3dmark11 Rank thread

Ok made a post because I thought it was interesting. Removed one 580 coz its air cooled (doesn't like 4.53GHz effective) ;)

3DMark11 Basic Scenario:



  1. Stock CPU as Physx dedicated - Stock GPU @ 1.1V (neede for BF3 stability).

    Untitled-4.png

  2. Auto physx standard oc

    Untitled2.png

  3. 2600K at 4.8GHz - Stock [email protected]

    Untitled3.png

  4. 2600K (55oC) with GTX580 at 906MHz at 1.15V (47oC)

    Untitled4.png

Im going to pump it to 5GHz and see if the CPU is bottling.. I doubt it.

*edit*

2600K @5GHz 1.5V 36oC 100% load (ambient 25oC) - GPU 906MHz (4.53GHz effective)@1.15V

Untitled5.png


2600K @5.2GHz 1.5V 36oC 100% load (ambient 25oC) - GPU 906MHz (4.53GHz effective)@1.15V

Untitled6.png
 
87dtna said:
No GPUz in the screenshot...
Click to expand...

I wasn't asking for verification but thanks anyway ;) I actually see a GPUz in the screenshot at the end... but anyhoo.

Two things that are interesting, I think by showing different states of CPU clock and GPU clock increases that PhysX plays a pretty significant role in 3DMark11. Auto selection in nVidia control panel seems to slow things down, but I only did one test. ;)

I also noticed that at 5.2GHz the CPU no longer is at 100% load during CPU testing etc in the benchmark. This also shows in the figures with the significant yet unrewarding return between 5.0GHz and 5.2GHz. I wouldve expected a bigger return, yet the GPUs could feed it faster.

Ive decided to watercool the other 580, so when thats in I'll test again. I think spesh had 13,000 in 3DMark11 with GTX580 TRI-SLI. If thats the case, thats an even more minute retrun on investment in this benchmark.

The data has a correlation coefficient of 0.94 which is pretty strong.
 
Here is my first 3DMark 11 score with my new toy. Everything is stock:



I just read the rules, so Ill run it again and post another screen. I just didnt think I would be in the top 5 with my CPU.

Here is my second run. A little lower, but then I had a steam convo running in the background so maybe thats why :p

[IMG=http://img220.imageshack.us/img220/782/3dmark112.jpg][/IMG]

(I decided not to re-size the screen as the text would be alot harder to read and it appears I did mess up the date with a 2 instead of a 6)
 
Last edited:
Not not at all. The 680 doesn't consume near as much power as a 570....in fact it's pretty close to 560ti wattage.
 
Last edited:
I wouldn't want to run a gtx 590 on less than 700w PSU....not because of the total wattage but because typically PSU's less than 700w wouldn't have enough on the 12v rail to support it.....a 590 can pull well over 30 amps if overclocked, they pull 29 stock.
 
Last edited:
SmileMan said:
yeah but doesn't the gtx680 requieres at least a 700+ psu with good amps?
Click to expand...

The TDP of the 2GB version of this card is 195W. I assume the 4GB is not much more. Compare that to my G80 which was 145W, and throw in the fact that I dont have a game that can fully utilize this card, Id say my PSU can handle it just fine.

nvidia does recommend a 550W PSU with a 32A 12v though, I have 3 x 22A 12v but I guess it works. Havent had a problem yet. I was also afraid my PSU would not be able to handle it, as its gettings rather old at around 5 years old already. Surprised its still holding up as well as it is, I keep my PC pretty clean though so the inside of my PSU still looks shiny and new so I think that has helped to pro-long its life by keeping it dust-free.
 
Last edited:
Ok so heres with some changes:





I turned off speedstep so to display what was my current CPU speed, which in turn turned off the turbo mode and I did not feel like manually overclocking it to 3.34Ghz, not that it would have made a huge difference. I also enabled XMP in my bios (not sure exactly if it made any difference either) and disabled Windows Aero for the test and AVG.
 
You must log in or register to reply here.
Back
Top