AMD Radeon HD 7950 delayed until February

By on January 3, 2012, 12:00 PM

Last week we posted an in-depth review of the Radeon HD 7970, the first in a series of upcoming graphics cards from AMD to utilize a 28nm fabrication process. Although we concluded that the card is extremely fast, it won’t be available to gamers worldwide until January 9.

This paper launch is uncharacteristic of AMD and led many media outlets to question the move. As a result, AMD has reportedly decided to delay the launch of the Radeon HD 7950 until early February. According to Guru of 3D, AMD made the decision to avoid similar concerns and have the card available on the market when it actually launches.

The Radeon HD 7950 is based on the same Tahiti core used in the higher-end 7970 except that four of the card's GCN (Graphics Core Next) compute units have been disabled. This will result in what AMD calls the Tahiti Pro core, where as the GPU on the 7970 is known as the Tahiti XT. According to NeoSeeker, the 7950 will feature 1792 Streaming processors and a 3GB GDDR5 frame buffer running on a 384-bit interface. The card will use the same video output configuration as the 7970, meaning it can support up to six screens.

The only unknowns at this point are the total TDP, clock speeds and price point that the 7950 will launch at. It looks like we might have to wait at least another month to get these answers unless AMD has something to share with us at CES 2012.




User Comments: 33

Got something to say? Post a comment
Ranger12 Ranger12 said:

Bummer. I'm waiting on the 7xxx series cards to come out so I can hopefully pick up another 6870 for cheap!

treetops treetops said:

same here ranger but make that a nvidia gtx 275

Ranger12 Ranger12 said:

Ya know, I thought about selling my 6870 and 5830 and buying an nvidia because I've never owned one. I've bought only ati cards since my first 9800 pro back in the day.

Dannyk0ed said:

Ya know, I thought about selling my 6870 and 5830 and buying an nvidia because I've never owned one. I've bought only ati cards since my first 9800 pro back in the day.

Everyone should at least try a NVIDIA Card

Guest said:

Is it a 'crippled' flagship card, like the HD 6950 is? Or is it a redesign, such as the original GTX 560 Ti from almost one year ago? Just curious... (unlockable ???)

treetops treetops said:

Ranger12 said:

Ya know, I thought about selling my 6870 and 5830 and buying an nvidia because I've never owned one. I've bought only ati cards since my first 9800 pro back in the day.

heh, im a bang for the buck kinda guy unless a product has burned me in the past, i have no loyalty to amd or intel. nvid or ati.

dividebyzero dividebyzero, trainee n00b, said:

Is it a 'crippled' flagship card, like the HD 6950 is? Or is it a redesign, such as the original GTX 560 Ti from almost one year ago? Just curious... (unlockable ???)

Salvage part like 6950. One-eighth of the shaders and TMU's inactive, lower specced memory IC's and slightly lower clock for core/shader.

Like the 6950, the inactive parts of the GPU core may not be damaged, they just don't meet the specification laid down by AMD for the top-bin 7970.

howzz1854 said:

if it's unlockable like the 6950, i'll get one.

Feb isn't bad, consider most people will be getting their tax return then. i can't wait til i get my tax return.

lawfer, TechSpot Paladin, said:

Ranger12 said:

Ya know, I thought about selling my 6870 and 5830 and buying an nvidia because I've never owned one. I've bought only ati cards since my first 9800 pro back in the day.

I used to own an ATI card, mind you, but if you've never owned an Nvidia card, the only thing I'd say is better is driver support. They usually release faster, more specific updates.

bandit8623 said:

lawfer said:

Ranger12 said:

Ya know, I thought about selling my 6870 and 5830 and buying an nvidia because I've never owned one. I've bought only ati cards since my first 9800 pro back in the day.

I used to own an ATI card, mind you, but if you've never owned an Nvidia card, the only thing I'd say is better is driver support. They usually release faster, more specific updates.

how do they have better driver support? each release 1 driver a month, both also release beta drivers/preview drivers.

driver support is equal. only thing different is the driver interface.

which i will agree nvidia wins there, but AMD still has better speed for the money.

amstech amstech, TechSpot Enthusiast, said:

AMD has released nearly double the driver updates to fill in a wide range in holes/bugs/gltiches etc over the past few years, compared with Nvidia.

As far as the article I think AMD is doing this so thier sales fare better, once people saw how close the GTX 580 competes with the 7970 at 1600P and above (the reason you buy a card like this) it killed any reason to upgrade for them and several others who own 570's and 6970's.

