What are good 3DMark scores?

Nazo

Nazo

2[H]4U
Joined
Apr 2, 2002
Messages
3,672
Ok, I KNOW this has been done before. Heck, it used to be stickied somewhere, but I can't seem to find it and the forum keeps telling me the search function is disabled, so I can't use that. (I strongly recommend fixing this!!!)

Anyway, my current best scores are in my signature. I'm a little worried that these numbers are a lot lower than they should be and I was a little dissapointed with this card's performance as I thought it was supposed to be better to replace my old card with when it started messing up. The scores are slightly higher, but primarily just because of the pixel shader tests where this card leaves the Ti4200 choking on dust. I noticed that with the high polygon tests, the 4200 actually seems to have done better and I've noticed that some things that ran tolerably before (Deus Ex: Invisible War) run much worse now. Ok, I had shadows disabled and all, but otherwise it was set mostly ok. Anyway, I was wondering what other people's scores are too. It would be useful to have something to compare with.
 
that score isnt abnormally low for a 9600 Pro. I wouldnt worry about it
 
Well, for one I just thougth I saw someone post before a much higher score. Then again, they could have had heatsinks and all on the ram. I think that's my most limiting factor. I could only overclock the ram by a mere 20MHz without introducing artifacts or even instability and my case is well cooled...

Also, I kind of wondered because I swear I saw a post once that said they had around 14K with a GF4 Ti4200 in 3dMark2001...

Anyway, I think it would be nice to have one of those 3dMark score topics stickied somewhere. Gives people a nice reference, plus some get bragging rights I guess d-:
 
I used to have 14K+ with my old ti4200. I had the Tt GF4 cool kit for it and it was overclocked to 300/600 +

I miss it already, I think I just wasted my money on this 9600pro. Too bad I nerfed the GF4 taking the HSF off of it.

:(
 
I'm afraid of the same thing. I thought I was pretty much upgrading, but I'm pretty dissapointed so far. Looks like it can do some decent overclocking though if you get good stuff for it. Someone in another thread said they saw 8K+ in 3dMark2003 after doing such things. I'm hoping I can at least get past my current crappy scores and get games like Deus Ex: IW running decently.

Oh, on the upside, turn up that FSAA and AF with performance mode. I'm running at 4x FSAA, 8x AF without a very noticable performance hit. That GF4 Ti4200 couldn't do that... I suppose it must use the more powerful pixel shader engine to do the FSAA or something.

EDIT: Oh, and on mine I nerfed the memory chips I think d-: Hehe, does anyone have a Ti4200 WITHOUT overclocking I wonder...
 
I just ran 3dmark03 on mine and it was like pulling teeth to watch it. 1680 for the score. phtpphtphtphtpht piece of crap.
 
1680? Something is wrong... Firstly, make sure FSAA and AF are on application preference and set the preferences as low as they go, that helps a lot since 3dMark2003 just requires every single ounce of power it can get out of a card like this and even though ATI has pretty good AF and FSAA, it doesn't help benchmarks any to use that. Also, be sure to overclock it a bit. I got my VPU up as high as 521 or so, but the memory would only go up to 317 and remain stable. You can see my score, so I really doubt yours should be so noticably below it.

Oh, and pulling teeth? Well, it doesn't get too much better even when you get this much. Actually, it's kind of funny watching the nature one. When it focuses on the water, things are just flying, then it starts to pan upward and you see all the grass and trees and the FPS takes a screaming nosedive. I admit to being impressed at ATI's pixel shader capabilities...
 
Yah, I got the slider all the way over to the performance side. I think it might have something to do with the fact that my fast write setting always shows off, even if I set it to on and then reboot. after the reboot its back on off. (its enabled in BIOS) I dunno, I'm pretty disappointed. I used that overclocker utility that I downloaded from another thread here and it found my OC max to be 457/371 and it still ran just fine (stock fan on this card is pretty beefy compaired to the made by ATI card, good thing I didnt get that ICEBERQ4 I was thinking about), but when I close it, it craps out and dies on my win2000 and I am not sure the settings are saved like that as the 3dmark01 benchmark is just the same. *sigh*

I dunno if I should jst get another ti4200 off of ebay and see if I cant get that back up to 14000+ 3dmarks.
 
Try the radlinker program I guess. It lets you set a bootable default clock rate.

Anyway, I had a really low rating about what you just mentioned in 3DMark2003 at first, but after a little playing around with settings, I got it to work a lot better. Oh, and I set mipmap and texture quality one down from the highest quality setting. That makes a difference too I suppose. I would say to check vsync, but if yours is like mine, I doubt you see high enough framerates for vsync to interfere.

BTW, you keep mentioning 14K, remember that 3DMark2003 and 2001 use completely different scoring systems. So the GF4 Ti4200 doesn't get a horribly different rating in 2003 than what I'm seeing I guess. It was a while ago when I tried my Ti4200 in the 2003 edition, but I think it was roughly around this same area.

If you want, I can dig through all of my settings and give you a list of everything so you have something to try.

EDIT: I guess further responses along this should be put in the topic you started about it.
 
That score doesn't seem unreasonable for a 9600 Pro and your CPU.

On a side note, one thing that I'm sure some of the top scorers do (and I do as well) (not that I'm a top scorer) is shut down pretty much everything else on the PC before the benchmark. All tray apps, all unneccessary services (for 2K and XP). I even have two batch files for stopping and starting services for this very purpose. Although this does net some improvement, I have found that above all it seems to give more consistent results.
 
How dependent is 2001se on processor speed

I upgraded from a 2100+ to a 3200+ (400MHzfsb) with 512MB PC3200 RAM and only got a 2000 point increase with a 9600XT

haven't tried tweaking anything yet, but I don't know how much of an increase I was expecting.
 
3DMark2001SE's low quality tests are based a lot on the processor speed. So that is where the 2000 points comes from. That is quite an increase in performance. I saw about the same from my stock 2500+ to the OC of 2.2 GHz.

And in response to the original post. Like I said, the low quality tests are based more on processor speed, so your 10.5K is a good score for that processor and the 9600 Pro.

And yes, shutting down any errroneus programs that are running will help your score probably about 250-750 points higher depending how much stuff is running.

I am getting about 12000 on my set up.
 
Well on my setup ...minus the dual channel 1gig ram...(im currently running single channel 512mb) I will soon run dual channel 1 gig. I can easily reach 15k on 2001se and 5-6k on 2003.

Your score seems just about right for that setup. Don't worry about it.
 
I have:

Abit IS7 MB
P4 2.4 @ 150mhz bus = 2.7X
512 megs generic pc2100
GeForce XFX 5900 video card

I get 14,000 in 2001se and 5500 in 2003. I can never enable the sound tests on 2003, i dont know if that is a factor or not.
 
You must log in or register to reply here.
Back
Top