Hard disk speeds and 3D Mark Scores

[jel]

Anyone know how much hard disk speed (seek time etc.) speed up 3D Mark Scores?
I am building a new system and am trying to decide whether I should fork over the money for the Raptors in a raid array or save the money and put it somewhere like better sound etc.

I see lots of reviews on how 3D Mark scores and other benchmarks improve with different chips or memory types but have not seen any comparisons holding all things equal and just putting in faster hard drives, raid arrays etc.

 

[txxxx]

I'd imagine like ..... 0 ? Loading screens exist for a reason

 

[jel]

But during a 3D mark test, doesnt the computer have to go to the hard drive for anything or is it all loaded into ram?

 

[Davegod]

i think every thing is loaded into ram before each tests starts, iirc there's something in the 3dmark faq (or somewhere) about it too, saying something about even systems without a lot of ram wont use pagefile.

 

[Sideswipe001]

Originally posted by: jel
Anyone know how much hard disk speed (seek time etc.) speed up 3D Mark Scores?
I am building a new system and am trying to decide whether I should fork over the money for the Raptors in a raid array or save the money and put it somewhere like better sound etc.

I see lots of reviews on how 3D Mark scores and other benchmarks improve with different chips or memory types but have not seen any comparisons holding all things equal and just putting in faster hard drives, raid arrays etc.

Not to be a troll here, but, doesn't it strike you as...stupid?...to buy a HD just for a higher 3dMark score?

Get a faster HD so your computer boots faster, or loads levels faster, that I can understand. But 3dMark?

 

[Pistolero]

it's his choice, lay off. ^_^
post in the overclocking forum or maybe try a different site.

 

[HeroOfPellinor]

Just to add some empirical testing to end any speculation. I happened to be building a new computer and had two disparate hard drives to choose from.

With a 120GB 8MB cache Maxtor my 3dmark2001 score was 13164. Identical setup with a Samsung 40GB 2MB cache was 13292. The discrepency is just statistical variance.