pretty standard within the industry), maybe sooner since nvidia has so much pressure on AMD right now (like when Nvidia popped out 5 series so quickly after 4 wasn't so well received).

Right now the ball is in AMD's court, Nvidia is in a stronger position. I'd expect 8 series around Christmas.


You have to remember this is their income source, you can't sit on tech in the tech world because the advantages of improved processes and tech are fairly large. It's like a car company, what else are they going to do to continue obtaining revenue if they aren't making a new model each year?

 

[SirPauly]

GK104 is for a supposed highend part akin to G80/GT200(b)/GF100/GF110 uncommonly small and efficient compared to its direct predecessors. To consider it highend is a bit far fetched. I would call it beefed up midrange.

Consider the Gk-104, similar to what the G-92 did, when it was their fastest GPU. Priced according to the competition.

 

[Homeles]

- The GTX680 has the largest die size out of the 6xx generation. What about die size do you not understand? The top end card of the generation is the one either company releases, not the one they wished they had released. Never heard this argument when the 2900XTX was canned "The 8800GTX cleans house, but AMD only released its MID RANGE CARD"

Holy crap, you're ignorant.

Not mid range card, mid range die. ~300mm2 is not large at all for a modern GPU. Since GT200, Nvidia flagship dies have been 500mm2+. ~300mm2 is an x60 level die size.

If you knew anything about GPUs, you'd know this.

 

[tviceman]

For the specs of the OP i dont expect more.

And im sure AMD will not want to raise power consumption more than current 7970/GHz.

Why not? The hd6970 wattage increased over the hd5870, both on the same node. The performance improvement was not "free" there.

 

[boxleitnerb]

Actually, perf/W decreased quite a bit with RV870 vs RV970. It will be interesting if AMD will commit to 250+W real world consumption (although the GHz ed. already reaches 250W at times).

 

[AtenRa]

Why not? The hd6970 wattage increased over the hd5870, both on the same node. The performance improvement was not "free" there.

HD5870 was bellow the 200W mark, HD7970 is above. They simple dont have the luxury this time to go any higher, unless they will like to follow the NVIDIA way, Huge monolithic dies of 250W+.

But 20% more shaders over the 389mm2 of Tahiti will not make a huge 500mm+ die.