AMD Radeon HD 7970 GHz Edition Review: Give Me Back That Crown!

[davemaster84]

Close regarding what? I think any comparisons made in here should be made thinking about gaming performance since is actually what most of the people who posted here is refering to.

 

[verbalizer]

quick breakdown:

Processor Arithmetic
Analysing...
Aggregated Score : 51.31GOPS
Result ID : AMD Phenom(tm) II X4 975 Processor (4C 4GHz, 2GHz IMC, 4x 512kB L2, 6MB L3)
Speed : 4018MHz
Capacity : 4Unit(s)
Power : 147.49W
Finished Successfully : Yes

Processor Multi-Media
Analysing...
Aggregated Score : 87.84MPix/s
Result ID : AMD Phenom(tm) II X4 975 Processor (4C 4GHz, 2GHz IMC, 4x 512kB L2, 6MB L3)
Speed : 4018MHz
Capacity : 4Unit(s)
Power : 147.49W
Finished Successfully : Yes
---------------------------------------------

Processor Arithmetic
Analysing...
Aggregated Score : 67.07GOPS
Result ID : Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz (4C 4GHz/2.67GHz 50% OC, 3.22GHz IMC, 2x 256kB L2, 8MB L3)
Speed : 4013MHz
Capacity : 4Unit(s)
Power : 95.00W
Finished Successfully : Yes

Processor Multi-Media
Analysing...
Aggregated Score : 153.32MPix/s
Result ID : Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz (4C 4GHz/2.67GHz 50% OC, 3.22GHz IMC, 2x 256kB L2, 8MB L3)
Speed : 4013MHz
Capacity : 4Unit(s)
Power : 95.00W
Finished Successfully : Yes
-------------------------------------------------

Processor Arithmetic
Analysing...
Aggregated Score : 76.61GOPS
Result ID : Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz (4C 4GHz/6GHz Turbo, 4GHz IMC, 2x 256kB L2, 6MB L3)
Speed : 4000MHz
Capacity : 4Unit(s)
Power : 95.00W
Finished Successfully : Yes

Processor Multi-Media
Analysing...
Aggregated Score : 214.32MPix/s
Result ID : Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz (4C 4GHz/6GHz Turbo, 4GHz IMC, 2x 256kB L2, 6MB L3)
Speed : 4000MHz
Capacity : 4Unit(s)
Power : 95.00W
Finished Successfully : Yes

 

[davemaster84]

That would better fit in the CPU forum's topic, what about some game benchmarks?

I get 65.5 fps in crysis 2 adrenaline benchmark, everything maxxed at 1920 x 1080 using dx 11; Tom's benchmark with the same card on a ntel Core i7-3960X (Sandy Bridge-E) system was 69.70 fps, not much of a difference hugh?

 

[verbalizer]

just can't please everybody can 'a' person... :/

 

[verbalizer]

 

[verbalizer]

just looking for the chart with the most CPU's in it...

 

[davemaster84]

Well dunno , I am aware that Intel outperforms AMD but I am just trying to say that the difference is not as large as people thinks at least in gaming (of course talking about deneb, thuban) . In my personal experience I have bf3 too, and I play at 64 player servers and it is almost 60 fps all the time, metro 2033 runs over 60 even seen 90, deus ex sucks (at least for me) so just dunno. But yeah I get your point , actually I am waiting for the 8350 fro piledriver to upgrade my system

 

[davemaster84]

[citation][nom]recon-uk[/nom]Crysis 2 is GPU dependant....BF3 multiplayer will find the real man out of the 2, 64 man server.Try out something at least a little demanding...[/citation]

Actually in the real gameplay crysis 2 is more demanding than bf3, maybe the benchmarks say it run an average of 65, but when you actually play it stays more in the 55 tange, although you may take a look at the new unreal benchmark I posted with my 955 at the threat you actually started

 

[davemaster84]

Sorry about my english, I meant thread, I posted a couple of benchmarks last night , I just checked them and I managed to have 5 frames more than mouse monkey using his same settings, and he is running a 2600k. I am not trying to be the best benchmarker or anything, I know bigmack would get much better numbers than me, as I said I am just trying to proof AMD cpus are not as bad as people are trying to put

 

[sayantan]

They should have rolled this driver much before the release of 680.

 

[ojas]

[citation][nom]DataGrave[/nom]100% acknowledgeAnd for the gamers: take a look at the new UT4 engine! Without excellent GPGPU performace this will be a disaster for each graphics card. See you, Nvidia.[/citation]
Wait, UT4 was shown off ON Nvidia hardware iirc.

 

[nebun]

lol...not good enough to take the crown just yet....maybe if they improve the code they could do it....time will tell

 

[arbiter1]

what should be issue is 84c, but is that in a case or is it open air test bench. If its open air 84c is optimistic temps. Could be 90c+ in end users machines or a lot higher fan noise. I have a gtx670 with cooler mine has even in a case highest it runs at load is 60-65c (no i didn't mistype). Even with power limit set to max 111% so it will boost clock to 1200mhz.

 

[eddieroolz]

I would call this a mixed bag. I see this as AMD's stopgap measure to stave off the GTX 680 while they prep the real contender, the HD8xxx series. As such the whole "GHz Edition" moniker seems very tacked-on and means little to enthusiasts.

