AMD Radeon HD 7970 3 GB

[Brod]

I believe there is an error with the 6870 and 5850 performance per watt figures. As you can see, their PPW relative to the 6850 differs wildly from Techpowerup's 6870 review.

 

[THE_EGG]

Wot, a 10% increase in performance over the GTX 580 justifies it being 10% more expensive than the GTX 580 card that was released ages ago? Nvidia produces better quality cards with much better quality drivers and support. In contrast, ATI has always offered better value for money, but we aint seeing that here.

Hear Hear, totally agree 580 killer my ass.

 

[magibeg]

Hear Hear, totally agree 580 killer my ass.

Well it all depends on how you look at it. If i was going to buy a new rig right now (i'm waiting until ivy bridge) I looks like it would be a no brainer to get the 7970 over the GTX580. At 1920x1200 res the 7970 maintains an easy 20+% advantage (according to anad, hardware canucks, etc etc). Uses less power too.

By all means the resolutions that gamers in the market for $500 cards in has the 7970 in quite the lead.

 

[pantherx12]

Well it all depends on how you look at it. If i was going to buy a new rig right now (i'm waiting until ivy bridge) I looks like it would be a no brainer to get the 7970 over the GTX580. At 1920x1200 res the 7970 maintains an easy 20+% advantage (according to anad, hardware canucks, etc etc). Uses less power too.

By all means the resolutions that gamers in the market for $500 cards in has the 7970 in quite the lead.

The reason it reads as 10% on here is because Wizz averages out the scores for the performance summary page, so one bench mark with a low score can bring down the entire thing.

I posted this on the last page but no one seems to of noticed

VS 6970
1920x1080
AVP : +26.26
Batman AA :+ 62.55
BFBC3: +51.57
Battleforge:+ 32.57
COD4:+11.81
CIV5:+43.73
Crysis:+33.51
Crysis2:+56.26
Dirt3:+15.64
Dragonage2:+19.14
Hard reset :+34.14
Metro:+27.23
Stalker cop:+40.07
Starcraft 2 :0.93
Shogun 2 :+39.51
Skyrim:+5.95
Wow: - 6.5 that's minus 6.5 so to be absolutely clear
3dmark 11:+47.43
Heaven:+97.81 (wooo improved tessellation)



In games with decent engines it maintains 20%+ and then some.

 

[THE_EGG]

The reason it reads as 10% on here is because Wizz averages out the scores for the performance summary page, so one bench mark with a low score can bring down the entire thing.

I posted this on the last page but no one seems to of noticed

VS 6970
1920x1080
AVP : +26.26
Batman AA :+ 62.55
BFBC3: +51.57
Battleforge:+ 32.57
COD4:+11.81
CIV5:+43.73
Crysis:+33.51
Crysis2:+56.26
Dirt3:+15.64
Dragonage2:+19.14
Hard reset :+34.14
Metro:+27.23
Stalker cop:+40.07
Starcraft 2 :0.93
Shogun 2 :+39.51
Skyrim:+5.95
Wow: - 6.5 that's minus 6.5 so to be absolutely clear
3dmark 11:+47.43
Heaven:+97.81 (wooo improved tessellation)



In games with decent engines it maintains 20%+ and then some.

hey can you do a comparison like this except vs. gtx 580 ? maybe the 3GB model as well ?

 

[THE_EGG]

Well it all depends on how you look at it. If i was going to buy a new rig right now (i'm waiting until ivy bridge) I looks like it would be a no brainer to get the 7970 over the GTX580. At 1920x1200 res the 7970 maintains an easy 20+% advantage (according to anad, hardware canucks, etc etc). Uses less power too.

By all means the resolutions that gamers in the market for $500 cards in has the 7970 in quite the lead.

I do agree it is better but still the 580 is over a year old now and I think a direct comparison review is needed with Nvidia's newest (maybe WHQL?) drivers. Along with the 3GB version of the card. TBH though I think it would have been a GTX 580 killer if it came out at the start of the year. Also from my experience Nvidia have always had more stable drivers etc.

 

[pantherx12]

hey can you do a comparison like this except vs. gtx 580 ? maybe the 3GB model as well ?

I can do vs the regular model of 580*, give me 10 minutes or so.

Have to manually input the data


*At 1920x1080 the extra ram shouldn't really come into play anyway, if it does difference won't be game changing.