Ranger12 Ranger12 said:

I'm just an ati fanboy so yah can't really trust me when I say this but I also think ati gives me more bang for the buck. That being said, I'm also emotionally attached to my video cards. I think some of y'all can understand that whether you like ati or nvidia.

lawfer, TechSpot Paladin, said:

bandit8623 said:

lawfer said:

Ranger12 said:

Ya know, I thought about selling my 6870 and 5830 and buying an nvidia because I've never owned one. I've bought only ati cards since my first 9800 pro back in the day.

I used to own an ATI card, mind you, but if you've never owned an Nvidia card, the only thing I'd say is better is driver support. They usually release faster, more specific updates.

how do they have better driver support? each release 1 driver a month, both also release beta drivers/preview drivers.

driver support is equal. only thing different is the driver interface.

which i will agree nvidia wins there, but AMD still has better speed for the money.

Since the unification of both mobile and desktop GPU drivers (Verde Drivers), Nvidia has literally tripled the speed of which updates are released.

Depending on some games, some beta drivers are released merely a week or so after previous WHQL drivers. For example, when Battlefield 3 was released, they pushed out the 285.62 driver, but merely two weeks later they released the 285.79 driver to improve performance. In comparison, ATI took longer.

Alternatively, putting Intel out of the picture, Nvidia holds more of the GPU marketshare (partly due to its ARM offerings), and is often sponsored by game publishers and studios (as seen when you launch a game and see Nvidia in the opening credits). When I bought my GTX 560 Ti, for instance, I got a code to download Batman: Arkham City for free from Steam. That, I'd argue, is better support.

Like I said, I owned an ATI card, but after switching to Nvidia, the only thing I like better is the faster rate in which drivers are updated. Other than that, I consider ATI card to be less expensive, and perform equally and/or better.

Leeky Leeky said:

Bummer. I'm waiting on the 7xxx series cards to come out so I can hopefully pick up another 6870 for cheap!

I can't speak for across the shores, but I recently picked up a HD6870 for £135, which considering the performance it offers its already in the bargain basement price category in my mind.

red1776 red1776, Omnipotent Ruler of the Universe, said:

AMD has released nearly double the driver updates to fill in a wide range in holes/bugs/gltiches etc over the past few years, compared with Nvidia.

As far as the article I think AMD is doing this so thier sales fare better, once people saw how close the GTX 580 competes with the 7970 at 1600P and above (the reason you buy a card like this) it killed any reason to upgrade for them and several others who own 570's and 6970's.

What reviews are you reading? more pap

Adding it up from Dividebyzero:

Qoute:

EDIT: Tech Report's review in

(2560x1600)

Batman:AC +9.38% over GTX 580 (average of previous 9 reviews 19.79%)

BF3 +19.35% over GTX 580 (average of previous 14 reviews 17.38%)

Civ 5 +14.29% over GTX 580 (average of previous 3 reviews 14.56%)

Crysis 2 +24.14% over GTX 580 (average of previous 13 reviews 23.49%)

TESV +9.43% over GTX 580 (average of previous 5 reviews 19.89%)

And as for the comment-

once people saw how close the GTX 580 competes with the 7970 at 1600P and above (the reason you buy a card like this)

The gap widens when you go to multiple monitors.

From [H]ocp: ( these numbers are against an overclocked GTX 580)

The Galaxy MDT GTX 580 which, by the way is $50.00 more than the 7970

[link]

dividebyzero dividebyzero, trainee n00b, said:

^^^ Hardly surprising.

40nm vs 28nm

1.5GB frame buffer vs 3GB frame buffer

Fermi arch = compute orientated (larger core/lower core speed), VLIW5/4/GCN = fillrate orientated

Surround/Eyefinity is valid, albeit for a minority of users at present, and a stronghold for AMD in single card setups...and which isn't natively supported by reference Nvidia cards (hence the lack of review benches -both from the niche perspective and the non-support by one vendor)

The gap widens when you go to multiple monitors

3-D gaming is probably no less valid as a niche market...and the gap closes quite considerably ...How many 7970 reviews ran benchmarks testing HD3D vs 3D Vision ?* And there wont be while AMD are giving out the hardware. Hardware Heaven used to include both multi-monitor and 3D multi-monitor benching...guess which one they dropped.

Imagine how John Bonham would have sounded banging on one drum. The music sounds much better with the whole kit.

