AMD Cayman info (or rumor)

Quote:

---

Originally Posted by hurrdurr

Also I believe Techpowerup had a review (of GTX 480 or 470?) that compared two exactly same clocked Fermi's, one with better custom cooling and one with stock cooling. The one with better custom cooling (and temperatures) consumed a lot less power, in spite of the fact that the two GPU's were identical.

---

If they compared two different cards (in attempt to figure out about the power consumption) to each other and the only difference was the cooling solution, that was a wrong way to go. The right way would've been to test both cooling solutions in both cards and measure. Twice. Most probably most of the difference was due to better bin GPU and better bin VRAM. But sure temperature does play a role in power consumption of a chip, there are many other factors too, which vary from card to card.

It would be quite interesting to see accurate measurements though.

Quote:

---

Originally Posted by Calmatory

If they compared two different cards (in attempt to figure out about the power consumption) to each other and the only difference was the cooling solution, that was a wrong way to go. The right way would've been to test both cooling solutions in both cards and measure. Twice. Most probably most of the difference was due to better bin GPU and better bin VRAM. But sure temperature does play a role in power consumption of a chip, there are many other factors too, which vary from card to card.

It would be quite interesting to see accurate measurements though.

---

I don't think so. After seeing that the overclocked, custom cooled 480 (which was Zotac's version I believe) actually consumed less power than the stock 480, the guys did another test: They changed the card's temperature (I think by controlling the fan speed) and recorded the temperatures. Results were clear, higher temperature meant more power drawn; and that was the source of the original differences in the temperatures of the stock and Zotac cards.

Quote:

---

Originally Posted by Calmatory

If they compared two different cards (in attempt to figure out about the power consumption) to each other and the only difference was the cooling solution, that was a wrong way to go. The right way would've been to test both cooling solutions in both cards and measure. Twice. Most probably most of the difference was due to better bin GPU and better bin VRAM. But sure temperature does play a role in power consumption of a chip, there are many other factors too, which vary from card to card.

It would be quite interesting to see accurate measurements though.

---

he did that. he took the card with better temps and better power draw and simply turned down the fan speed and watched it go right back up. he then tested it every few C and got us a curve. about every 1.5C reduced power by 1W

it shouldnt be too hard to simply search all of W1ssard's posts in here and look for a fermi or 480 thread

That would make sense since heat is the leakage and would need more power to compensate. The higher the temps the more leakage will happen and cause you to use more power for the same purpose or job. I think I made sense... :shrug:

Quote:

---

Originally Posted by Manicdan

he did that. he took the card with better temps and better power draw and simply turned down the fan speed and watched it go right back up. he then tested it every few C and got us a curve. about every 1.5C reduced power by 1W

it shouldnt be too hard to simply search all of W1ssard's posts in here and look for a fermi or 480 thread

---

The power draw/temperature relation is quite substantial due to the excessive power usage of the GTX480, but yeah power consumption definitely does affect power draw.

found it:

Quote:

---

Originally Posted by W1zzard

http://www.techpowerup.com/reviews/Z...owervstemp.gif
there is the data, linear fit is y = 1.1975x + 215.29

basically for every °C that the card runs hotter it needs 1.2W more power to handle the exact same load

---

Yeah the chip was big enough that each centigrade changed the internal resistance enough to register on a watt meter. That's seriously impressive, it gives you a idea of how huge Fermi is. 1 watt/C is pretty shocking.

2 connectors. Rumour says it's Antilles.

http://i0.sinaimg.cn/IT/cr/2010/1022/1336383876.jpg

that looks different.. enlarged photos anywhere?

that looks too small for antilles given thats a zalman vf3000 cooler there meant for 5870 and should be pin compatible with 6970

Quote:

---

Originally Posted by Jowy Atreides

2 connectors. Rumour says it's Antilles.

http://i0.sinaimg.cn/IT/cr/2010/1022/1336383876.jpg

---

It's not. http://www.xtremesystems.org/forums/...postcount=1013

Quote:

---

Originally Posted by GoldenTiger

It's not. http://www.xtremesystems.org/forums/...postcount=1013

---

thats a killer 6850!
and o correct myself cayman is not confirmed compatible with 5870 coolers. rather 6870 is.

Antilles is a dual Cayman card, PERIOD. How can AMD achieve that, what kinda crippling that needs to be done to do that ? I think they know better than the rest of us. We can speculate, but Antilles as a dual Cayman card is a FACT known from HD 68xx series launch presentation slide.

IIRC, GTX 275 is a 215 w TDP card, while GTX 295 (dual GTX 275 card) is a 296 w TDP card, so it CAN be done, and i can see this kinda analogy in work for Antilles (speed & voltage binning, reduced clock & voltage).

Quote:

---

Originally Posted by spursindonesia

Antilles is a dual Cayman card, PERIOD. How can AMD achieve that, what kinda crippling that needs to be done to do that ? I think they know better than the rest of us. We can speculate, but Antilles as a dual Cayman card is a FACT known from HD 68xx series launch presentation slide.

IIRC, GTX 275 is a 215 w TDP card, while GTX 295 (dual GTX 275 card) is a 296 w TDP card, so it CAN be done, and i can see this kinda analogy in work for Antilles (speed & voltage binning, reduced clock & voltage).

