Overclock.net banner

Team Green vs. Team Red - The Friendly OCN Battle Royale

35K views 516 replies 79 participants last post by  MKHunt 
#1 ·
Team Green vs. Team Red

Update - October 2nd, 2011 - I am going through some potential..... RMAs right now. Some of you may or may not know, I sold my 590s. I do have a 6990, and possibly a 6970 for benching purposes. MKHunt has taken the huge responsibility of aiding me in keeping this thread alive. Who knows, we may keep it going for the 7xxxx and 6xx series. Once we develop a reliably oiled machine.

THANK MKHUNT! Thats an order!!!!


Backdrop:
Its come up in numerous threads, its about time we had a little bit of a friendly competition. This thread was entirely dedicated to the 590 vs 6990. I felt that the 590 wasn't properly represented in the majority of benchmarks, nor do I deem the card a "failure." This thread was in good spirits to show how the two of the baddest boys on the block faired against one another, and each other. Due to a lack of enthusiasm from the Red Camp for whatever reason, I have decided to expand the testing to a proper Team Green vs. Team Red fashion. While the thread is always in progress, 580 and 6970 results will be hosted in Single/Dual/Triple/Quad fashion, separately so that they may shine in their single/multi-monitor glory.

Note: As of now, the 580 and 6970 Spreadsheets and results have not be posted yet. I apologize for any hopes being raised, my whole system currently down, and I am not a fan of spreadsheet creation ala Netbook.

Rules:
Failing to follow the submitting procedures means we will disregard any score, no matter how good.

I ask for a good clean fight, discuss in a civilized manner, that means no flaming, no flame bating, trolling, etc.

Benchmarks Running:
3DMark 11 v 1.0.2 - Performance
(Started August 1st, 2011)

Heaven 2.5 (Max Settings (Tess Extreme) 1920x1080)
(Started August 4th, 2011)

Heaven 2.5 (Max Settings (Tess Extreme) 5760x1080)
(Started August 6th, 2011)

Metro2033 (Max Settings 1920x1080)
(Started August 6th, 2011)

DiRT 3 (Ultra DX11 Ranked Settings 1920x1080)
(Started August 9th, 2011)

DiRT 3 (Ultra DX11 Ranked Settings 5760x1080)
(Started August 12th, 2011)

Stock RUN Explained:
When I asked for a stock a run, I pertain to ONLY the GPU's at stock. I wasn't trying to be annoying and make people drop their CPU/System overclocks for no reason. It was to compare the difference of stock vs. overclock.

Settings Restrictions:
All Users - All Benchmarks:
CPU Max: 4.4GHz +/- 25 MHz due to Bclk variation.
Max CPU Clock Reasons:
  • There is not much bottleneck (if any) past 4.4GHz.
  • Not everybody (i7 9xx vs 2500K/2600K) can get past 4.4GHz.
  • These results represent an atypical enthusiasts setup.
  • Keeps people on air cooling capable of benching.
There are no other CPU/System restrictions at this time. I will be testing out various Uncore/QPI speeds as I go a long to see if it makes a difference as well. Later on we may take one or two select benchmarks and make them "suicides", a true best of the best. I have a feeling this will be real popular.

nVidia Users - All Physx Benchmarks:
Physx must always be disabled when applicable. Contemplating restricting driver use to 267.85 and above (due to safety concerns.) 267.5x and earlier may indeed not have certain PDL restrictions but also may cause damage. For now I will leave it open as it may be the only way to benchmark with higher than 720~730 MHz in certain applications such as 3DMark11.

ATi/AMD Users - All Benchmarks:
Must submit a benchmark with Driver Level Tessellation enabled and 100% AS well as one that is disabled, or for Heaven, submit a screenshot with cobblestones or rope, or spikey dragon showing tessellation is indeed on. Sorry, but thats the way it is. The same goes for the initial stock run. HWBot may be okay with people manipulating scores to have higher benchmark scores by disabling tessellation but I am not. Lets help keep each other honest.

Results:
ALL results will be recorded and ranked, similar to HWBot's system, multiple submissions will be kept, they do not overwrite each other! Every result counts! You are not just doing this for bragging rights, but for the good of the community! If you submit one result only, you don't really do anything but show off e-peen. If you submit trials and incremental overclocks, people can use those as templates.