(*It's less than 1...somewhat odd since HD3D features quite prominently in the reviewers guide. See page 17)

You look hard enough you can make the numbers say anything you want. Talking of which...

The Galaxy MDT GTX 580 which, by the way is $50.00 more than the 7970

The Galaxy card is $549.99 or $539.99....the 7970 is selling for $490-500 ?...I'll take two please if international shipping is an option.

the only thing I like better is the faster rate in which drivers are update.

You could also look upon it as every driver release means fixing or adding compatibility. AMD keep to a monthy driver schedule, Nvidia do not. In fact Nvidia's latest whql (285.62) debuted on the 24th October. You could also argue that issuing a raft of beta's/hotfix's shows a manufacturers commitment to it's user base-both Nvidia and AMD usually issue a non-whql for every game release. Overall what matters is the functionality - in this respect, there isn't a lot to choose between CCC and Forceware, and the gap narrows as each utilize the features of the other.

red1776 red1776, Omnipotent Ruler of the Universe, said:

^^^ Hardly surprising.

40nm vs 28nm

1.5GB frame buffer vs 3GB frame buffer

Fermi arch = compute orientated (larger core/lower core speed), VLIW5/4/GCN = fillrate orientated

Surround/Eyefinity is valid, albeit for a minority of users at present, and a stronghold for AMD in single card setups...and which isn't natively supported by reference Nvidia cards (hence the lack of review benches -both from the niche perspective and the non-support by one vendor

I agree, thats why the 580 is not "on it's tail"

Are you getting many requests for 3-D? I am not at all. i am (in my anecdotal corner of the world) getting a lot of Eyefinity/multi work though.

If the GTX 780 comes out and bests the 7970 by 5%, this guy (and his ilk) will be the first ones to scream that it "crushes the 7970".

The Galaxy card is $549.99 or $539.99....the 7970 is selling for $490-500 ?...I'll take two please if international shipping is an option.

From [H]:

The Galaxy MDT GTX 580 is more expensive than the Radeon HD 7970 by about $50 This will make an interesting test as we see if the less expensive HD 7970 can provide a gameplay advantage over the Galaxy MDT GTX 580 here :here

I will be happy to send them to you...if i can find them. Told ya that a long time ago :p

on another subject, do you remember where that nifty multi GPU WC piece was that you showed me last year? I now need it.

[link] it was similar to this.

dividebyzero dividebyzero, trainee n00b, said:

I agree, thats why the 580 is not "on it's tail"

Are you getting many requests for 3-D? I am not at all. i am (in my anecdotal corner of the world) getting a lot of Eyefinity/multi work though.

Mostly still single monitor, although more people are heading towards 120Hz -not necessarily for 3D gaming though. Nvidia's 3D vision is too expensive, and AMD's HD3D isn't really working. Add in the headache issues and those are the reasons.

Likewise, most people here have to opportunity to slap together 2-3 TN panels or buy one IPS 2550x1440/1600. Most with that budget go with the better colour quality rather than the screenspace unless they are flight sim junkies.

I'd think if the multi-monitor uptake were all pervasive that 1. reviews would concentrate on 5040x1650 and higher, and 2. the Steam Hardware survey might reflect it's marketshare. I'm assuming that multi-monitor makes up the bulk of the 4.50% "Other" under the screen res (along with 2048x1152 and 2560x1600) , and of that 4.5%, a little over half would be made up of sub-4 mega pixel resolutions (under Multi-monitor category). 2 and 3 screens are obviously gaining some traction, but most people I know would go with a one larger screen with better colour reproduction.

If the GTX 780 comes out and bests the 7970 by 5%, this guy (and his ilk) will be the first ones to scream that it "crushes the 7970".

Most assuredly.Likewise, the red camp tend to take potshots to the same extent. Witness the mass howling over Bulldozer benching vs Intel...the fanboyism now seems non existant when it comes down to the benching the 7970 almost exclusively on Intel systems...even more amusing, the same people decrying Intel's "cheating ways" now start squealing that the 7970 should be benched exclusively using a PCIe 3.0 motherboard. You could add in forum trolls demanding nothing less than 8xMSAA, whereas I seem to recall (I can cite if req'd) those same people saying that 4xAA was more than adequate and benching anything higher was a designed solely to show the HD 5870 et al in a bad light (see also tessellation, extreme/DoF/HBAO/SSAO). The arguments tend to stay the same, they just get uttered by alternating camps depending on whom it suits.

And....

