Information

Dear forum members,
We are delighted to inform you that our forums are back online. Every single topic and post are now on their places and everything works just as before, only better. Welcome back!



Discussion

Discussion on Article:
EVGA GeForce GTX 680 Superclocked Signature 2 Graphics Card Review

Started by: er_wendigo | Date 09/04/12 01:35:45 PM
Comments: 15 | Last Comment:  09/09/12 10:34:36 AM

Expand all threads | Collapse all threads

[1-1]

1. 
boys, you must check your collection of games for the test, 5 of 8 are "gaming evolved" titles, and 4 of them have a large byass to AMD.

Sleeping dogs and sniper elite, both of them, have techniques like supersampling+other antialiasing (I don´t know what, itsn´t specified) for make a clear advantage to the cards with the best memory subsystem (bandwidth).

Oh, and some "little" compute shaders in the equation.

Dirt Showdown uses some heavy techniques of compute shaders too.

"DiRT Showdown – AMD Benchmark Guide":

http://blogs.amd.com/play...wn-amd-benchmark-guide/2/

And the last one, Alien vs Predator, another game with heavy byass for "the gaming evolved" way of AMD life:

From de AMD blog:

"One of the key new features of DirectX 11 that Rebellion has incorporated into their new game is support for Direct Compute. Direct Compute lets developers harness the massive parallel processing power of modern GPUs to speed up advanced post processing effects like Depth of Field and tone mapping."

"That, combined with the more realistic shadows created with DirectX 11’s High-Definition Ambient Occlusion"

And in all of them, a very little of tessellation.

Boys, are you biased?

If you want to look as a neutral site, you can´t make a testbench with this heavy byass to AMD graphic cards. 5 of 8, 4 heavy adepts to the "guides of AMD" about what is DX11.
4 4 [Posted by: er_wendigo  | Date: 09/04/12 01:35:45 PM]
Reply
- collapse thread

 
It is not necessary to do unfounded statements. We don't arrange the tests list under someone.

Offer games for tests, and, the main thing, a testing technique in them without FRAPS use.
3 0 [Posted by: Jordan  | Date: 09/04/12 11:31:23 PM]
Reply
 
It isn´t unfounded statements, are facts:

5 of 8 games of your list are "gaming evolved" games.

4 of these games shows a huge byass to AMD graphic cards, via the 3D techniques promoted by AMD.

The use of techniques like supersampling AA, without any alternative like the traditional multisampling (sniper, sleeping dogs), are part of this byass in this games. This games combines supersampling with postprocessing "AA", very convenient for AMD. It´s another fact.

From the games that are TWIMTBP partners (3), one has a moderate byass (crysis 2 and its pointless and excesive tessellation), but the other games doesn´t have this clear byass (metro 2033 is a TWIMTBP game but with a clear advantage with modern AMD graphic cards, and Battlefield 3 don´t use any 3D technique in a suspicious way).

Dirt Showdown is a game that uses the same game engine that Dirt 3 (with new techniques implemented via gaming evolved, of course), but shows a very bad performance with nvidia cards, and so for it is a very suspicious choice.

Your list of games includes TWO games of the same developer and, this developer has a strong relation with AMD and its "gaming evolved". The developer is Rebellion. And the game engine is the same, Asura.

Why do you use two games with the same engine (but different version), and a engine that isn´t in the engine market like the Crytek engine, the unreal engine, etc? A intenal game engine for the use of Rebellion, what a very useful and redundant information!! ¬¬


It´s a bad choice your battery of games, is biased and this isn´t a statement, is a sustained fact.

One solution for this problem is the use of a list of games with more titles, and with a balance between TWIMTBP and gaming evolved titles. Or you could avoid the games that have a huge byass to one side (dirt showdown, as example, has a very HUGE byass, that it´s more evident if you look the performance of its predecesor, Dirt 3, with nvidia and AMD cards).
3 4 [Posted by: er_wendigo  | Date: 09/05/12 03:31:20 AM]
Reply
 
It not facts, it only your personal preferences.

You don't take that fact that is only reduced list of our test games in attention, and full look in other articles.