---

As mentioned.. TDP != power consumption, particularly with AMD's ratings

Quote:

---

Originally Posted by Iconyu

Yeah the chip was big enough that each centigrade changed the internal resistance enough to register on a watt meter. That's seriously impressive, it gives you a idea of how huge Fermi is. 1 watt/C is pretty shocking.

---

temperature dependence of transistor properties is more complex than that. many different properties change with temperature from leakage, to velocity saturation, and different voltages may reverse these effects.

dont take ratings for granted. amd said 150 watts and >150 watts for 6850 and 6870 but reality is around 120w/145w respectively.

so if they say 300w for antilles im more inclined to believe 270w real draw and 225w for cayman xt

Quote:

---

Originally Posted by zerazax

As mentioned.. TDP != power consumption, particularly with AMD's ratings

---

Still AMD has strived to keep its own definition of TDP, whatever that is, under 300W. If Cayman is 255W, the amount of cutting down AMD has to do, if it still wants to stay under 300W, is insane.

Quote:

---

Originally Posted by hurrdurr

Still AMD has strived to keep its own definition of TDP, whatever that is, under 300W. If Cayman is 255W, the amount of cutting down AMD has to do, if it still wants to stay under 300W, is insane.

---

Well, i don't think Cayman XT's TDP will be far north of 200 w level (210-220 w seems probable). Considering the suggested architectural change compared to Bart & die size compared to Cypress, i certainly hope that AMD will successfully pull a better performance/watt in this new flagship, highend single chip of them.

Bart XT needs to be clocked so high to be rated @151 w TDP, yet against HD 5850 that's rated the same, the actual power usage is slightly smaller. If Cayman actually delivers a new, more efficient architecture compared to what was applied in the current products, can be clocked lower & given lower voltage, i think they can achieve that, and make Antilles a worthwhile card valuewise (not just king of e peen, faster but not good enough to warrant its price tag).

Quote:

---

Originally Posted by hurrdurr

Still AMD has strived to keep its own definition of TDP, whatever that is, under 300W. If Cayman is 255W, the amount of cutting down AMD has to do, if it still wants to stay under 300W, is insane.

---

its pci express compliant so they can put pci express stickers on the boxes .. and it leaves non refference design headroom for oc version ...

or oc it yourself and break the 300watt tdp that amd had to obey

Quote:

---

Originally Posted by Sn0wm@n

its pci express compliant so they can put pci express stickers on the boxes .. and it leaves non refference design headroom for oc version ...

or oc it yourself and break the 300watt tdp that amd had to obey

---

AMD cannot release a severely cut down card that will barely be faster than its single GPU version (if 6970's really 255W).

Quote:

---

Originally Posted by hurrdurr

AMD cannot release a severely cut down card that will barely be faster than its single GPU version (if 6970's really 255W).

---

i know it wouldnt make sense .. but even if its only 40% faster then a 6970 its still a good card ... and it might leave us with oc refference or non refference design ...


ill wait after the nda ends to see if its a worthwile purchase or should i get 2 6850

HD5870's TDP was 180W, 5970's was 300W. And yet 5970 was only 40% percent faster than 5870 on average.

If the case is 255W vs 300W, even with improved CF scaling I don't think a 40% figure is possible.

Quote:

---

Originally Posted by hurrdurr

HD5870's TDP was 180W, 5970's was 300W. And yet 5970 was only 40% percent faster than 5870 on average.

If the case is 255W vs 300W, even with improved CF scaling I don't think a 40% figure is possible.

---

they said they had some sort of buffer chip on the boards and other stuff so it might be possible to have such performance .. but i guess we will have to wait for proper reviews

If its 255W is a big question. The slides only said > 225W, which can mean anything from 226W up to 300W of course.

That said, the 6850's and 6870s are getting a lot of scaling in CF, so who knows if some hardware changes were made to allow for this. If Cayman is a superset of Barts' features, then the scaling may be there, even with downclocked/cut-down cores.

That said, it seems like downclocking might be the biggest difference in power draw. For instance, they can set the 6970 to be clocked very high at stock, say 900 or 950 MHz, and put Antilles at 725 with the same overall core. It depends greatly on how much they want the 6970 to perform as a single GPU. They might be pulling out all stops to claim the single GPU crown, and so Cayman XT might be far higher clocked than it might otherwise need be.

Quote:

---

Originally Posted by zerazax

If its 255W is a big question. The slides only said > 225W, which can mean anything from 226W up to 300W of course.

That said, the 6850's and 6870s are getting a lot of scaling in CF, so who knows if some hardware changes were made to allow for this. If Cayman is a superset of Barts' features, then the scaling may be there, even with downclocked/cut-down cores.

That said, it seems like downclocking might be the biggest difference in power draw. For instance, they can set the 6970 to be clocked very high at stock, say 900 or 950 MHz, and put Antilles at 725 with the same overall core. It depends greatly on how much they want the 6970 to perform as a single GPU. They might be pulling out all stops to claim the single GPU crown, and so Cayman XT might be far higher clocked than it might otherwise need be.

---

that's exactly what they did with 5970, down-clock 2 5870 cores to meet tdp's

Anyone wanna speculate what the Cayman is going to be vs current GTX 480. I think 20% more than a GTX 480 would be be kinda of 'meh' in between disappointing and impressed.

