Information

Did you know you could become an X-bit labs fan on Facebook or follow us on Twitter?



Discussion

Discussion on Article:
Palit GeForce GTX 680 Jetstream 2 GB Graphics Card and GeForce GTX 680 SLI Performance Review

Started by: BestJinjo | Date 05/26/12 10:43:08 AM
Comments: 4 | Last Comment:  05/28/12 02:10:23 AM

Expand all threads | Collapse all threads

[1-1]

1. 
Under Just Cause 2 gaming benchmark, you instead have a 2nd Metro 2033 graph.

Also, it would be great if you guys could do a review of GTX670 and do a GTX670 vs. 680 vs. HD7970 at maximum overclocks on air with GPU voltage @ 1.25+ on 7970!

My other recommendations are to get rid of Lost Planet 2 and HAWX 2. Those games are too old and not popular while L4D2 is not demanding enough to test new GPUs.

You could add Mass Effect 3 or Portal 2 and test GPUs ONLY in Super Sampling mode for those 2 games to show us how newer GPUs can perform with Super Sampling enabled in titles that permit this type of anti-aliasing without bringing performance to the knees.

You could also substitute Lost Planet 2 / Hawx 2 / L4D2 with:

1) Alan Wake
2) Tom Clancy's Ghost Recon: Future Soldier
3) Diablo 3
4) Max Payne 3
5) The Witcher 2 (very demanding and great game!)

Also, you could replace Dirt 3 with the newer Dirt Showdown.
1 1 [Posted by: BestJinjo  | Date: 05/26/12 10:43:08 AM]
Reply
- collapse thread

 
1)
Alan Wake
2) Tom Clancy's Ghost Recon: Future Soldier
3) Diablo 3
4) Max Payne 3
5) The Witcher 2 (very demanding and great game!)

Also, you could replace Dirt 3 with the newer Dirt Showdown.

How to test is these games?
Can you tell me?
1 0 [Posted by: Jordan  | Date: 05/26/12 01:39:53 PM]
Reply
 
If you played a particular game and remember a very heavy fighting/boss/scene with lots of explosions/characters on screen/action, etc. you could go there and run FRAPs utility for about 2-3 minutes.

For example, here is a sample of how Hardware Canucks does it. They record sequences from within a game and record the frames with FRAPs:
http://www.hardwarecanuck...uperclocked-review-2.html

Alternatively, you can just run a 5-10 minute game sequence which minimizes the margin of error of trying to find a particular demanding scene. Over a 5-10 minute game sequence you get a wider sample/average of in-game performance. Although I do realize this would take longer.

GameGPU.ru does this. For example in Alan Wake they recorded the first 10 minutes of the game:
http://gamegpu.ru/action-...s/alan-wake-test-gpu.html

By doing that, they are testing exactly how a GPU will run a game. You don't necessarily need to rely on pre-built benchmarks.

Also, while some benchmarks such as Far Cry 2 of Just Cause 2 or Dirt 3 are very good at representing real world performance, other times the "canned" benchmarks are not demanding enough. Remember Crysis 1 and how much more demanding the game got when you got to the ice/snow levels?
0 1 [Posted by: BestJinjo  | Date: 05/27/12 11:51:41 AM]
Reply
 
They do it manually, instead of automatically on a demo records. Therefore it is impossible to trust such results and I won't use them.
1 0 [Posted by: Jordan  | Date: 05/28/12 02:10:23 AM]
Reply

[1-1]

Back to the Article

Add your Comment


Our Sponsors

Crucial.com Western Digital G.Skill