Once again I will repeat: in tests we are always impartial and we do not give preferences on selection of games to one or other producer.
In this article only fresh and most hard benchmarks are left, and that entered into them – not your business.

Secondly, you only are ready to criticize, and to offer games and, the main thing. Testing techniques in them as we see, aren't capable.
4 0 [Posted by: Jordan  | Date: 09/05/12 03:43:01 AM]
Reply
 
No, they are facts, not preferences:

5 of 8 games are "gaming evolved". This is a fact, not a preference.

And you can´t deny the large byass to AMD cards of some of these titles.

If you look to Dirt 3 and see the performance of the game, and then you look to Dirt Showdown... well, if you says that is a normal election, then you must be blind.

And yes, I´m ready to criticize your work when it looks like biased.

I´m a reader of your articles (a ooold one), not a editor.

Your tests maybe are created with the idea of the impartiality, but your battery of games isn´t impartial.

Other Games? Of couse:

The old and good Crysis or Crysis:WH (well, you have 2 rebellion titles with the same and internal game engine, a reference title like these two have more sense that 2 rebellion games, and yes, they are heavy gpu-games yet).

You can choose the best videogames of 2011-2012 too, like:

Batman Arkham City (the very best of 2011). It doesn´t need FRAPs.

mm... Skyrim?? the game that rockets the industry? yes, it doesn´t have a integrated bench, but... IT´s the GAME.

Civ5 or Anno 2070 (another strategy game in the list).

The witcher 2 (a game that likes AMD cards, but hell, it´s a game with little or nothing of byass), portal 2, dragon age 2, etc.

Hell, you can make a simple search of the top games of 2011-2012 and make a list of games searching a equilibrium.

With your 5 gaming evolved games, you could add to the lists some games with heavy byass to nvidia, like Lost Planet 2 or HAWX2.

They aren´t worst choice that Dirt Showdown (the name.. THE NAME speaks high!).

It´s posible make a better list of games, but it´s your work.

Darksiders 2, Prototype 2, Max Payne 3, etc.


Well, a best explanation of your "fault" is here:

http://www.xbitlabs.com/a...radeon-hd-6950-toxic.html

Before, you made reviews with 14 games and 3 sinthetic tests, and now you make reviews with 8 games and one sinthetic test.

Before you chose a better balance in the list games (some TWINTBP, some gaming evolved, many games, etc). Now you choose a worse balance (less games and with more titles with heavy byass to a side of the market.

Your problem, because to the eyes of many readers it´s a problem, is the reduced game list with a preference (5/8) with tittles with a seal of a "developing program" of one part of the market.

You made better reviews, come on, you could return to make them.

I do appreciate your taste for good analysis and its good explanations in there. But I dislike your new battery of tests, itsn´t a neutral battery.
3 4 [Posted by: er_wendigo  | Date: 09/05/12 11:06:47 AM]
Reply
 
Dirt Showdown, Sniper Elite V2 and Sleeping Dogs are latest titles. I agree with you that AvP is no longer relevant but just because Kepler is so slow at DirectCompute lighting, doesn't mean gamers don't want to see those results. This was the same situation when GTX470/480 outperformed 5850/5870 cards in games with tessellation. It was helpful for us gamers to know that GTX470/480 were better for tessellation.

Now, we also want to know that Kepler cannot provide good performance in modern titles that use DirectCompute shaders. That's useful for people who play other games besides BF3 and WOW.

What you need to remember the reviewers jobs is to show how GPUs compare in latest and most demanding titles and this review did exactly that (excluding AvP).

Actually, it's about as favourable as it gets towards GTX680 by the inclusion of Battlefield 3, Shogun 2 and Crysis 2.