3DMark11 Results Area:

Submission Format Example For 3DMark 11 - Performance:
Quote:
1st Result (for this benchmark)

Verification URL
http://3dmark.com/3dm11/1609418

CPU 4.4 / QPI 3.6
DRAM 1600 / Uncore 3.6 / PCI-E 110
GTX 590 SLI - 650 - 1300 - 864
GPU Voltage: 0.938
Driver Version: 280.19
3DMark11 Score: P16017
Proof (of score AND settings):
Screenshot
Screenshot Requirements:
The more information you can display the better, but for the bare minimum, you need 3 CPU-z's open (displaying CPU/Mainboard/Memory Tabs), at least ONE GPU-z open. The only one setting I don't ask you to prove, because not all motherboards display this info outside of the BIOS, is your PCI-E Frequency. In essence, if you claim a setting, you have to show proof of it. We also need to make sure people are following the CPU frequency ceiling.

Unfortunately taking a screenshot of a website can actually be faked within a few minutes by those that know JScript and HTML, so you must have your score on the 3DMark11 Program, showing all the sub-scores as well. I recommend including your name in a Notepad, so nobody has any doubt.

Example Screenshot:
attachment.php?attachmentid=222733&stc=1&d=1312945833


3DMark11 - Performance v 1.02 Results
http://spreadsheets.google.com/pub?key=0AgLiKaiBbWsXdFlTN0VVT055b3V6czJnUmpRMzl5N3c&w=93&h=500&gid=0]The Performance Runs[/URL]

Heaven 2.5 Results Area:
Heaven v 2.5 - 1920x1080 Settings:
API: DX11
Shader: High
Tessellation: Extreme
AF: x16
AA: x8
Full Screen: Yes
Resolution: 1920x1080

Heaven v 2.5 - 5760x1080 Settings:
API: DX11
Shader: High
Tessellation: Extreme
AF: x16
AA: 0xAA and 4xAA (590s can't run this resolution at 8xAA)
Full Screen: Yes
Resolution: 5760x1080

Submission Format Example For Heaven 2.5:
Quote:
1st Result (for this benchmark)

CPU 4.4 / QPI 3.6
DRAM 1600 / Uncore 4.2 / PCI-E 115
GTX 590 SLI - 630 - 1260 - 864
GPU Voltage: 0.925
Driver Version: 280.19
Heaven Score: 2272
Proof:
Image 1
Image 2
Example Submission: http://www.overclock.net/14460933-post181.html

Screenshot Requirements:
Okay two screenshots necessary, one showing the settings either pre/post test run, I prefer post as GPU usage gets included, but its up to you. You need to show your claimed settings obviously, exception still being PCI-E Frequency. The other one is after your benchmark has ran with the score on screen. You will need to use a image capturing program like Afterburner and take a screenshot of BOTH the score window AND the background having tessellation. It doesn't have to be the mr. dragon, it can just be simple cobble stones. Doing so may prevent AMD users from having to RUN it twice with Tessellation enable/disabled. However, AMD also allow increments of tessellation, so I don't know how I feel about this we will see how it goes. I hope the good nature of one or two guys will call out scores that seem incorrect.

Heaven v 2.5 - 1920x1080 - Extreme/Max Settings
http://spreadsheets.google.com/pub?key=0AgLiKaiBbWsXdENEVEFFM0ptTUNIYnI4MDhMdjlteVE&w=93&h=500&gid=0]The Heaven Runs[/URL]

Heaven v 2.5 - 5760x1080 - Extreme/Max Settings, 4xAA/0xAA
http://spreadsheets.google.com/pub?key=0AgLiKaiBbWsXdHItTl9DS0FjQm93QzNYaEw5NE5XQ1E&w=93&h=500&gid=0]The Heaven Runs[/URL]

Metro2033 Results Area:
Metro2033 - 1920x1080 Settings:
DirectX: DX11
Resolution: 1920x1080
Quality: Very High
AA: MSAA 4X
Texture Filtering: AF 16x
Full Screen: Yes
Enable Advanced PhysX: No

DX11:
Tessellation (Enabled by Default)
DOF: Yes

Scene: Frontline
Runs: 3

Settings Image:
m2033settings.png


Submission Format Example For Metro2033:
Quote:
1st Result (for this benchmark)

CPU 4.4 / QPI 3.6
DRAM 1600 / Uncore 3.6 / PCI-E 115
GTX 590 SLI - 630 - 1260 - 864
GPU Voltage: 0.925
Driver Version: 280.19
Avg. FPS: 71
Proof:
Image 1
Image 2
Screenshot Requirements:
Okay two screenshots necessary, one showing the settings either pre/post test run, I prefer post as GPU usage gets included, but its up to you. You need to show your claimed settings obviously, exception still being PCI-E Frequency. The other one is after your benchmark has ran with the score on screen. The second image will consist of the benchmark results HTML page displaying all three runs and the average. You may have to shrink it in browser (CTRL+ Mouse Wheel) to get it all in a screenshot. As long as you can read the data it should be fine. I suggest keeping the actual benchmark documents in your "My Documents" folder in case someone questions the results, or, it can't be visible.

