[H] Core 2 Gaming Benchmarks

[Geo]

Holy thread resurrection, Batman!

[H] has finished tallying the votes from upstate, and would like to revise and extend their findings. . .


http://enthusiast.hardocp.com/article.html?art=MTI2MiwxLCxoZW50aHVzaWFzdA==

 

[Skrying]

I would highly suggest that Kyle just simply opt out of actually speaking or writing. Seriously, the guy can't even help write something without being himself on a platform and believing he is the savior of benchmarking (wait...).

I also had a nice giggle with the term "lagging" used in regards to offline play, now that must be one heck of a slow net connection.

 

[RobertR1]

Yep, real world gaming is all about 30inch widescreen with Sli'd 8800GTX's.

 

[cam]

I don’t understand, if you’re comparing like for like why run certain games on a slightly higher setting for one platform and not the other.

Surely if one system bottlenecks at certain level and the other one doesn’t, you use the level the one bottlenecks at and apply it to both, or am I missing some basic testing fundamentals?

 

[Skrying]

HardOCP does that. They determine if a card can handle certain settings and then give the FPS results. So if they feel a 8800GTX can handle full settings plus 4xAA but the X1950XTX can only do full settings with no AA then they'll give the results for those settings and simply not test the same results on each card.

 

[pjbliverpool]

I don't get why they run some tests with MSAA on the Intel and not on the AMD. Surely whether you have MSAA turned on or not should have no baring on CPU performance!

Also, is it just me or does the suggestion that 768 MB of graphics memory not being enough for Oblivion just seem wrong? Almost seems like a plug for "something else" which may be coming along soon.

 

[Skrying]

I don't get why they run some tests with MSAA on the Intel and not on the AMD. Surely whether you have MSAA turned on or not should have no baring on CPU performance!

Also, is it just me or does the suggestion that 768 MB of graphics memory not being enough for Oblivion just seem wrong? Almost seems like a plug for "something else" which may be coming along soon.

Its a CPU vs CPU article, they are trying to suggest that not being able to run MSAA would be the CPU holding you back with all other things equal.

 

[pjbliverpool]

Its a CPU vs CPU article, they are trying to suggest that not being able to run MSAA would be the CPU holding you back with all other things equal.

Yes but MSAA performance is 100% GPU reliant as far as im aware so a different powered CPU should make no difference to it at all. In fact if anything you should be turning on MSAA with the slower CPU since its more "free" in that case as the game is already CPU bound, thus you want to increase the GPU's work load to bring it in line.

 

[Moloch]

I like how the article talks about feeling things, feeling the framerate is lower, feeling like more videoram is needed (my fav).
Truthyness galore!

 

[KimB]

Yes but MSAA performance is 100% GPU reliant as far as im aware so a different powered CPU should make no difference to it at all. In fact if anything you should be turning on MSAA with the slower CPU since its more "free" in that case as the game is already CPU bound, thus you want to increase the GPU's work load to bring it in line.

Well, if everything is working as it should, then some frames should be GPU-limited, and others CPU-limited. Thus, when you turn up GPU-bound things like MSAA, you will be lowering your framerate on GPU-limited frames, which in turn lowers the average framerate. If the benchmarker is only paying attention to the average, he/she is likely to conclude that you want lower graphics settings on a system with a slower CPU.

This is, as you mention, largely nonsense because doing this probably isn't going to have much impact on your minimum framerate, but will rather more affect the maximum, and will thus not be as noticed.

 

[Skrying]

Yes but MSAA performance is 100% GPU reliant as far as im aware so a different powered CPU should make no difference to it at all. In fact if anything you should be turning on MSAA with the slower CPU since its more "free" in that case as the game is already CPU bound, thus you want to increase the GPU's work load to bring it in line.

I know this.

I was pointing out why HardOCP does it. Did I say it makes sense? No, because it doesn't, its just part of the program for HardOCP...

 

[poopypoo]

HardOCP does that. They determine if a card can handle certain settings and then give the FPS results. So if they feel a 8800GTX can handle full settings plus 4xAA but the X1950XTX can only do full settings with no AA then they'll give the results for those settings and simply not test the same results on each card.

I'm fairly certain that they still test the games at each resolution and IQ setting, but the trouble is that ofc ourse they don't publish any but the results they find most meaningful -- leaving you without the ability to make any informed decisions for yourself. It' a shame, since their whole max/min performance graph approach was truly revolutionary and great -- then they ruined it with this subjective garbage. Ah well. (ps Skyring I know you already knew this -- just trying to clarify in case anyone missed the obvious. I'm prob just resurrecting a dead horse -- if so, sorry...)

 

[ANova]

The guy obviously doesn't (or won't) grasp the concept of hardware specific bottlenecks and how it relates to overall performance.