 

[Vipeax]

i rebenched the hd 5970 and uploaded new summary graphs with the corrected results

Thank you W1zzard. Highly appreciated that you toke the time to fix it.

 

[magibeg]

I do agree it is better but still the 580 is over a year old now and I think a direct comparison review is needed with Nvidia's newest (maybe WHQL?) drivers. Along with the 3GB version of the card. TBH though I think it would have been a GTX 580 killer if it came out at the start of the year. Also from my experience Nvidia have always had more stable drivers etc.

http://www.hardwarecanucks.com/foru...s/49646-amd-radeon-hd-7970-3gb-review-13.html

I dunno about the drivers, but he 3GB model doesn't appear to gain all that much.

From those numbers it looks basically like the regular 580 loses by 24% and the 3GB loses by about 20%.

 

[dom99]

hmmm I must admit I was expecting a bit more from this chip, plus for the price they are asking there is no way I am trading in my 6950. If they priced it at £300 I would be sorely tempted but £450 is ridiculus. Once Nvidia release their 28nm cards im sure we will see a huge price drop. It must be AMD trying to make up for the Bulldozer fail and claw back more profits

 

[pantherx12]

1920x1080
7970 vs 580
Percentage increase or decrease
AVP : +25.40
Batman AA :+8.71
BFBC3: +19.69
Battleforge:+4.61
COD4:+3.46
CIV5:-7.3
Crysis:+23.64
Crysis2:+20.57
Dirt3:-7.53
Dragonage2:+12.44
Hard reset :+3.06
Metro:+10.03
Stalker cop:+17.82
Starcraft 2 :+2.84
Shogun 2 :+47.36
Skyrim:-1.4
Wow:+2.5
3dmark 11:+32.18
Heaven:+27.76

Now we have a fair comparison, highlights any weaknesses and strengths of the new card : ]

 

[Winston_008]

Well, based on actual capacity of my 2TB HDD, compared to previous 360GB one, it should cost me $850. So at $85 it was a steal, "which is part of the reason prices for it never stabilized at that low a level".

You really should use the quote button when you try to mock what someone has said, people just see you as stupid otherwise.

Great review btw, i actually read the entire thing which is unusual for me. Top job

 

[Aceman.au]

So it's going on shelves on January 3rd?

 

[mastrdrver]

Anyone know if the knew eyefinity features a 7000 series only?

Or if it's a software thing.

( talking specifically about independent resolutions)

As if it's only 7000 series only, I MUST have one lol.

Will be able to have a 2.39:1 aspect set up

It is software from what I've read. Suppose to come to all 5xxx and 6xxx cards when the 12.2 Catalyst driver shows up.

 

[khanman125]

I am really surprised how well the 5970 has held up over the past 2 years, but i think its time for me to move to a single GPU config finally. But, a 7990 sounds pretty good to me. It seem like the scaling issues from the 5xxx to the 6xxx have been improved quite a bit. I hope the 7xxx follows suit.

 

[pantherx12]

It is software from what I've read. Suppose to come to all 5xxx and 6xxx cards when the 12.2 Catalyst driver shows up.

Thanks man, now I just need a hdmi/display-port to vga and to find two el cheapo 16:10 monitors

 

[mastrdrver]

That's what I read over on Anand. I have not seen anyone else comment on it so ymmv.

Source

So what is Eyefinity Technology 2.0 composed of? We’ve already seen several new features starting with Catalyst 11.10, such as 5x1 portrait and landscape support and flexible bezel compensation support. The next step is going to be integrating Stereo 3D (or as AMD likes to call it, HD3D) into the mix, similar to how NVIDIA has 3D Vision Surround. Catalyst 11.12 introduced the ability to use HD3D with an Eyefinity display setup, and Catalyst 12.1 (preview out now) added support to do that in a CrossFire configuration. The final step is going to be with Catalyst 12.2 in February, which will add support for custom resolutions and the ability to relocate the Windows task bar to an arbitrary screen, two features that users have been asking about for quite some time. Again, all of these improvements are driver side, but they are a major component of AMD’s marketing for Southern Islands.

 

[H82LUZ73]

Review your facts again


6870->6970 18%
5870->6870 -7%
4890->5870 26%
3870->4870 46%
2900XT->3870 -2%
Overall performance summaries

As I don't see anything that exceeded 50% let alone came close to 100%.