Example Submission:
http://www.overclock.net/14489620-post210.html

Metro2033 - 1920x1080 - Extreme/Max Settings
http://spreadsheets.google.com/pub?key=0AgLiKaiBbWsXdHRoZGFFQTlDT0NaMVBrYUJWOTI0OVE&w=93&h=500&gid=3]The Metro 2033 Runs[/URL]

DiRT 3 Results Area:
All DiRT3 scores can be validated here:
http://benchzone.adrenaline.com.br/

Simply add the ID# to the end of the address for example:
http://benchzone.adrenaline.com.br/262


DiRT 3 - 1920x1080 4/8xAA:
Game: DiRT 3
Quality: Ultra
Resolution: 1920x1080
Anti-Aliasing: 8x or 4x
Route: Norway/norway_rally/route_0
Cars: 8
API: DirectX 11

Ranked: Yes
Signed In/Joined: http://benchzone.adrenaline.com.br/
Benchmark Download: http://benchzone.adrenaline.com.br/downloads/downloads.html
Submit Results for ID#

Settings Image:
dirt3settings.png


Submission Format Example For DiRT3:
Quote:
1st Run DiRT3 1920x1080 4xAA

CPU 4.4 / QPI 3.6
DRAM 1600 / Uncore 3.6 / PCI-E 115
GTX 590 SLI - 630 - 1260 - 864
GPU Voltage: 0.925
Driver Version: 280.19
Avg. FPS: 164.0
Proof:
Image 1
Screenshot Requirements:
Okay using the Adrenaline benchmark tool, you just need to submit scores online, and take a screen shot with GPU-z, and the 3 tabs of CPU-z open. Once the results have been submitted, you will be given an ID#, this is similar to the 3DMark11 verification, and a really cool feature, but you must be signed into Benchmark Adrenaline.

Example Submission:
http://www.overclock.net/14526674-post224.html

DiRT 3 - 1920x1080 - DX11
http://spreadsheets.google.com/pub?key=0AgLiKaiBbWsXdEtLcmxmTmpfNEhtVnp4Q2lsOHk4VlE&w=93&h=500&gid=1]The DiRT 3 Runs[/URL]

DiRT 3 - 5760x1080 - DX11
http://spreadsheets.google.com/pub?key=0AgLiKaiBbWsXdFdFd1VKd2lmbUZxRkJWa0xBVzd1Z3c&w=93&h=500&gid=2]The DiRT 3 Runs[/URL]

Reference 580 + 6970 Results Section
http://spreadsheets.google.com/pub?key=0AhBvVst5fpYydG9DSDljUXFEM1BhTlFLSHA5Zmk3Q1E&w=93&h=500&gid=0]580s vs 6970s Runs[/URL]
 
See less See more
3
#4 ·
subbed.

Regardless of results, I hope people can be respectful and not turn this into a flame fest.

Let the results speak for themselves.

I suggest adding...

Mafia 2 since it has a benchmark tool as well, run it with Physx off

What other games have benchmarking tools?
 
#6 ·
Quote:
Originally Posted by 2010rig;13901137
subbed.

Regardless of results, I hope people can be respectful and not turn this into a flame fest.

Let the results speak for themselves.

I suggest adding...

Mafia 2 since it has a benchmark tool as well, run it with Physx off

What other games have benchmarking tools?
Yes, please don't flame, friendly trash talking won't make me butt hurt, as long as you don't trip up over my benchmarks
wink.gif
DO NOT TROLL or FLAME!

I do not have Mafia, but if I get any other requests that push for it, I might pick up Mafia off of Steam.

You can view a list of everything I have on STEAM, plus any other benchmarks you guys might be able to download, I don't want this to be a victory over not enough games to test ya know?

http://www.steamcalculator.com/id/RagingCain

Thats everything I have on Steam, feel free to suggest others too.
Quote:
Originally Posted by ReignsOfPower;13901331
I B4 Levesque my-1k-setup-is-better-than-green.
Goodluck
tongue.gif
I'd play but I'm not using the BIOS fix
frown.gif
Use the bios fix then? Or use 267.85 drivers, I am going to have to for the Heaven benchmark. I don't know if we can "undo" the power draw limit, so we will be limited in 3DMark11
frown.gif
 
#9 ·
#10 ·
Quote:
Originally Posted by windfire;13901519
There are a 590 club and a 6990 Club.
Perhaps OP can go over to the 6990 Club and invite them over here for a friendly competition.
http://www.overclock.net/amd-ati/968084-amd-radeon-6990-club.html
Beat you by about half an hour
wink.gif


Its late on a Thursday, its college night, I am sure all of us are picking up chicks....