You seem to be VERY vocal about lack of other games. Here I compiled reviews that have all of these games. Even if you add even more games, GTX680 still loses:

- Brink
- Batman AC
- Crysis 1 / Warhead
- Anno 2070
- Skyrim (including ENB Mods)
- Arma II / Day-Z / Operations
- Risen 2 Dark Waters
- Bulletstorm
- Mafia 2
- Deus Ex Human Revolution
- Serious Sam 3
- Darksiders 2
- Guild Wars 2
- STALKER: COP
- Alan Wake series
- Witcher 2
- Trine 2
- Civilization V

^ In all of these, the 680 would still lose. The only games left where 680 would have an advantage are old (Lost Planet 2, HAWX 2), not very good (Dragon Age II) or among the good games: Max Payne 3, Hard Reset, WoW, Just Cause 2 and Project Cars (that last one is not an official game you can buy in retail).

If you want reviews with way more games, head over to these websites and you'll get exact same results Xbitlabs got:

TechPowerup tested 18 games, 7970 > 670, 7970 GE > 680:
http://www.techpowerup.co...D_7970_Toxic_6_GB/28.html

Computerbase tested 15 games, 7970 > 670, 7970 GE > 680:
http://www.computerbase.d...on-hd-7950-mit-925-mhz/3/

H4Tu tested 17 games, 7970 > 670, 7970 GE > 680:
http://ht4u.net/reviews/2..._sc_3_gb_test/index39.php

TechReport tested 6 games, 7970 > 670, 7970 = 670 Amp!, 7970 GE > 680 AMP!
http://techreport.com/rev...on-hd-7970-ghz-edition/11

Sweclockers tested 8 games, 7970 > 670, HD7970 GE > 680:
http://www.sweclockers.co...hertz-edition/17#pagehead

etc. This is what 90% of reputable reviews show.

Sorry, but NV lost this generation on price/performance and absolute performance for singe-GPUs. Until they lower prices on GTX670/680, they are not competitive right now.

....

Every reputable professional review on the Internet has shown the exact same thing Xbitlabs did:

HD7950 800mhz < 660Ti after-market
HD7950 925-950mhz > GTX660Ti after-market < GTX670
GTX660Ti 1300mhz < HD7970 925mhz = GTX670 (depending on the source 7970 > 670)
HD7970 1Ghz = GTX680 > after-market GTX670
HD7970 GE 1.05Ghz > GTX680
HD7950 1165-1175mhz = 1250-1300mhz GTX670 = 1100mhz HD7970 > afer-market GTX680
HD7970 1150-1175mhz > any GTX670 OC
HD7970 1250mhz > 1290-1300mhz GTX680

max vold-mod on air
HD7970 1280-1290mhz < 1380-1390mhz GTX680 (only 2 cards can do this = $550 MSI GTX680 Lightning or EVGA GTX680 Classified for $720 with EVBot

max vold-mod on water
HD7970 OC > GTX680 OC on water

At stock speeds, HD7970 = 670, 1Ghz HD7970 = 680 > after-market 670s, HD7970 GE > 680.

Right now, it's not even a contest since you can buy an 1100mhz Gigabyte Windforce 3x HD7970 GE for $470:
http://www.newegg.com/Pro...aspx?Item=N82E16814125439

Sorry, but Xbitlabs is not biased. It seems you are still stuck reading March 2012 reviews with outdated drivers or are the biased one since you won't accept that AMD has the faster single-GPU card this generation.

BTW, Xbitlabs already tested HD7970 GE vs. GTX680 in 13 games and 680 lost:
http://www.xbitlabs.com/a...-ghz-edition_9.html#sect0

When nearly every professional review disagrees with you no matter how many games they use, it's time to re-assess the accuracy of your own information before calling Xbitlabs biased.

It sounds to me you are the one who doesn't have correct information about current performance of GPUs.
6 6 [Posted by: BestJinjo  | Date: 09/05/12 12:56:47 PM]
Reply
 
" I agree with you that AvP is no longer relevant but just because Kepler is so slow at DirectCompute lighting, doesn't mean gamers don't want to see those results. This was the same situation when GTX470/480 outperformed 5850/5870 cards in games with tessellation. It was helpful for us gamers to know that GTX470/480 were better for tessellation."

Well, in the very reduced list of "gaming evolved" games, and a someone more, that uses Xbitlabs NOW, please:

Can you tell me how many of them use heavy tessellation?