Regarless of anything else, if they can get a single chip to perform as good as, or close to it's previous dual chip and be able to bring that down to the $200-$300 market, that's a pure win. But it probably wouldn't hit $200 for a while.


[edit]

And if all that were true and pigs could fly, the 32nm version might have been untouchable. I still think it makes sense though.

Quote:

---

Originally Posted by zerazax

If its 255W is a big question. The slides only said > 225W, which can mean anything from 226W up to 300W of course.

That said, the 6850's and 6870s are getting a lot of scaling in CF, so who knows if some hardware changes were made to allow for this. If Cayman is a superset of Barts' features, then the scaling may be there, even with downclocked/cut-down cores.

---

According to Dave Baumann at B3D there is no changes made to crossfire in Barts and AMD would be stupid if they didn't promote changes that made CF scale better.

Quote:

---

Originally Posted by marten_larsson

According to Dave Baumann at B3D there is no changes made to crossfire in Barts and AMD would be stupid if they didn't promote changes that made CF scale better.

---

Good to know then that changes are likely in drivers and/or bottlenecks-removed in the architecture

Quote:

---

Originally Posted by kadozer

Anyone wanna speculate what the Cayman is going to be vs current GTX 480. I think 20% more than a GTX 480 would be be kinda of 'meh' in between disappointing and impressed.

---

480 plus 20 percent makes it on par with HD5970. If that's the case, it's still very impressive that this was actually done in the same node as 5870. It's a little bigger but if it's really 40% faster, that's something everyone's gotta congratulate AMD for. The guys are incredible on performance / mm2.

Quote:

---

Originally Posted by zerazax

Good to know then that changes are likely in drivers and/or bottlenecks-removed in the architecture

---

Well, it's kind of hard to tell. It was a response to a question whether or not the incredible CF scaling with Barts was to be seen with 5000-series as well.

Quote:

---

Quote:
Originally Posted by Unknown Soldier View Post
Question Dave: Will the crossfire benefits also come to 58xx and older hardware with the new drivers? I know you said everything is set default but I don't see the 5850 scaling as well as the 6850 in crossfire. Is that because reviewers used old drivers for the 5850?
Posted by Dave Baumann:
Personally I'm not aware of any Crossfire changes in the driver, other than having the latest profiles, which are generic across all Crossfire capable boards.

---

The post is here:
http://forum.beyond3d.com/showpost.p...postcount=3960

Quote:

---

Originally Posted by hurrdurr

HD5870's TDP was 180W, 5970's was 300W. And yet 5970 was only 40% percent faster than 5870 on average.

If the case is 255W vs 300W, even with improved CF scaling I don't think a 40% figure is possible.

---

Calm down Cayman will not consume 255w of power that is simply the absolute figure the chip could draw with 8+6 pin power connectors.
Im certain the figure will be around 220 max.

so cayman will be faster and cost less, wow... lots of useful speculation in this thread ^^

It also :banana::banana::banana::banana:ing prints money...

"I don't care"

:D


The people AMD/ATI rely on have let them down and forced them to tweak a design for 40nm instead of moving to 32nm. AMD/ATI have done this and are looking at having a big % boost in performance at a dis-proportionately smaller increase in power consumption and a similar price point to where the 5870 is now.

...whilst facing the possibility of lower profit-per-wafer because the die size has increased

How can anyone be disappointed with that?

The top single-GPU of a new generation usually beats/matches the double-GPU of last generation in performance and heat/power, with almost same price of the last gen single-GPU. It has been the rule of thumb in the past.

When AMD chooses to play nVidia's number-game in this round, then Cayman should beat the 5970 in performance and heat/power usage, with almost the same price of 5870 . If they can do that, then it's a good move, but that's going to be hard on the same 40nm node.

It will be inserting to see if they can pull off a "new generation" successfully, or it will be a re-fresh, re-brand, or evelution.

Quote:

---

Originally Posted by Sam_oslo

The top single-GPU of a new generation usually beats/matches the double-GPU of last generation in performance and heat/power, with almost same price of the last gen single-GPU. It has been the rule of thumb in the past.

When AMD chooses to play nVidia's number-game in this round, then Cayman should beat the 5970 in performance and heat/power usage, with almost the same price of 5870 . If they can do that, then it's a good move, but that's going to be hard on the same 40nm node.

It will be inserting to see if they can pull off a "new generation" successfully, or it will be a re-fresh, re-brand, or evelution.

---

Go away please. :shakes:

How would YOU have done the naming? HD 5860 and HD5890? HD 5960/5990? HD5999 for Antilles? X2? Why X2 when there is no X2 for dual Cypress board? Doesn't the 4+1 -> 4D transition warrant a new gen naming? What about the rumoured upcoming Fusion products "polluting" the naming scheme for lower end SKUs? What about the possible future plans to introduce Caicos and Turks to lower end parts to replace Redwood and Cedar? The fact which Nvidia leaked that OEMs want new gen stuff badly to ship with their products just gives more of a reason to not keep them as HD5xxx parts. Do you honestly believe that it would've been better for consumers to have 4-6 new cards in HD5xxx series during the next few months? And that Nvidia would've gotten away with next gen stuff while AMD couldn't have delivered it until 28nm? What do you think that these "average joes" who supposedly feel pissed because of HD68xx would have felt about AMD because they "failed to bring out new stuff" while Nvidia would push GTx 5xx series?

