How do y'all benchmark 3DMark11?

I just bought 3DMark11 in addition to my 3DMark (13) that I have, when I launched it, it gave me 3 options for benchmark the default is P (performance) and then E (Extreme).........so what do you all use? is there a standard? It wouldn't be fair for someone to compare to diff. scores of P or E so I wanna know the default that everyone here uses?

 

Go to basic tab and then tick run benchmark tests only. It should be on performance (P) only.

Good luck

 

^ The standard is Performance, and folks will typically state outright if the numbers they posted were run at a different quality setting.

 

Selecting the Stretched option under Display Scaling Mode versus Centered also seems to provide a slightly higher score.

 

Yes, definitely do stretched. With centered there is a framerate limit... I think it is like 120 FPS limit, don't recall the exact value. If you have a multi-GPU system powerful enough to exceed that, your benchmarks will be capped. If you have weaker GPUs or single GPU, the impact on the score will be less because they struggle to reach the framerate where it gets capped off.

It is also useful for isolated testing of GPU and CPU performance, overclock stability, etc. I use it frequently for testing CPU overclock stability (Physics Test). Just go to the advanced tab, check only the test you wish to run, and go. If you are pushing GPUs to the edge it is also useful to find out if your overclock will pass individual tests. For example, what might pass in test #3 with flying colors might freeze or cause a BSOD in test #4. You can find out easier this way and focus your effort on optimizing for the problematic test. The folks using the free version do not have access to this functionality.

When you run individual tests for checking stability and whether tweaks help or hurt, you can also run them in window mode. You can watch clocks and temps using HWiNFO64 when you do this. You are testing and evaluating the effect of tweaks, so running full screen to try to achieve a better score is not important in that scenario. You can still measure the up or down influence of tweaks in window mode while monitoring your PC's vital signs.

 

Does the stretch method help in 3Dmark (non-11)?

 

Thanks everyone

 

TBH, I don't know.

But based off of the way it (3DMark 13) behaves by default, it goes full screen on its own. At least when I have ran it, it does. I have never bothered to run it Windowed.
The only reason I even stumbled across it is because when I had my M18x R2 with 680m SLI, my 3DMark11 GPU scores were always 1000-1200 points lower than others that were running the same clocks as I was. It drove me batty. Until I realized that I was running the benchmark in Windowed Mode. Once I switched it to Scaled, my scores were in line with everyone else.