NO one. And no, Crysis2 doesn´t use HEAVY tessellation... ¬¬

And, Can you tell me how many of them use massive computing shaders for "illumination"?

Al least four of them (metro2033, avp, sleeping dogs and sniper elitev2).

"When nearly every professional review disagrees with you no matter how many games they use, it's time to re-assess the accuracy of your own information before calling Xbitlabs biased."

What? Man, speak for you, not for other people. You make a blatant and tiresome defence of "the AMD cards are superior", but it´s your opinion, not facts. I´m reclaiming a more HONEST list of games, not a very picked one.

What´s more, you say:

"- Brink
- Batman AC
- Crysis 1 / Warhead
- Anno 2070
- Skyrim (including ENB Mods)
- Arma II / Day-Z / Operations
- Risen 2 Dark Waters
- Bulletstorm
- Mafia 2
- Deus Ex Human Revolution
- Serious Sam 3
- Darksiders 2
- Guild Wars 2
- STALKER: COP
- Alan Wake series
- Witcher 2
- Trine 2
- Civilization V

^ In all of these, the 680 would still lose. "

LoL... XD You don´t know whar are you saying.

You make a picked list with selected titles (came on, you don´t include Crysis 2 ¬¬, as a fast example of your byass), and conditions (with mods when the game doesn´t favor to AMD cards), And yet, you are lying.

I tested a GTX 670 (max boost 1215) and a 7950 (1200 MHz), and in some of your "titles" the performance is better in the nvidia card.

Crysis Warhead (a little), Stalker CoP (very much), civ V and others.

you, boy, are the one who doesn´t have any idea of this serious problem of byass that Xbitlabs has. Because you want to have "faith" that this tests are correct, because favors your favorite brand.

The best proof of this are your words:

"The only games left where 680 would have an advantage are old (Lost Planet 2, HAWX 2), not very good (Dragon Age II) or among the good games: Max Payne 3, Hard Reset, WoW, Just Cause 2 and Project Cars (that last one is not an official game you can buy in retail). "

Man, you don´t have any gaming taste. You said that the crap Hard Reset game is "good" (itsn´t bad, it´s worse)... and Max Payne (not bad at all, but this isn´t the same that a good game).

And man, the "old games" that you said, they use more advanced game engines that the "good ones" of your choice (more of them are crappy DX9 ports/poor game engine).
2 4 [Posted by: er_wendigo  | Date: 09/08/12 01:25:02 PM]
Reply
 
You can't read can you? I linked you 5-6 reviews where all the games I listed HD7970/7970 GE wins against GTX680.

I said Dragon Age 2 is crap, I put hard reset as a good game. PLEASE re-read my post.

Finally, a reviewers job is to test many different games. I listed from 5-6 reviews the ONLY games where 680 wins.

If you don't agree with professional reviews all over the internet, I can't help you.

I don't even know what you point is. In BF3 and Crysis 2, GTX680 won by less than 10% at 1080P and lost at 1600P.

See I am a gamer so I compare benchmarks across 20-30 games. You OTOH, only look at specific games where 680 wins and claim them to be the best/good games LOL!
4 5 [Posted by: BestJinjo  | Date: 09/09/12 10:30:03 AM]
Reply
 
Jordan just ignore him. He obviously has no concept that a good reviewer should show us gamers the games in which AMD and NV both perform well and perform poorly. What if I want to play Sleeping Dogs, Dirt Showdown, Sniper Elite and Metro 2033? Sorry to say, I am going with the AMD card. That's the whole point of these types of reviews: to show us which cards run faster at what games.

In the end the best advice is to buy the GPU for the games you personally play. The reviewer's job is to test the games and comment on the performance and prices of current GPUs. Xbitlabs did an excellent job with this review since it looked at more recent games, while older reviews already looked at 12-13 games, and the conclusion doesn't change: HD7970 GE is the faster and cheaper card overall.

Things in the marketplace change all the time. From March to June, GTX670/680 were the better buys but after Catalyst 12.7 and the launch of newer titles, AMD has regained the lead in Dirt 3, Skyrim, Batman AC, and completely closed the gap in BF3 and Crysis 2, leaving NV with almost no games where they are leading.