From [H]

The Galaxy MDT GTX 580 is more expensive than the Radeon HD 7970 by about $50 This will make an interesting test as we see if the less expensive HD 7970 can provide a gameplay advantage over the Galaxy MDT GTX 580 here :here

You know many people that would shop at Amazon in order to pay more than they would pay at Newegg given the choice?....and even if those people exist, why would you use an Amazon link when virtually every other build recommendation link you've ever posted here on Techspot leads to Newegg ? Sudden change in affiliation ?

Just as well we weren't referencing Amazon's HD 6990

BTW : The Amazon marketplace link is showing $565.35 as the cheapest on Amazon, which would still require that the 7970 sell at $34 under MSRP. Maybe it's true, since it's highly unlikely Kyle and others would weight their editorial to suit their own agendas.

red1776 red1776, Omnipotent Ruler of the Universe, said:

You know many people that would shop at Amazon in order to pay more than they would pay at Newegg given the choice?....and even if those people exist, why would you use an Amazon link when virtually every other build recommendation link you've ever posted here on Techspot leads to Newegg ? Sudden change in affiliation ?

You realize that was the link from Kyles article right?

[link]

Silly me, I just assumed it was because it was not availible at Newegg, look what I found

[link]

Most assuredly.Likewise, the red camp tend to take potshots to the same extent. Witness the mass howling over Bulldozer benching vs Intel...the fanboyism now seems non existant when it comes down to the benching the 7970 almost exclusively on Intel systems...even more amusing, the same people decrying Intel's "cheating ways" now start squealing that the 7970 should be benched exclusively using a PCIe 3.0 motherboard. You could add in forum trolls demanding nothing less than 8xMSAA, whereas I seem to recall (I can cite if req'd) those same people saying that 4xAA was more than adequate and benching anything higher was a designed solely to show the HD 5870 et al in a bad light (see also tessellation, extreme/DoF/HBAO/SSAO). The arguments tend to stay the same, they just get uttered by alternating camps depending on whom it suits.

and that there is the "better gaming experience" that almost anything can fall under, or be included or excluded from testing.

And....

venomblade said:

I'm holding out for a 7890, or the Nvidia equiv.

dividebyzero dividebyzero, trainee n00b, said:

on another subject, do you remember where that nifty multi GPU WC piece was that you showed me last year? I now need it.

Try this. You just leave the taps in the ports you don't use. Are your cards reference layout? If not, you might have a problem sourcing full cover blocks. Don't think I'd recommend GPU-only blocks for a 69xx card.

EDIT:

If this is due to your system going into the Cosmos 2, I'd closely check the clearance in the roof. The one gripe that came out of the July showing- and doesn't seem to have been remedied, is that there really isn't much provision for a decent radiator in the roof. The slimmer rads seem OK, but the better performers are 54mm+ deep and there seem to be some issues with rad depth + fans.

ET3D, TechSpot Paladin, said:

dividebyzero said:

I'm assuming that multi-monitor makes up the bulk of the 4.50% "Other" under the screen res (along with 2048x1152 and 2560x1600)

Since that category is defined as "Primary Display Resolution" I imagine you're wrong. It would be interesting to know what the multi-monitor percentage is, but I assume that isn't it.

Even if it was 4%, it would be higher than the percentage of users with highest end cards. The question isn't what percent of the population has multiple monitors, but what percent of users with cards like the 7970 or 580 (or SLI) have. My bet would be it's much higher than those with 3D.

ET3D, TechSpot Paladin, said:

Ouch, forgot the ending quote mark.

red1776 red1776, Omnipotent Ruler of the Universe, said:

Speaking of the Steam hardware survey earlier, the results of it are to be taken with a grain of salt if not thrown out the window. i noticed that over my last four quad fire machines, it detects them as single card "no crossfire/SLI' detected" The email back from Steams customer service basically told me that yes, there are certain problems and we are working on them...this was over a year ago. There are also a number of other hardware aspects that it 'mis-detects' as well. too bad , i would actually like to have an reliable source for numbers on multi monitor. I have a suspicion it's higher than most think. i am seeing 3 wide benches starting to crop up in testing, kit Guru, Overclockers.com, [H]ocp,Hardware Heaven and others.

Leeky Leeky said:

Steam even came up with a warning saying it didn't recognise my HD6870 either. I don't recall ever changing a GPU and not seeing it tbh. I just assumed it was some stupid bug low down on the list of more important issues they're working on.