Tahiti does better in compute as anyone would have expected, but in the gaming space it again comes down to what games you'd play.

 

[verbalizer]

+1

 

[_zxzxzx_]

Haven't quite got the crown yet but perhaps the HD7990 will take out the dual-gpu crown?

 

[youssef 2010]

So, AMD has done it once again. This company is incredible. The idea to release the GHz edition might not win everybody. This edition will be particularly valuable to people who want to get the extra speed without the loss of warranty. Also, I've had two modern Radeons now; the 4870 and the 6950. Since catalyst release 9.2, I've been using manual fan control to balance temperatures and acoustics. I can't imagine why anybody would tolerate the default fan map when you can run the fan @ 45-50% duty cycle and have great thermals and tolerable acoustics (and that's in a 35C ambient temperature).

AMD's driver team is the real hero today. They should take the weekend off and take a significant bonus for saving AMD's current generation of cards.

 

[verbalizer]

Comparing your benchmark to another persons is very difficult to do. There are a variety of settings, variables and hardware combinations/optimization that could effect the outcome. I am telling you that under a very controlled scenario (all of the benchmarks I run are exact except BF3 MP which is obviously impossible - you just need a large sample size), with as close to identical secondary hardware as possible (obviously different motherboards), an i5 is a lot faster than a PII in gaming. I ran my 960t as high as it could go stable, 4.2ghz, and it did not even compete with the i5 750 @ 3.6ghz.

+1 to that and I have been saying the exact same thing all along..

 

[davemaster84]

Yeah mate that's true, but that opens the old debate about the "perfect benchmarks scenario" , it's well known that some people know how to "cheat" while benchmarking, so actually I'm talking about real life scenarios . I bought my hardware, out it together (hardly ) and I doubt I've something special over those I mentioned, in practical situations the differrence is not as huge , trust me. All I can do is shor you my benchmarks from a common system config, there's nothing weird or super expensive on mine.

 

[verbalizer]

I'm just getting at the fact that you can take a Deneb C3 and clock it above 4.2GHz and it still gets housed by an i5-2400.
and the difference between that example is over 1.1GHz difference at least.
what are the factors for this, doesn't matter IPC, latency - whatever.
that is an entirely different thread in itself.

bottom line is that's sad.
I also have these chips mention in-house and see them daily.

 

[JackNaylorPE]

[citation][nom]blazorthon[/nom] The 580 also has less VRAM and inferior scaling, so for multi-GPU setups, the 6970 wins substantially. [/citation]

Huh ?

Guru3D uses the following games in their test suite, COD-MW, Bad Company 2, Dirt 2, Far Cry 2, Metro 2033, Dawn of Discovery, Crysis Warhead. Total fps (summing fps in each game @ 1920 x 1200) for the various options in parenthesis (single card / SL or CF) are tabulated below along with their cost in dollars per frame single card - CF or SLI:

6970 (526/825) 526 825 (57% scaling)
560 Ti - 900 Mhz (495/862) 495 862 (74% scaling)
570 (524/873) 524 873 (67% scaling)
580 (616/953) 616 953 (55% scaling)

The 6970's 825 fps in CF got beat by the 5580's 953
The 6970's 825 fps got beat by the 570's 873
The 6970's 825 gets beat by the 560 Ti (900Mhz) 862

 

[verbalizer]

Huh ?

Guru3D uses the following games in their test suite, COD-MW, Bad Company 2, Dirt 2, Far Cry 2, Metro 2033, Dawn of Discovery, Crysis Warhead. Total fps (summing fps in each game @ 1920 x 1200) for the various options in parenthesis (single card / SL or CF) are tabulated below along with their cost in dollars per frame single card - CF or SLI:

6970 (526/825) 526 825 (57% scaling)
560 Ti - 900 Mhz (495/862) 495 862 (74% scaling)
570 (524/873) 524 873 (67% scaling)
580 (616/953) 616 953 (55% scaling)

The 6970's 825 fps in CF got beat by the 5580's 953
The 6970's 825 fps got beat by the 570's 873
The 6970's 825 gets beat by the 560 Ti (900Mhz) 862

I know jack I know.
he'll say one logical thing and then he blows it with something more dumber
than what he said previously was smart..
:/

 

[youssef 2010]

The 7950 ICQ Turbo IS available on Newegg:

http://www.newegg.com/Product/Product.aspx?Item=N82E16814161415

Now, It might not have been available at the time of writing. But I think you should update the article to avoid providing outdated info.

DHE FTW.

 

[Mousemonkey]

Sorry about my english, I meant thread, I posted a couple of benchmarks last night , I just checked them and I managed to have 5 frames more than mouse monkey using his same settings, and he is running a 2600k. I am not trying to be the best benchmarker or anything, I know bigmack would get much better numbers than me, as I said I am just trying to proof AMD cpus are not as bad as people are trying to put

Which set of settings? I have posted the results of several settings.

 

[JackNaylorPE]

I know jack I know.
he'll say one logical thing and then he blows it with something more dumber
than what he said previously was smart..
:/

You do have a hard time arguing with published data ..... calling what someone quoted dumb doesn't change what the reviewer published.