And people still do not understand what maximum board power and TDP stands for, and what's the actual difference between those two even though it has been discussed multiple times here on XS and other boards during the past few years or so, and was clarified by SKYMTL recently. Sigh. :(

Quote:

---

Originally Posted by Calmatory

Go away please. :shakes:

How would YOU have done the naming? HD 5860 and HD5890? HD 5960/5990? HD5999 for Antilles? X2? Why X2 when there is no X2 for dual Cypress board? Doesn't the 4+1 -> 4D transition warrant a new gen naming? What about the rumoured upcoming Fusion products "polluting" the naming scheme for lower end SKUs? What about the possible future plans to introduce Caicos and Turks to lower end parts to replace Redwood and Cedar? The fact which Nvidia leaked that OEMs want new gen stuff badly to ship with their products just gives more of a reason to not keep them as HD5xxx parts. Do you honestly believe that it would've been better for consumers to have 4-6 new cards in HD5xxx series during the next few months? And that Nvidia would've gotten away with next gen stuff while AMD couldn't have delivered it until 28nm? What do you think that these "average joes" who supposedly feel pissed because of HD68xx would have felt about AMD because they "failed to bring out new stuff" while Nvidia would push GTx 5xx series?

And people still do not understand what maximum board power and TDP stands for, and what's the actual difference between those two even though it has been discussed multiple times here on XS and other boards during the past few years or so, and was clarified by SKYMTL recently. Sigh. :(

---

Learn to respect others ideas, it's not up to you to invite or dismiss others, kid.

If AMD is going to use a new generation-number (with a new first-digit) then have to deliver the performance/power accordantly, otherwise a re-fresh, re-brand, or evolution should keep the generation-number.

It remains to be seen if Cayman can beat/match the performance/power 5970, and if it can't it would be a disappointment for many who are expecting to see a a real new generation.

But it was you who said you needed a die shrink to call it a new generation. By your logic AMD should have never named Barts or Cayman as the 6xxx series; in spite of the fact that there are at least 5 new cards on this series. 5750 5770 5830 5850 5870 5970... where do you add five more cards? And maybe even more difficultly, how do you name your new cards so that the performance is in accord with the older cards and their performance?

Plus, by your logic the 4800 series shouldn't be called a new generation because they have been on the same node with 3800s. Yet 4800s performance blew 3800's away completely.

Quote:

---

Originally Posted by hurrdurr

But it was you who said you needed a die shrink to call it a new generation. By your logic AMD should have never named Barts or Cayman as the 6xxx series; in spite of the fact that there are at least 5 new cards on this series. 5750 5770 5830 5850 5870 5970... where do you add five more cards? And maybe even more difficultly, how do you name your new cards so that the performance is in accord with the older cards and their performance?

Plus, by your logic the 4800 series shouldn't be called a new generation because they have been on the same node with 3800s. Yet 4800s performance blew 3800's away completely.

---

I personally mean you need a shrink to archive a real new generation. Because it's really hard to beat the performance and power of the dual-GPU of the last generator with a single-GPU based on the same 40nm.

Vi discussed this matter widely in Nvidia confirms the GTX 580 , and I'm not going to repeat it, but the final conclusion/undressing was that you need the architectural-changes who can archive a new level of performance/power to call it a "new generation".

It has been the rule of thumb in the past that, the top single-GPU of a new generation usually beats/matches the double-GPU of last generation in performance and heat/power, with almost same price of the last gen single-GPU. I want to see Cayman doing it, otherwise it will be a "fake" new generation.

Fortunately the semiconductor industry does not work on what one person thinks.
It has always worked on what is possible at any given time while being financially viable.
Even Intel, which has :banana::banana::banana::banana:loads of money and invests heavily in its own fab does this. They release new architecture on old fab and then shrink it mid way in its life.
This whole screw-up of 40nm and delayed move to 32nm was bound to happen sometime in GFX industry. The pace at which this industry has moved is the reason we are where we are. This was just waiting to happen.
The whole double the performance / frequency every generation bubble burst in CPU industry a while back. It is waiting to burst in GPU industry. It will be a miracle and incredible feat on AMD's part if Cayman ends up being as fast as 5970.
At the moment 32nm cant happen. Period.. Move on and concentrate on what we might get.

It will be an achievement in itself if they can manage even 10% performance increase over GTX 480 while being a smaller, less power hungry chip. Enough has changed on the chip level to call it a new series. Enough time has passed to call it a new series. Nvidia was late to the party, not AMD.

here we go!

http://img221.imageshack.us/img221/5...7478021102.png

http://img233.imageshack.us/img233/2...8290332101.jpg

comments are welcome :)

Quote:

---

Originally Posted by Funky

Fortunately the semiconductor industry does not work on what one person thinks.
It has always worked on what is possible at any given time while being financially viable.
Even Intel, which has :banana::banana::banana::banana:loads of money and invests heavily in its own fab does this. They release new architecture on old fab and then shrink it mid way in its life.
This whole screw-up of 40nm and delayed move to 32nm was bound to happen sometime in GFX industry. The pace at which this industry has moved is the reason we are where we are. This was just waiting to happen.
The whole double the performance / frequency every generation bubble burst in CPU industry a while back. It is waiting to burst in GPU industry. It will be a miracle and incredible feat on AMD's part if Cayman ends up being as fast as 5970.
At the moment 32nm cant happen. Period.. Move on and concentrate on what we might get.