Guest said:

Long Live AMD! they rock! Drivers and cards are great. All XBOX360's have AMD video chipsets. Gogogo AMD!

dividebyzero dividebyzero, trainee n00b, said:

Since that category is defined as "Primary Display Resolution" I imagine you're wrong. It would be interesting to know what the multi-monitor percentage is, but I assume that isn't it.

Hence the "assuming" preface to my comment

Even if it was 4%, it would be higher than the percentage of users with highest end cards.

True. CF/SLI of lower tier cards would make up a reasonable percentage. Whatever the true number is, if Steam's figures are indicative it would mean that even if you factored multi-monitor gaming as a significant percentage, over half of the display combinations combine for less pixels than a single 2560x1600 display.

More to the point, you could take the membership of a site such as Techspot (which would still produce skewed results in favour a more enthusiast setup than the average) and note how many use multi-monitor setups. (DONE. Let's see how enthusiastic the enthusiasts are)

The question isn't what percent of the population has multiple monitors, but what percent of users with cards like the 7970 or 580 (or SLI) have.

Good luck trying to find an estimated sample OF an estimated sample.

Here's Mercury Research's estimate to help you on your way. According to their figures, enthusiast AMD cards sit at ~5% of sales, and while Nvidia cards are almost double that figure, you do need two cards for multi-monitor gaming at 4+ Mpixels

[link]

My bet would be it's much higher than those with 3D.

Assuredly. But then, that wasn't the point. The point was that multi-monitor and 3-D were both niche markets, as is CUDA acceleration in CS and other apps for instance, or AMD's hit-or-miss approach to OGL based content creation apps.

++++++++++++++++++++++++++++++++++++++++

Now...

This whole line of posting is pretty much beside the point. In this thread, red posted a reply to a comment I made in another thread (see post #18) and decided to put an entirely different spin on it.

Note: like culling of the final line in my post

"Bench results seem in line with previous reviews - allowing for the reviews that used some very odd methodology or custom IQ."

The original premise was that red stated that he "[had] heard from a normally reliable source that the first round of review cards may have been throttling back a bit prematurely." (see post #3). I doubted that post-launch performance would differ appreciably due to a number of factors (see the posts in question). The Tech Report review was the first HD 7970 review post-launch, and I posted the condensed results as a comparison to the performance in launch reviews (using the GTX 580 as a baseline since it features in comparison in virtually all 7970 reviews).

So what was in all actuality, a comparison of before-and-after performance of the same HD 7970 card, has been taken and manipulated into a one-sided fanboy flamewar by red adding a multi-display comment to a quote taken out of context from another thread...and while I must admit I was interested to ascertain the rationale behind turning an AMD vs AMD post into an AMD vs Nvidia flamewar based on gaming res., it's now come down to the fanboy stronghold of attempting to prove a negative. Nice to see ET3D weigh in. I was hoping that red was going to be bolstered by Mosu as well.

ET3D, TechSpot Paladin, said:

dividebyzero said:

and while I must admit I was interested to ascertain the rationale behind turning an AMD vs AMD post into an AMD vs Nvidia flamewar based on gaming res.,

You certainly fanned the flames with your comment about 3D vs. multi monitor, not to mention discussing the flame war itself and taking pot shots at the red camp in particular. "Ascertain the rationale"... right.

dividebyzero dividebyzero, trainee n00b, said:

You certainly fanned the flames with your comment about 3D vs. multi monitor

Versus ? What I wrote was that both technologies were of a gaming minority....

3-D gaming is probably no less valid as a niche market

...please enlighten me as to where I stated that one of these technologies should be considered above the other

not to mention discussing the flame war itself and taking pot shots at the red camp in particular. "Ascertain the rationale"... right.

I believe the flame war that I was "discussing" was in reply to the supposition by another poster.....

If the GTX 780 comes out and bests the 7970 by 5%, this guy (and his ilk) will be the first ones to scream that it "crushes the 7970".

.....I answered the supposition with a statement of fact- after agreeing with the poster- a somewhat more even-handed approach I would have thought than having my original post taken out of context in order for someone to build a platform for their hobby horses.

As for taking "pot shots and fanning flames", I have no qualm about either- you won't see me losing any sleep over the process- "Fanboy Standoff In GPU Thread Causes Dispruption"...well that never happens does it? . If you'd care to re-read what I've written and use some modicum of written comprehension, you would note that my objection was to have a post from another thread edited and taken out of context to initiate the process.

red1776 red1776, Omnipotent Ruler of the Universe, said:

Which is why I bowed out of it all. All I have done is:

1) Wonder if there was any merit to the "premature throttling issue' I had heard about, and state that I would be watching with great interest the Jan 9th release for confirmation or dissolution of this.

