• 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.

[bit-tech] Radeon HD 7000-series rumoured for May production

Page 11 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Sorry, but I have to laugh at the original post about 100% performance increase between generations...didnt happen.
 
Sorry, but I have to laugh at the original post about 100% performance increase between generations...didnt happen.

The cards not OUT yet? how has it not happened?? And the tsmc reporting of 45% improvement is in CORE clock's Mhz ONLY

It doesnt factor in any design changes like twice as many pipelines ext...

They can pretty much double everything on the chip with the move to 28nm and ontop of that add 45% greater clocks

so at the end of the day we may see a 145% increase in performance.
 
The cards not OUT yet? how has it not happened?? And the tsmc reporting of 45% improvement is in CORE clock's Mhz ONLY

It doesnt factor in any design changes like twice as many pipelines ext...

They can pretty much double everything on the chip with the move to 28nm and ontop of that add 45% greater clocks

so at the end of the day we may see a 145% increase in performance.

No they cannot. The 45% clock increase would use up all the power budget. If you were to double everything, you would almost double power consumption.
Power-wise it is inefficient to increase only clock speeds. It will be a combination of new architecture, more units and a little clockspeed increase.
 
It's comical looking over the first two pages of this thread and reading how everyone thought that August and September would bring high end 28nm GPU's out. It reminds me of when Silverforce started posting regularly in the forums on how the 6870...errr....6970... would be as fast as the hd5970 and would outperform the gtx480 in Unigine's Heaven by 50%.

The good old days!
 
Last edited:
Honestly I cant believe the 7xx0 will be in may, it was tapped out how many months ago and TSMC sent out 28nm wafers last month. Theres no reason it will take 6 months unless there is a redesign somewhere.
 
Honestly I cant believe the 7xx0 will be in may, it was tapped out how many months ago and TSMC sent out 28nm wafers last month. Theres no reason it will take 6 months unless there is a redesign somewhere.

happy medium necro'd this thread started last April. I guess because a rumor was wrong? 🙄
 
No more waiting i got tired of all the people telling other people to wait for bulldozer.

Look how that pile of sh*t turned out....was like waiting for DNF yes i have to use this analogy cause so far that's been the biggest disappointment of my life besides bulldozer as far as tech goes.

If you need something now if your starving on a 9800gt or some other unfortunate setup replace it NOW if you can afford it...
 
No more waiting i got tired of all the people telling other people to wait for bulldozer.

Look how that pile of sh*t turned out....was like waiting for DNF yes i have to use this analogy cause so far that's been the biggest disappointment of my life besides bulldozer as far as tech goes.

If you need something now if your starving on a 9800gt or some other unfortunate setup replace it NOW if you can afford it...

How does the Bulldozer debacle relate to SI? That aside I agree with you and believe that waiting for something down the road isn't the best advice. If you need something today, go and buy it.
 
How does the Bulldozer debacle relate to SI? That aside I agree with you and believe that waiting for something down the road isn't the best advice. If you need something today, go and buy it.

No hard evidence on how 28nm will perform for all you know it could be a marginal upgrade or perhaps a rebadge.

When the reviews come in on mature cards and they are good i willl eat my words and you can repost this to me and call me a idiot if that makes you feel better but high expectations for me anyways are low.

5770/67700 rebadge nvidia 400 series to 500 technically the same thing with lower power consumption nothing ground breaking.

Biggest move of the last 5 years was the move from dx9 to dx10 with nvidia based cards every other series has been marginal ever since...

Don't get me wrong i do hope its a vast improvement but i got my doubts sorry .
 
No hard evidence on how 28nm will perform for all you know it could be a marginal upgrade or perhaps a rebadge.

When the reviews come in on mature cards and they are good i willl eat my words and you can repost this to me and call me a idiot if that makes you feel better but high expectations for me anyways are low.

5770/67700 rebadge nvidia 400 series to 500 technically the same thing with lower power consumption nothing ground breaking.

Biggest move of the last 5 years was the move from dx9 to dx10 with nvidia based cards every other series has been marginal ever since...

Don't get me wrong i do hope its a vast improvement but i got my doubts sorry .

I doubt I'll even remember this post, never mind call you an idiot or rub your face in anything. 🙂

There have been a lot of rumors. Everything from GCN top to bottom to mostly re-badges except for Tahiti/NZ. I'm hoping it's more the prior than the latter and leaning that way myself. It shouldn't be a lot longer to wait now.
 
Well, the rumors are popping up here and there such as:

http://www.techpowerup.com/155491/Noises-About-Radeon-HD-7900-Series-with-XDR2-Memory-Grow.html

which could point to a Xmas release considering AMD officials went on record saying 7XXX by EOY 2011.


Some interesting stuff there but this bit I cannot see being accurate.

In related news, other sources told TechPowerUp that AMD could adopt a "top-to-bottom" strategy with the high-end portion of its next-generation of products. This means that AMD could launch the dual-GPU "New Zealand" graphics card first, followed by single-GPU SKUs.

Every time a new node release of GPUs comes out, they do not have a good supply at the start. Why would they release such a premium niche card that will gobble up two gpus and sell for less than those two gpus would on their own card separately ?

On past history with the 5870, if they release again before nvidia on the next node - 28nm, there is going to be huge demand for the cards. Probably larger than there was for the 5870s/5850s with people not wanting to get caught out waiting as long as they did for nvidia to release last time.