It will be an achievement in itself if they can manage even 10% performance increase over GTX 480 while being a smaller, less power hungry chip. Enough has changed on the chip level to call it a new series. Enough time has passed to call it a new series. Nvidia was late to the party, not AMD.

---

If you have been following, I'm not blaming AMD only, I've been making a scene in GTX 580 thread too. It's matter of "rules and definitions" that these guys (both of them) are redefining in this round. But AMD has started this new number-game with a "fake" new generation of 68xx, in this round.

What Cayman can do remains to be seen, and I'm not concluding it's a "fake" generation" yet. I'm sure it's a improvement, achievement, evolution etc.. of 40nm node. I'm still hoping AMD can deliver what they promising. I will have the same expectation about nVidia's promises for GTX 580 too actually.

But Cayman needs to beat the performance/power of 5970 to become a "real" new generation. I hope it can do it, and I believe many are expecting it, and will get disappointed otherwise.

http://tof.canardpc.com/view/16926cc...4f94bb4c83.jpg

~ HD5970 level

Quote:

---

Originally Posted by SpuTnicK

here we go!

*snip*

comments are welcome :)

---

Something is telling me that is only 6850 or 6870 crossfire :shrug:

Well the performance wouldn't suprise me if the rumors about 255W TDP for this card would be correct, compare that to 151W of Barts XT, for 100W more it better be like at least close to HD5970 or beating GTX 480 by like 15%.

Quote:

---

Originally Posted by Sam_oslo

If you have been following, I'm not blaming AMD only, I've been making a scene in GTX 580 thread too. It's matter of "rules and definitions" that these guys (both of them) are redefining in this round. But AMD has started this new number-game with a "fake" new generation of 68xx, in this round.

What Cayman can do remains to be seen, and I'm not concluding it's a "fake" generation" yet. I'm sure it's a improvement, achievement, evolution etc.. of 40nm node. I'm still hoping AMD can deliver what they promising. I will have the same expectation about nVidia's promises for GTX 580 too actually.

But Cayman needs to beat the performance/power of 5970 to become a "real" new generation. I hope it can do it, and I believe many are expecting it, and will get disappointed otherwise.

---

I still dont understand your fixation about the "fake" generation even with 68xx.
I see it as a strategic move. Its made not by looking at the past, but the future. As I see it, it is first step in to making way for full fledged Fusion takeover of low end / mainstream graphics requirements of consumers.
There might still be a 67xx for this series, but not in the future. the x7xx and x6xx performance range might be taken over by Fusion chips.

Quote:

---

Originally Posted by Olivon

image
~ HD5970 level

---

what is this? maybe its HD5970, or 470 SLI, Who knows?

Quote:

---

Originally Posted by SpuTnicK

here we go!

comments are welcome :)

---

The Heaven screenshot has 6800 GPU in it. Nothing to discuss. Fake.

Quote:

---

Originally Posted by Sam_oslo

It remains to be seen if Cayman can beat/match the performance/power 5970, and if it can't it would be a disappointment for many who are expecting to see a a real new generation.

---

Good, that means those who are dissapointed (you) can skip this serie and leave more cards for those who will be happy to get a card that might be about only 10% away from 5870 X-fire performance (or rather act like 5850 X-fire on steroids without micro-stuttering and scaling drama).

I expect HD6950 little under performance of GTX480, TDP 200-220W and price under 300$
HD6970, +20% on GTX480 (in game), TDP 250-275W and price under 400$.

Good lord, why are the simpletons always the stubborns too ? :shakes:

Facts are given, mistakes are shown, yet we're always back to square one, :dammit: and keep :horse:.

:rolleyes:

Quote:

---

Originally Posted by Sam_oslo

I personally mean you need a shrink to archive a real new generation. Because it's really hard to beat the performance and power of the dual-GPU of the last generator with a single-GPU based on the same 40nm.

Vi discussed this matter widely in Nvidia confirms the GTX 580 , and I'm not going to repeat it, but the final conclusion/undressing was that you need the architectural-changes who can archive a new level of performance/power to call it a "new generation".

It has been the rule of thumb in the past that, the top single-GPU of a new generation usually beats/matches the double-GPU of last generation in performance and heat/power, with almost same price of the last gen single-GPU. I want to see Cayman doing it, otherwise it will be a "fake" new generation.

---

By your logic Nvidia should still be on the 200 series of cards since the 300 series did not out pace the GTX 295. Also, since the 400 series has no dual gpu variant, the 500 series can't possibly exist in any shape or form until a 400 series dual gpu is made. Only after Nvidia produces that dual gpu variant, can they proceed to make a high end single chip 500 series by your logic.

Quote:

---

Originally Posted by spursindonesia

Good lord, why are the simpletons always the stubborns too ? :shakes:

Facts are given, mistakes are shown, yet we're always back to square one, :dammit: and keep :horse:.

:rolleyes:

---

+1

I'v always wondered this to...