2) Ponder multi monitor adoption rates (while prefacing it by saying my customer experience is admittedly anecdotal)

3) Note yet another fanboy who (if it was the other way around) would say that the current numbers are a crushing defeat for AMD....or triumphant victory for Nvidia.

Dividebyzero said:

So what was in all actuality, a comparison of before-and-after performance of the same HD 7970 card, has been taken and manipulated into a one-sided fanboy flamewar by red adding a multi-display comment to a quote taken out of context from another thread...and while I must admit I was interested to ascertain the rationale behind turning an AMD vs AMD post into an AMD vs Nvidia flamewar based on gaming res., it's now come down to the fanboy stronghold of attempting to prove a negative. Nice to see ET3D weigh in. I was hoping that red was going to be bolstered by Mosu as well.

A.

I noted that as the res went up, so did the performance spread of the two cards, according to benches i have seen. Given that these are top level $500 cards, 'ultra res/multi monitor can be a reasonable factor in the assessment of these parts success and purchasing motivation..ie...do people who buy these really play at 1600 x 1200? etc

and...

B.

I the noted that Steam is not a very reliable survey of system hardware BY THEIR OWN ADMISSION, and In two years , and four quad GPU systems, they have never detected more than one GPU in my system. ( not to mention it has gets my ram wrong 50% of the time as well)

ET3D said:

You certainly fanned the flames with your comment about 3D vs. multi monitor, not to mention discussing the flame war itself and taking pot shots at the red camp in particular. "Ascertain the rationale"... right

How is it a shot at the red team? it is a valid question/factor and AMD's 3D is at best of questionable functionality 'go-go-go!'

....I guess that now does not count as bowing out. oh well.

dividebyzero dividebyzero, trainee n00b, said:

Which is why I bowed out of it all. all I have done is:

1) wonder if there was any merit to the "premature throttling issue' I had heard about, and state that i would be watching with great interest the Jan 9th release for confirmation or dissolution of this.

Which was worthy of discussion in the appropriate thread. My keeping an eye on post-launch review benches was precisely to test the theory you postulated. For the crime of using the GTX 580 as baseline comparison for HD 7970 at launch and post-launch, my post was ported over to this thread to make a comparison between the two. I believe I'm on record here as stating that the 7970 is new top dog. I certainly don't believe I put forward a case for the GTX 580....in fact, I believe my only statement on the card was to note that Kyle Bennett cherry picks pricing to show whatever product he's pushing into the best possible light

2) Ponder multi monitor adoption rates (while prefacing it by saying my customer experience is admittedly anecdotal)

Pondering is all anyone will ever do. As an exact science it ranks with ascertaining accurate RMA and failure rates and reliable breakdowns of sales by SKU. My argument is that you test for the market segment the product is aimed at, and for the enthusiast sector that encompasses Multiple GPU's, resolutions at 1080p and above (including multiple displays), 3-D gaming and compute functionality. You either isolate the benches for the masses, or you include the whole gamut. Every group has it's adherents, whether it be gaming at extreme resolution or distributed computing. I'd be reluctant to say that one preference is valid whilst another is not.

3) note yet another fanboy who (if it was the other way around) would say that the current numbers are a crushing defeat for AMD....or triumphant victory for Nvidia.

...who was bleating on about drivers and some such. One day I'll post an entire page of self-serving contradictory statements penned by some of the people here just for the lulz

I noted that as the res went up, so did the performance spread of the two cards, according to benches i have seen.

A self evident truth. Performance with a new arch and process should always outperform the old. If it didn't then you're doing it wrong. I don't see anyone disputing this fact (aside from amstech, so ask yourself).

Given that these are top level $500 cards, 'ultra res/multi monitor can be a reasonable factor in the assessment of these parts success and purchasing motivation..ie...do people who buy these really play at 1600 x 1200? etc

I'm running GTX 580 SLI and HD5850 CFX. I game at either 1980x1200 or 1980x1080. If I could afford it I'd ditch my two monitors (and possibly the TV) in order to buy a 30" 2560x1600. Does that answer your question?

I the noted that Steam is not a very reliable survey of system hardware BY THEIR OWN ADMISSION, and In two years , and four quad GPU systems, they have never detected more than one GPU in my system.

