I don't even have the damn program yet as the mirrors are swamped, but if your one of the lucky few... go ahead, make us sick.

The numbers will not be as high as they were in 3DMark 2001, especially if you do not have a DX9 capable card!

*glares at geforce3*

Stop it, your depressing me..

get it off kazaa...my score 884 1ghz duron with a ati 8500

The servers are SWAMPED.
I'm 56K too, so ill be last to get it.
Maybe next month I can report a score:(

1633
... :-/

I'd post the link but their site is still being hammered.

I'm not even having any luck with Kazaa:(

we are uploading it to our download section as i type this for members only.

Whoo Hoo!
Go 4peeps:)
Will it be up by 10:00pm(est)'ish?

pap i was just thinking the same thing

Tre, can you take a screenshot of the individual break down of the scores, or isn't that there anymore?
Like how 2001SE used to show you the scores you actualy got in each test.

916!! I AM SO GOD DAMN PISSED OFF RIGHT NOW

It's at 92% here. I would have waited to get it from here but I had already went looking for it before I saw that you guys were upping it. I just hope this one I have coming is in English

I got 5,020 on my first try, that was on Cat. 3.0a. Tried every easy tweak I could think of and it went down a 100. Installed the new Cat.3.1's and lost another 200. Unfortunatly, I didn't save the first score because I thought I could do a lot better. My computer is completely stock.

This test sure doesn't run very well on my computer, sometimes got frame rates as low as 3 or 4. I guess they left a lot of room for improvment or either I've got a problem, but I really don't care about the test as long as games run smooth and they do.

http://service.futuremark.com/compare?2k3=43460

My card is showing its age.

By the way, welcome to 4peeps olefarte

Ditto the welcome to olefarte! From what I've seen so far 5000+ is a respectablere. Remember that there will be absolutely NO correlation of the scores in 3DMark 2001 SE and 3DMark03! The tests are all completely different. Also, if you're not running a DX9 capable card (which means any currently available GF4 and below) your scores will suffer for it!

I know im going to be haded a bad beating because I can only run the one test for a score(the first one) since my GeForce4 MX doesn't do pixel shaders. Oh well, it runs Unreal Tournament 2003 nice for me, thats all tha matters:)

If you want to see a pretty comprehensive breakdown of 3DMark03, check out this article at Hard|OCP:

http://www.hardocp.com/article.html?art=NDI4

I want to see some compare links. I know you guys with the 9700's are itching to rub our faces in this :P

thats all you got teck???
Don't think I want to do mine then ether

I started downloading 3DMark03 before I left for work this morning. Hopefully it will be done and intact when I get home this afternoon. After what I read on Hard|OCP ( http://www.hardocp.com/article.html?art=NDI4 ), I'm not sure this is a terribly valid benchmark, anyway. It has only one hybrid DX9/DX8 game test, two DX8 tests, and a DX7 test that are used to generate the final score. There is only one of he other tests that is for DX9 and that is the Pixel Shader 2.0 test. I think we'll be seeing video cards compared using 3DMark 2001 SE Build 330 for a while yet, at least until 3DMark03 gets a few more DX9 tests in it, IMO.

The article does make me feel somewhat better about the screwy scoring. Especially since they say they're not going to be using it.

I'll run it on my gaming rig to see how things go, but I'm not going to bother with keeping track of the score generated. The weighting of the different test results is kind of screwy to me. The fact that each test is worth a different number of points isn't what I have a problem with, it's the vast difference in the weights and the seemingly random nature of the weights (other than the Game1 Test weight) that I question.

One other thing to note, which is also mentioned in the Hard|OCP article, is that you NEED DX9 installed to run 3DMark03!

Kyle Bennet, of [H]ard|OCP fame, has written an interesting editorial on the state of benchmarking in general and synthetic benchmarking in particular. It's worth a read at http://www.hardocp.com/article.html?art=NDMw

last nites upload failed , but it is going fine now with a ftp client.
If anyone wants the benchmark,get a hold of myself or ace at this point
just pm us.

yeah, good article.

I got a Geforce3 so as you can imagine, it really suffered, 1-4 fps through most of it.

Feels like when I had a voodoo 3000

Okay, well... I wanna see a score with a GeforceFX, that ought to be pretty high since that is it's whole league (pixel shaders, bump mapping etc)

any member want it?

get it in downloads.

Cool deal, Pappy! My download this morning died at around 116MB.