Of course NV is not standing still and I expect a very fast videocard from them with GTX780
5 6 [Posted by: BestJinjo  | Date: 09/05/12 04:32:45 PM]
Reply
 
Thanks for understanding!
5 1 [Posted by: Jordan  | Date: 09/05/12 08:31:38 PM]
Reply
 
show the post
2 5 [Posted by: er_wendigo  | Date: 09/08/12 12:48:27 PM]
Reply
 
You distort the facts, interpret them absolutely incorrectly and don't understand events essence.
Time to you only your point of view is important, it is not necessary to make comments on articles, can talk to itself.
3 0 [Posted by: Jordan  | Date: 09/09/12 01:30:56 AM]
Reply
 
er_windigo, you are not coming off as an objective gamer. Us gamers play a variety of games, and we can't change the fact that AMD or NV worked closely with the developer to optimize it. This is the WHOLE POINT of testing both AMD Gaming Evolved and TWIMTPB titles. Either way, I listed already the handful of games where NV wins. It's very limited.

There may be 5-6 new games where GTX680 will beat a 7970 in the Fall 2012 - Winter 2013 lineup and then what you are going to claim that Jordan suddenly turned NV-biased if he includes 6 games where 680 wins? LOL!

You also forgot to mention that NV has lost in all the major games that launched recently:

- Guild Wars 2
- Diablo 3
- Darksiders II
- I Am Alive
- Dirt Showdown
- Sleeping Dogs
- Sniper Elite V2

If Jordan included even more recent titles, GTX680 would have performed even worse.

Dirt 3 does not favor NV. HD7970/GE beats 680 in that game too, and in SKyrim and in Batman AC. 680 barely pulls a win in BF3 and Crysis 2 and that only happens at 1080P.

Go read some reviews if you don't believe me! All of your information is outdated.
http://www.computerbase.d...deon-hd-7950-mit-925-mhz/

or

http://www.techpowerup.co...e/HD_7950_Vapor-X/28.html

Sorry, it's pretty clear HD7970 > 670, HD7970 GE > 680.
You keep crying about the choice of games by HAWX 2 and Lost Planet 2 are games from last generation.

Once Metro Last Light, Far Cry 3, Borderlands 2, Medal of Honor Warfighter, Dishonored, Bioshock Infinite, Tomb Raider launch, I am sure Xbitlabs will test as many as it can and not care if they are TWIMTBP games or AMD Gaming Evolved. It seems you are only interested in seeing 680 winning not what us gamers care about: What performance do 680/7970 deliver in various games.

Also, you argument is akin to a child crying. When I buy a videocard to play games, I don't care if a game is AMD Gaming Evolved or TWIMTPB title. I just want good performance and I already showed you 5-6 other reviews where over 20-30 games, HD7970 GE was still faster. It is the fastest single-GPU this generation, accept it and move on.

Go read some older reviews by Xbitlabs with 12-13 games, 680 still lost! So either accept it or stop with your NV fanboy trolling because you claim this review used "too many" unfair games that put your precious 680 in bad light. Guess what if I owned a 680, I sure would love to know that it tanks in certain games. That's the whole point of reviews, it's not a pissing contest but an information highway for us gamers to see if we need to upgrade.

In this case not only is the 680 slower, but it costs more. Here is 1100mhz 7970 GE for $450 on Newegg:
http://www.newegg.com/Pro...aspx?Item=N82E16814125439

Until NV drops prices to $419-429, GTX680 is overpriced for new gamers because it performs worse overall (unless all you play is WOW, Project Cars, Lost Planet 2, HAWX 2, BF3 and Crysis 2 and Max Payne 3). If you play 100s of games, AMD delivers faster performance for less $. That's a win-win this generation.

Picking games only you like perfectly OK when you go out and buy a videocards for the games you play. However, that changes nothing that GTX680 is not the fastest GPU overall this round.
4 5 [Posted by: BestJinjo  | Date: 09/09/12 10:34:36 AM]
Reply

[1-1]

Back to the Article

Add your Comment