Granted, the figures aren't reliable. Do you know of any published figures that are more so?

I tend to use the information at hand. I'm more than willing to use a better source if one presents itself.

How is it a shot at the red team? it is a valid question/factor and AMD's 3D is at best of questionable functionality

ET3D doesn't seem in agreement with AMD, who are actively pushing 3D gaming via HD3D -as I noted regarding the reviewers guide pdf for the HD 7970, and as seen on AMD's site. Maybe ET3D see's Nvidia as synonymous with 3-D gaming which would say more about the rudimentary nature of some peoples technical breadth and public perception than any degree of actuality. On the other hand maybe using the Bulldozer example-even if accurate- constitutes heresy

...I guess that now does not count as bowing out. oh well.

done and done

red1776 red1776, Omnipotent Ruler of the Universe, said:

Dividebyzero:

Which was worthy of discussion in the appropriate thread. My keeping an eye on post-launch review benches was precisely to test the theory you postulated. For the crime of using the GTX 580 as baseline comparison for HD 7970 at launch and post-launch, my post was ported over to this thread to make a comparison between the two. I believe I'm on record here as stating that the 7970 is new top dog. I certainly don't believe I put forward a case for the GTX 580....in fact, I believe my only statement on the card was to note that Kyle Bennett cherry picks pricing to show whatever product he's pushing into the best possible light

This is the aspect i do not understand. As far as i know, the line I "culled' was an aggregation of survey/bench statistics. If they were accurate in post #18, they are accurate here. That was aimed at amstech. I don't believe I asserted that you put forth a case for the GTX 580.

If your point is , you do not like me using/quoting your findings or putting a handle on them...done. it's not as if I changed the numbers or content of your findings.I merely took your findings at your word, and used them as an example.

Dividebyzero:

Granted, the figures aren't reliable. Do you know of any published figures that are more so?

I tend to use the information at hand. I'm more than willing to use a better source if one presents itself.

I do not, and do not use them for that reason since learning about this flaw in Steams hardware detection system.

Dividebyzero:

A self evident truth. Performance with a new arch and process should always outperform the old. If it didn't then you're doing it wrong. I don't see anyone disputing this fact (aside from amstech, so ask yourself).

Which I agreed with, twice. and who (amstech) it was aimed at.

Dividebyzero:

I'm running GTX 580 SLI and HD5850 CFX. I game at either 1980x1200 or 1980x1080. If I could afford it I'd ditch my two monitors (and possibly the TV) in order to buy a 30" 2560x1600. Does that answer your question?

if mine is anecdotal, so is yours. Perhaps a good start to a survey...hey!....

dividebyzero dividebyzero, trainee n00b, said:

This is the aspect i do not understand. As far as i know, the line I "culled' was an aggregation of survey/bench statistics. If they were accurate in post #18, they are accurate here....If your point is , you do not like me using/quoting your findings or putting a handle on them...done. it's not as if I changed the numbers or content of your findings.I merely took your findings at your word, and used them as an example.

As I'd previously noted, the bench results were a small subset of total. In the interests of clarity these are the full results.

31 Reviews (Anand, Bit-tech, ComputerBase, Guru of 3D, Hardwareluxx, Hardware Canucks, Hardware France, Hardware Heaven (2), Hardware.info, Hardware Secrets, Hexus, Hot Hardware, HT4U, iXBT, Kitguru, Neoseeker, OC3D, OCC, PCGH, PC Perspective, Rage 3D, TPU, Tech Report, TechSpot, Tom's Hardware, technic3D, tweakpc.de,Tweaktown, VR Zone, Xbit)

Site reviews not used:

Legit Reviews and HardOCP ( used OC'ed GTX 580), motherboards.org ( GTX 580 3GB used in comparison)

Benches only from highest game IQ level used.

5760x1080 resolution HD 7970 CFX and GTX 580 SLI

Game....................>Gain over GTX580 @ 1920..........2560..........5760

AvP.........................................................
+21.15%......+22.90%

Anno 1404...............................................+31.25%..
...+37.16%

Anno 2070...............................................+53.33%..
...+52.76%.....+44.44%

ArmA II.....................................................+27.1
%......+38.24%

Batman:AA...............................................-10.
9%.........-8.57%

Batman:Arkham City................................+13.97%......+19.10%....
-42.19%

Battlefield:Bad Co.2.................................+24.18%.......+36.27%

Battlefield 3..............................................+12.43%......
17.30%....+33.63%

