Which 3dmark program to use?

[GaryKing]

Which 3dmark program should I use? There are at least 3 of them (2001, 2003, 2005) and I don't know what the difference is (besides the year numbers). I notice that people still use versions like 2001 etc. so I'm not quite sure which one I should get for myself.

Thanks!

 

[ts3433]

What are you using this for? Just benchmarking? Actual games are usually the best benchmark, though you can still use 3DMark to make sure what you've built is set up properly by comparing your score to systems with the same specs.

 

[imported_Kensai]

I like using 3DMark05 and the Doom 3 timedemo.

 

[GaryKing]

Also, my system is a P4 1.8 GHz but 3DMark05 supports 2.0 GHz and higher; so should I stick with 3DMark03 instead?

 

[bigboxes]

With only a 1.8 P4 your system will barely function with 3dmark05. It's meant to tax even the newer systems and will kill yours. Try 3dmark03 or 3dmark01.

 

[GaryKing]

Okay thanks, I'll give those a shot instead.

What's the difference if I use 03 and 01 anyways?

 

[mwmorph]

what video card do you have? i tests different things. i think 3dmark01 is dx7 an 3dmark03 is dx8

 

[fstime]

I like 03. 01 is mostly for benching the cpu....

 

[Matthias99]

Originally posted by: GaryKing
Okay thanks, I'll give those a shot instead.

What's the difference if I use 03 and 01 anyways?

Please use the search feature. Futuremarks' benchmarks are a frequent topic of discussion...

They're just newer versions, and the newer ones are MUCH more demanding on your video card.

3DMark01 -- uses DirectX8.1, only uses shaders in one test, VERY CPU/RAM limited with any reasonably recent video card (a GF4/R9XXX and up). So at this point this is more of an overall CPU/RAM/system test, unless you have a very old video card.

3DMark03 -- uses DirectX9, heavy shader load in some of the tests. Starting to get CPU/RAM limited with today's fastest cards and/or SLI.

3DMark05 -- uses DirectX9.0C, VERY heavy shader load in all tests. Crushes even today's fastest video cards.

Also, as stated above, benching with real applications is generally more useful if you want to know about performance in said applications. Synthetic benchmarking *is* useful in establishing baseline performance and comparing against it as you tweak things, or comparing against other systems to see how much performance increase you might see in an upgrade.

 

[GaryKing]

Okay thanks. Here's my rig info if it helps to determine which one I should use: http://www.pcpitstop.com/techexpress.asp?id=0ZKRMW4BNSQS0A46

But I suppose '03 is best for me.

 

[mwmorph]

3dmark 03, but iw will be incredibly low, albeit accurate
3dmark01 thoughi it'll be halfway decent but not as accurate.