Poll: Do you consider 3DMark a valid benchmark?

[Odin30]

I think i agree with the stability statement.
Good to test stability bad for measuring system performance.
Everytime i run it i get different scores and ive gottin as high as 7500 marks and now am staying around 6800 even though im running at a higher clock speed and better settings.
i think it will stress your system, whick is good to find stability issues but the bragging rights bulls**t is worthless.

 

[Noriaki]

I don't care much for 3DMark. I'm not sure if it's a vaild benchmark or not, but it's not one I give a damn about.
People say stuff like well UT is a bad benchmark becuase it doesn't scale well in CPU speed increases, yadayadayada... Is that true? Probably. Do I give a sh1t? no not really. I only have one CPU so what do I care how well it scales, it's a good game I like to play so I want to know how my system performs in it.

In my opinion 3Dmark tells me nothing...it might be good for seeing well did that driver set up my performance at all? But I use my system for gaming, I want to know how well it performs in real games, not synthetics, I don't care about the difference between 90fps and 100, I just want my game to run playably.

But it's not like I play 3Dmark games, and so what it uses D3D, wowee...it's not like it's a D3D egine that's used anywhere, I'm much rather see how a real D3D engine works, like the Unreal engine that is used in Unreal, UT, DNF, Deus Ex, etc, etc. And every engine is different, you can't benchmark a generic D3D engine and say this is the be all and end all of D3D performance...bench one that's used.

Final answer: Is 3Dmark valid? Maybe, depends how you define valid. Is it useful? No.

 

[Modus]

The objective of any good benchmark is to guage a product's real world performance. Thus, a real world benchmark like Quake3 Arena, UnrealTournament, or Expendable is inherently better than a synthetic test like 3DMark 2000.

3DMark does have some limitted uses, most of which have been mentioned above. Like all synthetic benchmarks, it's a great tool for telling you why a certain card might be performing differently from another in certain situations. Its detailed tests on T&L performance, fillrate, and texture transfer can give you clues as to what is most limitting the performance of a card under a certain game.

But as a stability test, 3DMark is probably not the best way to go. There have been too many reports of anomalous crashes even when the rest of the system is rock solid. It just doesn't appear to be well written software.

Modus

 

[Taz4158]

I rely more on Villagemark which gives you more of a test of what the future in games MIGHT be.

 

[BenSkywalker]

"The Score is just a general score figure, what u want one number to tell you everything?"

They shouldn't report it. They should not give a value, it is too easy for people to look at that number and judge the test based on it(look to this thread for evidence). The bench is the best at what it does, easily, but too many people look at that singluar score as if it means anything to them.

"But it's not like I play 3Dmark games, and so what it uses D3D, wowee...it's not like it's a D3D egine that's used anywhere"

The game should be shipping not too long from now. It is an actual game based engine, not a generic one created for a bench and not an extremely poor port of an engine based on another API.

 

[Muerto]

I've never really used it too much as a benchmark. It makes for a decent stability test though, and the demo looks kinda cool. Although I think that the CPU test is fairly good. It seems to scale relatively well with CPU performance.

 

[vvvvvv]

yes

 

[apoppin]

I like it.

When used together with other benchmarks (WinTune, Prime 95, SiSoft Sandra, etc.), it gives a good indication of the performance and stability of your system (in relation to other systems and to changes you have made in your system). It also places particular emphasis on the video card and is useful when pushing the O/C'ing envelope.

I can't wait for 3DMark2001.

 

[JellyBaby]

Another useful feature of 3DMark is the online database. Once you submit your system's test results to their server, you can see how other PCs with the same video card but different CPUs perform relative to yours. You can also see how other PCs with the same CPU but with a different video card compare to yours.

 

[Syborg1211]

I use 3dmark as a system tester to see how good my settings are. I really find the test a very good tester, but I think that 6400 for my geforce 256 ddr is a bit high while my friend with a gts gets 6900