Well I'm going to call Antilles the ATI 9800 series11.2X40 (naming scheme as follows - 9xxx denotes a herald to the old performance king, the x8xx denotes performance range, xx00 denotes i needed a longer name, series 11.xxxx denotes the DX11 nature of the architecture, xx.2xxx denotes 2gpucores and the x40 denotes manufacturing process of 40nm so there).
All who want my 9800 series 11.2X40 sign up below :rofl:, buy it now and I'll throw in a 9800 series 11.2X40SE for free when they are released 2 months down the line :ROTF:

for those that are anxious for Cayman, this will be the 9600 Series.

Quote:

---

Originally Posted by Sam_oslo

I personally mean you need a shrink to archive a real new generation.

---

Quote:

---

Originally Posted by Sam_oslo

But Cayman needs to beat the performance/power of 5970 to become a "real" new generation. I hope it can do it, and I believe many are expecting it, and will get disappointed otherwise.

---

So which is going to be before you'll indulge in calling the 6-series "new gen"? New process or new performance level?

And may I come with a kind suggestion? Stop bickering about a couple of digits. Numbering/naming is a marketing move and thus not really relevant on a tech forum.
They could call them "The Pink Love" -series for all I care... features and performance (or performance/watt) is what matters by the end of the day.

Also just out of curiosity: Nvidia has also been playing with numbers at least for the past 3-4 gens yet you don't seem to be so adamant about it when that happens - Are you biased?

in the time it took ATI to go from 2xxx to 6xxx, nvidia went from 8xxx to 5xx
thats 5 vs 8 generations. even though its only been 4 years?

8 generations?? Maybe 5 if you want to count GT300 series, wich is for OEM's:

8800 series
9800 series
200 series
*300 series*
400 series

And from ATi:

HD2000 series
HD3000 series
HD4000 series
HD5000 series
HD6000 series

IMHO ATi is 1 generation ahead, or at least 6 months ahead.

Quote:

---

Originally Posted by Andi64

8 generations?? Maybe 5 if you want to count GT300 series, wich is for OEM's:

8800 series
9800 series
*100 series*
200 series
*300 series*
400 series

And from ATi:

HD2000 series
HD3000 series
HD4000 series
HD5000 series
HD6000 series

IMHO ATi is 1 generation ahead, or at least 6 months ahead.

---

Actually they're about equal.

for the 100 series and 300 series, there where actually some OEM only GPUs being made, like GT 310 or 320

and the rumor is a 580 is coming, so i counted that too, even though its not 100% sure, but matches their style.

so 5 vs 7

You have

1: 8xxx - retail and OEM
2: 9xxx - retail and OEM
3: 1xx series : OEM only, GT 120 for example

http://www.nvidia.com/object/product...gt_120_us.html

4: 2xx series: retail and OEM
5: 3xx series: OEM only, GT320 for example

http://www.nvidia.com/object/product...gt_320_us.html

6: 4xx series: retail and OEM
rumoured 5xx series would be nr. 7

While ati has only 5 series, from 2k to 6k.

Quote:

---

Originally Posted by Andi64

8 generations?? Maybe 5 if you want to count GT300 series, wich is for OEM's:

8800 series
9800 series
200 series
*300 series*
400 series

And from ATi:

HD2000 series
HD3000 series
HD4000 series
HD5000 series
HD6000 series

IMHO ATi is 1 generation ahead, or at least 6 months ahead.

---

for nvidia it was more like

8800 series
8800 series
8800 series
9800 series
9800 series
9800 series
9800 series
9800 series
9800 series
200 series
*300 series*
400 series

:D:D

na, i'm just kidding. i don't get why people get so angry about product naming and new generations.

6870 are the same to the 5870 as the 5770 was to the 4870/4890. just because it doesn't follow the x770 naming scheme everyone is going nuts. seriously guys, get a grip!

look at the price/performance and you'll see where these cards belong. there's a good fight between the 6850/6870 and the gtx460/oc'ed in the mainstream gpu market - both, ati and nvidia, offer decent performance for a very good price :shrug:

Is this the "AMD Cayman info (or rumor)" thread or the "I am pissed about the Number Scheme" thread?

I am hoping Cayman can reach or exceed 5970, that's about it really. We need some leakage
to get this thread moving back on topic.

Quote:

---

Originally Posted by Gilgamesh

I expect HD6950 little under performance of GTX480, TDP 200-220W and price under 300$
HD6970, +20% on GTX480 (in game), TDP 250-275W and price under 400$.

---

I think HD6950 will at least match GTX480 if Cayman is really > Cypress size. Barts gets 5850 performance where Cypress was 30% larger. Granted, Cayman apparently has more changes done to it than Barts, and a bigger memory controller, but even assuming the same efficiencies, Cayman XT should be pushing 5970+ territory. Of course, Antilles is going to be nuts :eek:

Quote:

---

Originally Posted by Florinmocanu

for the 100 series and 300 series, there where actually some OEM only GPUs being made, like GT 310 or 320

---

Yeah, GT 300 series was used for notebooks as well

Quote:

---

Originally Posted by RaZz!

na, i'm just kidding. i don't get why people get so angry about product naming and new generations.

6870 are the same to the 5870 as the 5770 was to the 4870/4890. just because it doesn't follow the x770 naming scheme everyone is going nuts. seriously guys, get a grip!

look at the price/performance and you'll see where these cards belong. there's a good fight between the 6850/6870 and the gtx460/oc'ed in the mainstream gpu market - both, ati and nvidia, offer decent performance for a very good price :shrug:

---

Most people don't care once they get... well except the trolls and diehard fans :mad:

I'm curious about something else guys - the naming scheme.

