Welcome Stranger to OCC!Login | Register

Asus EAH4870 Dark Knight Review

tacohunter52    -   April 20, 2009
» Discuss this article (4)

Lowest Prices

Testing:

It's time to find out if this "Dark" card can live up to its title of "Knight." I'll be putting this card through a series of game tests and synthetic benchmarks in order to see how the EAH4870 performs against its opponents. The "Dark Knight" will be paired up with the OCC test rig listed below. The CPU will be clocked at 3GHz and all video cards will be set at the driver defaults. All cards will be tested under the same game conditions in order to provide as few variants as possible.

 

Comparison Video Cards:

Overclocking:

Overclocked Settings:

  • ASUS EAH4870 DK 808/1015

I decided that I would try and push the "Knight" as far as it would go using the Catalyst Control Center. However, after I had gotten to a core clock of 850 and was still able to pass CCC's built in test, I decided something probably wasn't right. Like any sane person would do, I switched over to Rivatuner. That is after all what ANY sane person would do, isn't it? I began by increasing both the core and memory clocks by increments of 20MHz.. I was able to jump twice with stability, but on the third jump, not so much. With a core clock of 810 and a memory clock of only 960, I was fairly certain the core clock was the source of instability. So I lowered it to 809 and did a quick 3DMark Vantage run. To my surprise it actually worked, so I decided to test out L4D. Of course, I was greeted with a few minor artifacts. This means the occasional green flicker on the screen. I then lowered the core clock to 808 and reran both 3DMark and L4D, this time everything worked perfectly. Now that I had gotten the core clock set as high as it would go, I began working on the memory. At this point I was getting tired. It was 4:37 AM and I had now been at this for 77 hours straight testing for you, so I just randomly set the memory clock to 1020. I ran the tests and...they failed. I then began lowering the memory clock by 1MHz increments until I could successfully run through all of my tests, as well as a little extra game time in Fallout 3 without any problems. The resulting clock speed was 1015Mhz.. I know that these aren't the best speeds, but I was going for stability rather than just a "screen shot clock." However, I did try to see how far I could push the "Knight" and still semi-run the games. This turned out to be a core clock of 840MHz and a memory clock of 1096MHz. Believe me, when I said "semi-run" the games, I meant "semi-run" the games. I'd have a total play time of about ten minutes before a crash. All in all, I'd say a core clock increase of 58MHz and a memory clock increase of 115MHz isn't all that bad. It also isn't the best, but hey I'll take it.

 

 

  • Video:
  1. Far Cry 2
  2. Crysis Warhead
  3. BioShock
  4. Call of Duty World at War
  5. Dead Space
  6. Fallout 3
  7. Left 4 Dead
  8. 3DMark 06 Professional
  9. 3DMark Vantage



  1. Introduction
  2. Closer Look (The Card)
  3. Closer Look (Drivers and Programs)
  4. Closer Look (Catalyst Control Center)
  5. Specifications & Features:
  6. Testing (Setup & Overclocking)
  7. Testing: Far Cry 2
  8. Testing: Crysis Warhead
  9. Testing: BioShock
  10. Testing: Call Of Duty World at War
  11. Testing: Dead Space
  12. Testing: Fallout 3
  13. Testing: Left 4 Dead
  14. Testing: 3DMark 06 Professional
  15. Testing: 3DMark Vantage
  16. Conclusion
Random Pic
© 2001-2014 Overclockers Club ® Privacy Policy

Also part of our network: TalkAndroid, Android Forum, iPhone Informer, Neoseeker, and Used Audio Classifieds

Elapsed: 0.0265271664