It SUCKS major ass anyway. Not worth the download.

I know it sucks, Ace. I've read several articles that detail how it generates it's score and what the tests are comprised of. It may get fixed in a patch, that's why I downloaded it.

5320 here stock,i think it is biosed.

Biased to what ?

I don't know,,,with the adds for allenware and all just spooks me.

I don't think I will download this.

From what I've seen so far, Pappy, your score is pretty good. The scores from 3DMark03 are roughly 1/3 of what they were in 3DMark 2001.

Yeh, Pappy, thats a great score. Right after downloading 3DM 03, I ran it with everything on my computer running just to see what it looked like, and got 5,020. Now everything I try to do to get it higher only makes it lower. Best I can get now is 4,779, sure can't figure out why.

I did not do it , but the peeps at madonion that post the highest scores format,reload just the drivers they need then run the benchmark.
fresh install is allways good for close to 1000 points on a hot machine.

Ya, that trick has been around a while.
You can make a logon for your machine, just give the user account power user privs, and you can set it to load only a chosen set of services and drivers. I've seen machines booting using only video drivers, no sound, network, help files, nothing. Just the bare basics so the machines can benchmark higher.

Also, with relation to the alienware thing, they did it with 3dm2001 with Falcon Northwest, rembember, "if you were running this on a Falcon, you'de be done by now".

Last, but not least, ill just hold off for a while then if scores are about 1/3 of what they are for 3dmark 2001, I score 4104 in 2001, so I can expect like, under 300 for 2003. BARF. Sounds like when I ran 2001 on my K6-2 450 with a Voodoo3 2000.

I'll settle for what I have, guess I'll keep playing with it though, as far as I'm concerned it's just another toy. The important thing to me is that all my games run okay. After reading about all the problems everyone is having just running 3DMark 03, I guess those of us that get it to run all the way to the last test should feel lucky.

~4,800 is the max i can get on my lowly 2.0a @ 2.4ghz p4 and 9700... so i'm fine with it.


um, thats not benchmarking, its called rigging a program to cheat the game... no one would ever play a game without internet or sound support. some people are 3dmark freaks, they've never played a real game in their life. i feel sad for them.

Possibly something like this http://www.overclockers.com/tips00286/

Vegetto, we actually agree on something!

I benchmark my system to see if something I did adversely my performance. I don't run benchmarks to prove to the world that my PC is faster than someone else's PC. In order for my benchmark numbers to be meaningful to me, I have to run the benchmark using the same configuration of services and programs that I would run while I am playing a game. Otherwise, it's just an exercise in mouse-clicking and waiting for the benchmark program to finish and spit out a number.

[img]http://groups.msn.com/_Secure/0RACvBCoTxnhf0o04bBod1wsRfSqra4rtmqeM6eRj4lGRNhz1jQRMa5WQPUiQzaABqp7QrJSrT7SIE57
PazuCX4Pmd2wmqqpTgF8svAt3YU/3dmark-03.JPG?dc=4675409536055982743[/img]

I ran it, just for fun, and got all of 4439. I'll post the link as soon as I can upload the project.

How about rigging drivers, thats the same deal. Rembember when the Radeon 8500 came out, ATi wrote into the drives that shipped with the card, special provisions that turned down the details in Quake 3, MipMapping junk, so the 8500 could post higher scores than the GeForce3's. After ATi was found out their driver "problem" was fixed quickly. It's hard to determine what cheaintg is anymore since everyone does it.

Just to get the other side of the coin in here, Nvidia did a little trick with 3DMark 2001 (non SE) where they would preload textures between game tests. This helped boost their cards' scores over ATI's. 3DMark put out the SE version in order to fix this, if I remember correctly.

the latest 3.x catalyst drivers seem to be living up to their names, nothing but good out of them. fsaa + aa looks fabulous (just like my voodoo5, except faster ) the reduction in mipmapping is an easy way to gain a few fps, and i wouldn't be suprised if nvidia does the same thing.

it didnt run at all on my box

3dmark03 vs GF2MX =

yeah, will as long as UT2003 can run, I'm happy.

Not impressed at all with 3dmark03.

Veg, did your 9700 totally bog down during the CPU tests? At the end of the second CPU test I was only rendering 1 frame per second. I don't know what they were intending to test there (probably software rendering mode, but I'm not sure), but it sure beat the hell out of my poor little 1800+!