If AMD keeps doing this, enxt series will be 7000 series, at the end 2011 8000 series, that means in ~24 months we will have Radeon 9800 ...again

I feel old.

Quote:

---

Originally Posted by KoHaN69

I'm curious about something else guys - the naming scheme.

If AMD keeps doing this, enxt series will be 7000 series, at the end 2011 8000 series, that means in ~24 months we will have Radeon 9800 ...again

I feel old.

---

Whats interesting is that the R300 came out in 2002, so being back to the 9800's in 2012 would mean 10 exactly generations in 10 years. Pretty awesome

Quote:

---

Originally Posted by SpuTnicK

here we go!
[...]

comments are welcome :)

---

this is the same heaven screenshot that appeared over a month ago and was proven faked/photoshopped. the chart just represents the fps from this nerfed screenshot.

we don't know if there's some truth to the screenshot, but it's obvious that it was edited (compare the fonts of the different results to a legit screenshot of the scores and you'll notice the difference)

Why are people expecting the 6890(6870?) > 5890?

The 5870 wasnt better then the 4870X2?

5870 is slightly faster than a 4870x2

Quote:

---

Originally Posted by Klarko

Why are people expecting the 6890(6870?) > 5890?

The 5870 wasnt better then the 4870X2?

---

4870 > 3870x2
5870> 4870x2

Quote:

---

Originally Posted by zerazax

Whats interesting is that the R300 came out in 2002, so being back to the 9800's in 2012 would mean 10 exactly generations in 10 years. Pretty awesome

---

and no more shrink to help us achieve better performance ... LOL i doubt it will move on much faster after 22nm ...

Quote:

---

Originally Posted by Sn0wm@n

and no more shrink to help us achieve better performance ... LOL i doubt it will move on much faster after 22nm ...

---

Are you flamebaiting? You don't think there will be any improvements in technology past 22nm? :ROTF:

Jeez, this thread is just chock full of epic fail.

Quote:

---

Originally Posted by eXceededgoku

Are you flamebaiting? You don't think there will be any improvements in technology past 22nm? :ROTF:

---

Don't you know the world stops when we hit 22nm? :shrug:

Quote:

---

Originally Posted by eXceededgoku

Are you flamebaiting? You don't think there will be any improvements in technology past 22nm? :ROTF:

---

there will be improvement for sure .. but we wont be getting them each year or 2 years .... the industry is starting to slow down ...

and no this wasnt a flamebaiting post

Quote:

---

Originally Posted by [XC] gomeler

Jeez, this thread is just chock full of epic fail.



Don't you know the world stops when we hit 22nm? :shrug:

---

so you think that 16nm will be when intel think its going to happen .. and after that 11nm and then 8 and 4 ???? where do you think intel will have problems as much as to push back their roadmap by maybe 1 or 2 years for that process ???

Quote:

---

Originally Posted by Klarko

Why are people expecting the 6890(6870?) > 5890?

---

you mean "why expect 6970 / cayman to be faster than 5970 / 2x cypress?"

okay so lets get a reality check:

http://www.pcper.com/images/reviews/1022/slide12.jpg

6800 (barts) is smaller than 5800 (cypress), but performs slightly better under the same power. which means if amd did nothing but use two 6800 barts in one product, it could ALREADY BEAT 5970.

but they aren't doing that. they are building a new, bigger, better chip. they can find NEW ECONOMIES OF SCALE. they can use NEW TECHNIQUES.

think of it like this. GF100 / gtx 480 is almost 2x the size and 2x the power of the competition. but nvidia knew this all along. nvidia designed a big, high power, high performance chip. but the performance didnt happen. the engineering failed to deliver the performance.

so now imagine amd trying this. trying their own huge chip, but not screwing it up.

Quote:

---

Originally Posted by bamtan2

you mean "why expect 6970 / cayman to be faster than 5970 / 2x cypress?"
*Snip*

---

Even better reasoning: Cypress is ~330mm^2. Barts is 255mm^2 but often matches (and probably will beat, given driver optimizations and game design changes over time) Cypress, while consuming less power.

Cayman is known to be larger than Cypress - conflicting reports put it at either largest ever for AMD (larger than R600's 420mm^2) or close to largest (confirmed its larger than Cypress though, so >330mm^2). Its also going to use more power than Cypress.

So *if* nothing else, and it reached the same scaling/optimization that Barts had, a Cypress sized Cayman should be a significant improvement. Now add in the fact that Cayman has a superset of the Barts architecture (i.e., more features/changes than Barts), and is designed from the get-go as a single GPU crown kind of GPU, then AMD has a known target and should get close to a 5970, if not beat it.

One thing is certain though: Antilles will be the fastest single video card you can buy

Given the power needed I'm convincing myself more and more that Antilles will aim for the extra-enthusiast market ($700+). Do not comply for the <300w.

The will use cayman chips so no loses.

Antilles options:
2x8Pin
2x8Pin + 1x6Pin

how many days left before the final news about 6900 series ????

