Is this a low 3D Mark 06 score for a GTX?

RandomDude

Distinguished
Feb 12, 2007
31
0
18,530
hey ppl,

just got a BFG 8800 GTX OC today. I benched it with 3D Mark 06, my specs are:-

C2D E6600 (2.4 GHz)
Corsair TwinX 2GB Dominator C4

I am scoring approx 9800 marks. My BFG 320 MB OC Card before this could hit about 8600.

Is my score too low?
 
Post your individual scores...CPU SM2 SM3
THat way we know if it is your CPU that is slowing you down.

If your CPU score is anywhere near 2000, then this is a VERY low score.
I am getting 12000+ with 2gts320's in SLI which is only marginally better than a GTX...by that logic you should be getting at least 11,000 or maybe 10,500 and more.

hmm.
 
oh by the way the version of your drivers can make a 20% difference!
And do not run ALL the tests, just the default ones.
Use an older version of the Nvidia drivers and you will get a better score if nothing else works.
 
3dmark is not a very good benchmark to gauge a system's true performance. I'd say the scores by 3dmark is accurate +/- 1000.

If you didn't overclock your system and ram and left your Bfg 8800GTX at it's factory speeds, I'd say that score is just right. However, the e6600 is bottlenecking the 8800GTX so your graphics card score should increase a lot more when you overclock your cpu. This is partly why I said the score is accurate +/- 1000.
 


I dont think the scores on 3dmark is in ratio with framerates, 12000:10500 seems in ratio with the SLI gts and the GTX framerates. The score should! be lower than 10500
 
Okay, a few more details.

I am running Windows Vista Home Premiium 32-bit (I know that Vista performs worse than XP).

Detailed 3D Mark scores:-

3DMark Score 9856 3DMarks
SM 2.0 Score 4668 Marks
SM 3.0 Score 4866 Marks
CPU Score 1991 Marks

This is on the basic (free) edition at default settings.
 
BFG 8800GTX OC + Core 2 Quad @ 2.4(3dmarks2006 set to run on 2 only)
Ram : 3.0 gigs @ 533 4,4,4,10

10233 3DMarks (well within the +/- 500 you can get from run to run....)
SM 2.0 Score 4836 Marks
SM 3.0 Score 4959 Marks
CPU Score 2120 Marks

Looks like your are in the right area after all..CPU adds a fair bit
--------------------------------------------------------
BFG 8800GTX OC + Core 2 Quad @ 2.4(3dmarks2006 set to run all 4)
Ram : 3.0 gigs @ 533 4,4,4,10

Main Test Results
3DMark Score 11803 3DMarks
SM 2.0 Score 4816 Marks
SM 3.0 Score 4992 Marks
CPU Score 3898 Marks

Hope this helps some....now back to City Of Heroes i go.... :) well that or ET Quake Wars Demo....
 

There can be variances with like systems with extra software running....

As many have said...it means nothings as long as the machine does what you want it to anyway....
 
I've already asked this like 3 or 4 times and people still don't give me a clear answer (or an answer for that matter)... Why does it matter your 3Dmark score at all? It's not like its going to do you any good. (I'm serious on this question)

And a system performance matters on games not on a synthetic benchmark, so don't start trying to give me an answer around that.
 
I thought 3D Mark scores were indicative of the power of the hardware you have. So if I get a GTX and expect high scores and don't get them, I immediately worry something is wrong. I could care less about the whole penis-envy thing that some ppl have with 3D Mark.

I just need to know that my hardware is not faulty in any way.
 
Honestly 3Dmark means squat, that's why I'm in a quest to seek the true answer to my question. Just benchmark with games, that's the real indicative of the power of your hardware.
 


Just because a benchmark result does not always translate to real world performance it does not mean that it has no value. 3dmark is an excellent controlled environment to use for tweaking. The OP can now see how his result compares to similar equipment. If something was working poorly on his setup that may be revealed by a low score when compared to the same equipment. He can also see the result of clocking up higher on both his processor and GPU. I guarantee you if he tweaks his system to score higher in 3dmark that it will translate to better performance in games. I'm talking strictly about tweaking equipment but not switching it out. It is well known that r600's do better in 3dmark06 than the real world and quad core cpu's bump your score far more than they impact current games.


 
R600's potential with optimized drivers is shown in 3dmark, but aside from that, you can compare GTX scores in there excluding CPU ones to see how your performance will meet up, comparing OC's or w/e.

9k is very low I get 11k with a GX2. So its probably a CPU Bottleneck. Looking at your CPU id say tahts exactly what it is, try OC'ing to 3ghz temporarily and see if your GPU score goes up.
 
it is a cpu bottleneck....see above :) i tested for just that....

@ Emp - Its a good way to see everything is working right......but nothing more.....

the difference from a gts to a gtx is at most 10fps and on average about 3-5 fps in 3d 2006....but real games can get more advantage....or no advantage....so what you play makes the difference...
 
I think 3dmark is great for comparing your own overclocks. I always look at the avg. fps and see how much more frames I get when I overclock. Then I divide it by the speed of the cpu or vidcard to find out the best fps/mhz and use that settings. If I were to use a real game instead, many variables are apparent and the results aren't as accurate. I mean if I score 1000 more points due to overclocking, it means I'm gonna get more frames right? 3dmark still has its uses, Emp.