3Dmark questions..

[MrEgo]

Alright I've been reading posts on here about people with Athlon XP 1900, 2000, 2100 with a GF4 TI4200 pushing 11,000 on 3Dmark 2001, but only about 1,000 to 1,500 on 3Dmark 2003.

Here's my specs:
2x Athlon MP 1900+ (not that they help gaming)
Radeon 9700Pro by ATI, updated drivers
Kingston PC2100 Registered ECC 512MB DDR
EPOX M762A motherboard
Windows XP Pro

However, even with a Radeon 9700 Pro, I get between 11,200 and 11,300 every time. But I do get 4,280 on 3Dmark 2003. Is 3Dmark 2003 a better benchmark (as in a better measuring stick)? It seems as though that most of the cards that score just as high as me on 3D '01 fall way below my score on 3D '03. Why is that?

Second of all, how much of a role would you guys rate CPU performance in your score? Like... would you say it played about a 50%, 25%, etc. role? Your thoughts are appreciated.

 

[paxman]

There have been a few articles ripping 3D Mark 2003 for relying too heavily on the video card while ignoring the CPU.

I have a 1.46 AMD (can you say slow!) and a ATI 9700 Pro and got almost 5000 points while others with faster systems and Geforce 4 cards are getting 2000-3000 points. That tells me that it is relying on my video card almost entirely.

I plan to stick to 3D Mark 2001 SE and other gaming benchmarks (UT 2003, Quake III, etc.) to gauge my system.

Also remember that you cannot compare 3D Mark 2001 and 3D Mark 2003 scores. Apples and Oranges. My understanding at this point is that 5000 is on the high score end of things (which is why it makes no sense that I would be getting such a good score). A good score in 3D Mark 2001 SE is 12,000 and up (I get about 10,000).

Read this for a bit more: http://www4.tomshardware.com/column/20030213/index.html

 

[LiquidPaper]

3dmark03 is a dx9 benchmarking suite. 3dmark01 is dx7. You have nothig to worry about with those scores. Besides, they're synthetic benchmarks and don't reflect real world perfomance. How many games take advantage of dx9? None. It's more a gauge of system performance looking 6 months to a year (2 years?) ahead, depending on implementation of dx9 in future games. doom III should be the first to use it to any degree. CPU and mem bandwidth probably account fo 20-30% of the scores, if that. You've got a good rig (if you had any doubts) and one sweet card.

 

[MrEgo]

Thanks fellas. Actually I am getting rid of the Athlon MP 1900's and the EPoX mobo as soon as I get my 8RDA, 2100+, SLK 800, and Corsair PC3200 xms in the mail. Do you think I would see significant amount of increases if I was able to get a decent overclock out of it (maybe 2.1GHz)? I don't believe that MP's affect 3D performance, so I assume that the single XP would work better under a higher frequency.

 

[LiquidPaper]

pump up that fsb! this 8rda+ rocks. if I had higher memory (like that xms pc3200 you're getting) I'd have the fsb over 200 just with my 1700 tbredA. don't know of how much improvement you'll see in 3dmark03 but your 3dmark01 score should shoot up. that's a wholesale system improvement you'll be dropping in and the epox will let you tweak to your hearts content. post your results as you go along, I'd love to see what it can do.