I unfortunately am neglecting my college work, while my chick is studying econ next to me on her dirty Mac....
Quote:
Originally Posted by ttoadd.nz;13901489
I nominate Megadeth
biggrin.gif
Me too, guys got a killer rig. I am also working with Khalam, to see if we can come to some sort of agreement on his Heaven score.
 
#12 ·
Quote:


Originally Posted by Pentium4 531 overclocker
View Post

Crysis 2 for sure, test Grand Theft Auto 4, and BlOPS as well (BlOPs = Black Ops)

We could throw in a Crysis 2, I heard they have a benchmarking tool out. Does GTA4 have any benchmarking tool to test? Thats a heavily CPU dependent game as well, a i7 980x might be better for it than i7 2600k.

Black Ops with a single 580 is 80 fps, Black ops with a SLI 580 was 80 fps, BlackOps with a single 590 was 80 fps, Black Ops with SLI 590s was 80 fps. If you guys want it and it has a tool we can do it, but its not a very GPU heavy game.
 
#13 ·
Quote:


Originally Posted by RagingCain
View Post

Yes, please don't flame, friendly trash talking won't make me butt hurt, as long as you don't trip up over my benchmarks
DO NOT TROLL or FLAME!

I do not have Mafia, but if I get any other requests that push for it, I might pick up Mafia off of Steam.

The Mafia II demo has the benchmark tool.


You don't have to buy it.
 
#15 ·
Quote:


Originally Posted by tsm106
View Post

F1 2010 has a tool. Imo, just skip heaven altogether.

We may end up doing that, or we can run it with Tessellation disabled.

IF tessellation can be disabled though by Catalyst 11.2/11.3 and beyond, we have to rely on the integrity of all ATI brothers in our DX11 games with tessellation.

Quote:


Originally Posted by 2010rig
View Post

The Mafia II demo has the benchmark tool.


You don't have to buy it.

Excellent news! I will add that to the list.
 
#16 ·
Interesting, subbed. Livin the dream, you are.
 
#18 ·
Since no one from the red team has come here yet, might as well address this.

Does anyone know how to disable / reset this so it's a fair test?

Quote:


Originally Posted by Shinobi Jedi
View Post

If you care about Bench scores then maybe go with the 6990's.

If you care about Performance and Image Quality, then you want to stick with the 590's..

Personally, I'm a gamer so I care more about Performance and Image Quality...

If you check Guru3D.com's reviews of the GTX 590 and similar cards, you'll see a tiny blurb where they mention that ATI/AMD employs optimizations in their drivers purposely to inflate bench scores at the cost of image quality.

-----------

Quote:


Originally Posted by RagingCain
View Post

Oh yes, the nerdy-white-can't dance-ex-navy and divorced-didn't learn my lesson and dating again-unemployed quality assurance analyst-back to college at 26-math programmer-dream!

Oh boy, if it wasn't for sex, I would have killed myself


A bit too much information, you forgot your Name, Address and SSN. LOL
 
#19 ·
Quote:


Originally Posted by 2010rig
View Post

Since no one from the red team has come here yet, might as well address this.

Does anyone know how to disable / reset this so it's a fair test?

I don't think there is anything we can do. What we can do is set up custom settings in the nVidia drivers. We can always lower our Image Quality and Compare Results.

What it comes down to it, I want to avoid as many shortcuts and tweaks you can do with the drivers. If AMD's proprietary drivers have 10% less image quality, then that's AMD's prerogative. We will be posting screen shots of our mayhem of both sides of the fence. If anybody feels like that someone is cheating or having an unfair advantage it should be easy to spot. I have seen the comparisons of Image Quality and I often have a hard time telling which is which to be honest.

Quote:


A bit too much information, you forgot your Name, Address and SSN. LOL

I am sure LulzSec is all over it
 
#25 ·
I will be watching this as well.... I would enter, but since I have an amd chip, my scores will be lower from the starting gate.....

I am currently attempting the 4ghz with my cpu, if someone with a similar system wants to go against me with a single stock 6990, then I would be game....
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top