The 5870s & 5850s were out of stock everywhere shortly after release and people were camping etailer sites to pick them up.

Would be a huge waste and a poor choice in terms of making more money to release a 7990 as the first card out the door. lol. There would probably be about 1000 people with one and everyone else would be stuck waiting.
 
times are a changing, bulldozer was a huge massive flop that can not be any bigger fail then it was.. but AMD's graphics core next will not, as with anything that "improves" on the older generation tends to be better.

unlike CPU's going from 6 cores on a short pipeline to 8 gimped cores on a longer pipeline

GPU's are going from 1600+ Cores to 2500+ Cores with each core improved to be more efficient

So even though they may not double the gpu core pipelines they will make each one of them work 30% faster..

In the end we "may" see 100% performance improvement over the previous generation

They could only do the later with the 6000 series GPU's as 40nm didnt allow them to add any extra cores and in the end they did a reshuffle of resources and performance improvement per core. to get the 15% improvement over the 5000 series
 
With Killebrew out and the 3 month delay on the "newer/better/faster/stronger" strategy, I'm very curious to see how hd 7000 series shapes up.
 
I am anxious, and those that point to the bulldozer fail, I have a lot faith/trust in AMD's GPU division than its CPU one.
 
With Killebrew out and the 3 month delay on the "newer/better/faster/stronger" strategy, I'm very curious to see how hd 7000 series shapes up.

I think it might be a telling sign that Apple has decided to go with Nvidia when ivy bridge hits. (Yes, this was verified beyond Charlie at this website: http://blogs.barrons.com/techtrader...hey-pushed-out-amd-at-apple/?mod=yahoobarrons) AMD and Nvidia are going to be launching 28nm GPU's in about the same time frame, with most people reporting / expecting AMD to come to market first, yet Apple decided to change from AMD based discrete offerings to Nvidia...it makes me think Nvidia made power draw and/or performance per watt priority number 1 with Kepler.
 
Rather Apple has no confidence AMD can deliver the quantity they require in a timely manner.

Apple and NV both have a lot of bargaining power over TSMC. <-- important factor when yields are rubbish.
 
Me: AMD You said new parts are due this year!!!! WTF!!!
AMD :We didn't say which parts. Jakes on you.
Me: FUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
 
No hard evidence on how 28nm will perform for all you know it could be a marginal upgrade or perhaps a rebadge.

Highly unlikely. Almost all professional website expected HD6000 and GTX500 series to be a minor 'refresh' of the previous generation. Sure HD6000 did a more drastic move from VLIW-5 to VLIW-4, but essentially it's based on the same fundamental architecture, just re-tweaked.

Contrary to this, NV's Kepler is not just a refreshed Fermi. The architecture is supposed to have 3-4x increase in dual precision efficiency. We should see a significant increase in performance from both camps, especially in regard to Tessellation. Similar, on the AMD side, the higher variants of the HD7000 series should constitute a brand new GCN architecture -- the largest architectural change on the AMD GPU side since HD2900XT series went with a unified shader architecture.

When the reviews come in on mature cards and they are good i willl eat my words and you can repost this to me and call me a idiot if that makes you feel better but high expectations for me anyways are low.

Not sure how you can have low expectations:

1) We are moving to a 28nm process from 40nm. This alone allows transistors to work at 40-45&#37; faster speeds within the same power envelope.

2) That doesn't take into account any increases in performance that will come from architectural redesigns/major enhancements.

3) There should be a huge wave of next generation DX11 games once next generation of game consoles launch. We might see those games in 2-3 years, but they will arrive eventually. I have no doubts that current cards will not be fast enough to run them.

5770/67700 rebadge nvidia 400 series to 500 technically the same thing with lower power consumption nothing ground breaking.

For starters, HD6000 and GTX500 series were never meant to be a new generation. Sure marketing upped the numbers denoting a new generation, but among the in the know hardware enthusiasts, HD6000 and GTX500 series are not a new generation. They are just refreshes like X1900XT was to X1800XT and 7900GTX was to 7800GTX.

Biggest move of the last 5 years was the move from dx9 to dx10 with nvidia based cards every other series has been marginal ever since...

Completely untrue. GTX280 was way faster than 8800GTX. GTX480/580 is way faster than GTX280/285. In each of these cases, the performance increase was at least 50%. The GTX580 is 2.8x faster than 8800GTX was.

On the AMD side, since HD2900XT/3870 sucked, the difference in performance was even more dramatic. HD4870 doubled the performance of the 3870, while HD6970 is about 68% faster than HD4890 is.

I don't consider these marginal performance increases.

graphicscardevolution.png

Source

The problem now is unless you slap very high AA, or want 60 fps min frames or play at 2560x1600 or use multiple monitors, even today's GTX480/570/580/6950/6970 generation of cards is fast enough for 90% of games. The problem is games aren't demanding enough, not that GPUs aren't getting faster.

But look at BF3, and it becomes clear that something from HD4870/GTX260 216 era is a lot slower than today's top cards. And that's not taking into account that those cards aren't even producing the same image quality...
 
Last edited:
Interested what prices these cards come in at,i payed &#163;182 for a 5850 many moons ago,the 6950 released for 198&#163; when these where relaeased later on(these are uk etailer release prices)i will not pay more than &#163;250 for a 7950 give or take a few pounds.the pricing rumours

are making me nervous:S.
 
Last edited:
Back
Top