Quote:

---

Originally Posted by zerazax

Even better reasoning: Cypress is ~330mm^2. Barts is 255mm^2 but often matches (and probably will beat, given driver optimizations and game design changes over time) Cypress, while consuming less power.

Cayman is known to be larger than Cypress - conflicting reports put it at either largest ever for AMD (larger than R600's 420mm^2) or close to largest (confirmed its larger than Cypress though, so >330mm^2). Its also going to use more power than Cypress.

So *if* nothing else, and it reached the same scaling/optimization that Barts had, a Cypress sized Cayman should be a significant improvement. Now add in the fact that Cayman has a superset of the Barts architecture (i.e., more features/changes than Barts), and is designed from the get-go as a single GPU crown kind of GPU, then AMD has a known target and should get close to a 5970, if not beat it.

One thing is certain though: Antilles will be the fastest single video card you can buy

---

I don't think that is better reasoning. I don't even know if that is reasoning at all

Quote:

---

Originally Posted by Nintendork

Given the power needed I'm convincing myself more and more that Antilles will aim for the extra-enthusiast market ($700+). Do not comply for the <300w.

The will use cayman chips so no loses.

Antilles options:
2x8Pin
2x8Pin + 1x6Pin

---

It's an official AMD product, they will be meeting PCI-SIG specifications.

Quote:

---

Originally Posted by bamtan2

I don't think that is better reasoning. I don't even know if that is reasoning at all

---

okay. if you say so :rofl::rolleyes:

Quote:

---

Originally Posted by Sam_oslo

Learn to respect others ideas, it's not up to you to invite or dismiss others, kid.

If AMD is going to use a new generation-number (with a new first-digit) then have to deliver the performance/power accordantly, otherwise a re-fresh, re-brand, or evolution should keep the generation-number.

It remains to be seen if Cayman can beat/match the performance/power 5970, and if it can't it would be a disappointment for many who are expecting to see a a real new generation.

---

No marketing firm has to comply with what Sam_oslo thinks... let-alone someone who doesn't even understand urach & logic.

Quote:

---

Originally Posted by Nintendork

Given the power needed I'm convincing myself more and more that Antilles will aim for the extra-enthusiast market ($700+). Do not comply for the <300w.

---

Well if they use two barts cores it should be able to shoehorn into a 300watt window and beat a 5970 with a cheaper BOM so power or pricing may not be lunacy. However without a directly competing product there is nothing in the market to put price pressure here either.

Quote:

---

Originally Posted by highoctane

Well if they use two barts cores it should be able to shoehorn into a 300watt window and beat a 5970 with a cheaper BOM so power or pricing may not be lunacy. However without a directly competing product there is nothing in the market to put price pressure here either.

---

something which all of us who love to oc would not like to see, I would prefer a more powerful chip being downclocked TBH, so that there is lots of headroom to oc

fake or not 6950 or 6970? Its from OBR blog...
http://2.bp.blogspot.com/_e5QeyTrfo3.../s1600/xxx.png

It can be anything...

I vote potato

he write this "This is not Dual GPU card or Dual VGAs setup score ... This is Score of Unreleased new VGA with only ONE GPU :)" http://obrovsky.blogspot.com/2010/10/neco-se-blizi.html

Quote:

---

Originally Posted by FlanK3r

he write this "This is not Dual GPU card or Dual VGAs setup score ... This is Score of Unreleased new VGA with only ONE GPU :)" http://obrovsky.blogspot.com/2010/10/neco-se-blizi.html

---

I put this score 2 pages ago and I think it's real. That's on par with HD5970 scores :yepp:

This slide does explain nothing. I can show you tons of such results and even with bigger figures. IMHo, FAKE!!!

Quote:

---

Originally Posted by highoctane

Well if they use two barts cores it should be able to shoehorn into a 300watt window and beat a 5970 with a cheaper BOM so power or pricing may not be lunacy. However without a directly competing product there is nothing in the market to put price pressure here either.

---

with barts they could easily fit a 225W TDP budget....

OBR is a known liar and troll. Just check his post history here. Always claiming to have new cards then delivering nothing or fakes.

I call fud.

OBR is maybe liar and boastful person, but he is also reviewer at www.pctuning.cz. Maybe its true

OBR is all talk and no show. No wonder i've had him IL'ed for years.

heh, he is banned on most czech forums :)

Take a look at this information. Now lets take a quick glance at newegg.

thermography of bart :

http://www.hardware.fr/articles/747-...thermique.html

( french Hardware.fr web site. )

LOL OBR, i wonder where his minion whose name started with "V" is right now, i guess it's not so fun anymore spreading unfounded FUDs if things turn out to be better for AMD at the moment, eh ?? :ROTF:

Quote:

---

Originally Posted by Vardant

Some news from OBR

Quote:

---

Launch date has been moved from 18th to 22nd
Most AIBs don't have the cards ready, some of them are still even working on the box design
Samples should arrive just before the launch
Limited supply of cards
No final spec. have been set yet
HD 6870 should be available in larger quantities in the middle of November, HD 6850 will follow later
It's going to be a paper launch
Nothing else planned till the end of the year
The rest has been postponed till January

---

---

BRAVO !! :clap:

EDIT:

OBR has a penchant of mixing his nVidia biased fantasy with reality. I think he does have early access to some new hardware -at times, too bad it's very hard to deliver the truth when you read the slides with certain colored glass. ;)