Battle Forge.............................................+ 2.17%.........+5.62%

Brink.......................................................
.+22.58%......+22.31%

Bulletstorm...............................................+4
.85%......+48.34%.....+29.47%

CoD4:MW....................................................-
.43%.........-6.46%

CoD:MW2................................................. +1.75%.........+5.69%

CoD:MW3..................................................+9.
0%........+18.46%

Call of Juarez...........................................+11.49%....
..+18.26%

Chronicles of Riddick................................+21.08%.......+18.69%

Civilization V.............................................+34.92%.......
20.33%......+6.80%

Crysis......................................................
+32.48%........+35.41%

Crysis 2....................................................+16.34%
.......+23.00%.....+21.13%

Crysis Warhead........................................+22.43%......
+28.90%

Dawn of War II:Chaos Rising....................+4.73%.........+13.30%

Dead Island 2............................................+24.91%

Deus Ex:Human Revolution......................+27.61%.......+28.68%

DiRT2.......................................................
.. -8.36%.........+5.44%

DiRT3.......................................................
..+8.63%.......+18.44%

Dragon Age 2..............................................+5.23%.......
10.40%

F.E.A.R.....................................................
.+5.91%..........+4.08%

F1 2010.....................................................+27
17%.......+37.90%

F1 2011.......................................................+
.38%........+23.44%.....+24.50%

Far Cry Regulator demo...............................-8.45%..........-1.65%

Far Cry 2.......................................................-2.9
%.......+11.58%

Grand Theft Auto IV....................................-19.34%

Hard Reset.................................................. +8.26%.......+12.78%

H.A.W.X.....................................................
.-28.58%.......-25.18%

H.A.W.X.2...................................................
-16.11%.......-10.42%

Homefront...................................................
14.52%

Just Cause 2..............................................+25.96%......
.+33.16%

Left 4 Dead................................................+16.72%.
.....+35.10%

Left 4 Dead 2.............................................+14.50%.......
24.73%

Lost Planet 2..............................................+14.49%......
.+25.00%

Mafia II.......................................................+17
79%........+32.35%

Mass Effect 2.............................................+53.45%.......
+62.72%

Metro 2033.................................................+21.15%
.......+25.87%....+27.91%

Portal 2.......................................................+10.
3%.......+18.15%

Project C.A.R.S...........................................-18.46%...
.....+7.41%....+10.00%

Quake 4........................................................+5.
0%........+41.56%

Resident Evil 5.............................................+10.66%.......
10.33%

Serious Sam 3............................................+27.54%.......+
5.20%

Shogun 2....................................................+34.40%
.......+28.56%....+56.25%

S.T.A.L.K.E.R. Call of Pripyat.....................+22.28%.......+37.01%

S.T.A.L.K.E.R. Clear Sky..............................+3.85%.........+4.01%

Starcraft 2.....................................................+0.51%
.........-0.48%

Star Wars: The Old Republic.......................+10.50%

Stormrise...................................................
.+66.55%.......+83.01%

Street Fighter IV...........................................-22.41%

Supreme Commander 2.................................-2.13%

Swarm.......................................................
..+48.67%.......+41.94%

TESV:Skyrim................................. ...............+9.45%......+17.96%....-12.96%

The Witcher 2...............................................+23.14%.....
.+31.74%

Track Mania...................................................-7.7
%.........+8.52%

Unreal Tournament 3...................................+21.74%

WoW:Cataclysm.............................................-3
24%.........-5.47%

Average.....................................................
..+7.76%.......+15.73%....+11.17%

it's not as if I changed the numbers or content of your findings.I merely took your findings at your word, and used them as an example.

Probably more a case of culling than editing. The post reads as though it is indicative of the relative performance between two cards. It is only indicative of the performance between two cards in five games at one resolution, whereas my earlier post in the two previous threads made the overall performance clear as well as indicating that some of the testing regimes could stand some scrutiny ( Note Hardware Heaven's testing of Batman:AC...GTX 580 SLI has an uphill struggle in an apples-to-apples comparison...while Hardware Heaven decided to further burdon the Nvidia cards with PhysX set to high). Questionable benches such as the HH Batman one are included in the overall figures.

Load all comments...

Add New Comment

TechSpot Members
Login or sign up for free,
it takes about 30 seconds.
You may also...
Get complete access to the TechSpot community. Join thousands of technology enthusiasts that contribute and share knowledge in our forum. Get a private inbox, upload your own photo gallery and more.