You Have this wrong my friend ,The 5870 was replaced with the 6970 not the 6870 replace 6970....Barts and Cayman came out last year,The 6870 was the first to hit but was a mid range card,The Cayman replaced the Cypress IE 6970-5870.So too see the % between the 5870-6970 you need to redo them.

 

[buggalugs]

The release price usually only lasts a couple of months anyway, when all the stores have stock the price will start to drop, then the cheaper non reference models start to appear.

 

[cdawall]

You Have this wrong my friend ,The 5870 was replaced with the 6970 not the 6870 replace 6970....Barts and Cayman came out last year,The 6870 was the first to hit but was a mid range card,The Cayman replaced the Cypress IE 6970-5870.So too see the % between the 5870-6970 you need to redo them.

Your like 5 pages late that's been gone over.

 

[Kaotik]

I believe there is an error with the 6870 and 5850 performance per watt figures. As you can see, their PPW relative to the 6850 differs wildly from Techpowerup's 6870 review.

It's not only that, 6870 was ~20% faster than 6850 earlier, now it says ~6% - at "all resolutions" it's exactly the same speed, and at 1024x768 it's actually ranked slower than 6850.

It leaves a bitter taste of the whole review, TPU has been "the place" for me to check the general speed of cards, but this incident makes one think, is rest of the graphs crooked too (as they're quite a bit different from previous ones, though that could be explained by new game choices - 6870 performance however can't be)

I just hope W1zzard will re-bench things when 7950 is released, if he's not going to do it now

 

[DannibusX]

I just hope W1zzard will re-bench things when 7950 is released, if he's not going to do it now

On the third page, you'll find the follwing:

not sure about those HD 5970 numbers, maybe CF didnt get enabled. I'll rebench.

no, i rebenched all cards with the latest drivers, check the test setup page

 

[Kaotik]

On the third page, you'll find the follwing:

Yes, and that still doesn't change the fact that at least 6870 results are still wrong.

HD5850 and 6870 are both VLIW5 parts, there's just explainable way that HD6870 goes from being ~7% or so faster than 5850 to 5850 beating it by nearly 4% at same resolution.

Even if you could pull some tricks from somewhere to explain why that happens, there is absolutely no possibility that HD6850 would outperform HD6870 in any circumstances, like it is doing now (1024x768 chart), or that it would be as fast (all resolutions combined chart).

6850 is weaker than 6870 on every single theoretical and practical meter, it uses the very same chip with parts disabled and has lower clocks too - 6850 can not outperform 6870, ever, yet in this review it does under certain resolution, with a difference (over 3%) which can't be even called error margin by a long shot.

 

[YautjaLord]

Whether this is the 1st time you use it Wiz, or not - thank you for running Crysis 2 benchie in DX11 & not 9; thanx for conductive benchmarking as usual. One more question remains: which level or timedemo you used? Times Square one or do you have access to built-in timedemo via console/cvar? Once got this game i wanna run it too.

 

[the54thvoid]

Yes, and that still doesn't change the fact that at least 6870 results are still wrong.

HD5850 and 6870 are both VLIW5 parts, there's just explainable way that HD6870 goes from being ~7% or so faster than 5850 to 5850 beating it by nearly 4% at same resolution.

Even if you could pull some tricks from somewhere to explain why that happens, there is absolutely no possibility that HD6850 would outperform HD6870 in any circumstances, like it is doing now (1024x768 chart), or that it would be as fast (all resolutions combined chart).

6850 is weaker than 6870 on every single theoretical and practical meter, it uses the very same chip with parts disabled and has lower clocks too - 6850 can not outperform 6870, ever, yet in this review it does under certain resolution, with a difference (over 3%) which can't be even called error margin by a long shot.

Why are you getting so upset?

How the 6870 and 6850 score relative to 7970.
All Res : Equal %
1024x768: 6870 2% lower (the 'wonky' result')
1280x1024 : 6870 2% higher
1680x1050 : 6870 3% higher
1920x1200 : 6870 3% higher
2560x1600 : 6870 2% higher

In one result it is 2% worse than the 6870 when cmpared to the 7970. It could have been a minor glitch and you want him to rebench? It's a review for the highest end single gpu. The low-medium mainstream cards aren't that relevant to the review.

Go here to feel better. http://www.techpowerup.com/reviews/HIS/Radeon_HD_6870/29.html. This is the 6870 original review. It is far more relevant if you want to compare the 6870 and 6850. This review is for the 7970.