3DMark on 8800GT under performing?

V

visuar

n00b
Joined
Mar 19, 2008
Messages
13
I'm trying to build a nice gaming system with plenty of storage, but the gaming part feels way too slow. I've been running 3DMark and I think the results I'm getting are not as they should be. I've searched around but couldn't find what kind of scores I should be getting with a 8800 GT (for the graphics tests). Some notes:

1) I'm currently running a Celeron cpu while I wait for my E8400 to show up. This apparently has a FSB of 133 MHz which no doubt is hurting things. I don't really care for the CPU scores in 3DMark. Am I correct in assuming the graphics tests shouldn't really be hurt due to this CPU?

2) I've run tests under Windows 2008 Server 64-bit, Vista Home Premium (both with and without SP1 RTM) 64-bit, XP Professional 64-bit and XP Professional 32-bit. All test results were virtually the same, results below are from running 3DMark on XP Prof 32-bit

3) I've run the tests with all the extra hardware in but also with none of the additional cards and only the main OS drive. No difference. Also tried disabling the legacy IDE (which the burner is connected to) and disabling the USB controller that has a shared IRQ with the graphics card, no difference either.

4) All Windows updates have been applied, running with the latest driver versions. Tried both a final and beta driver from NVidia as well as one from EVGA. No difference in scores. Results below are with the 169.21 driver.

Main hardware:
- Antec earthwatts EA500 ATX12V v2.0 500W PSU
- Gigabyte GA-P35-DS3L LGA 775 Intel P35 / ICH9 ATX Motherboard
- Intel Celeron D 356 3.33GHz LGA 775 (FSB = 133 MHz??)
- 2x G.SKILL 4GB (2 x 2GB) 240-Pin DDR2 800 SDRAM (PC2 6400) 5-5-5-15, so 8GB total
- EVGA 512-P3-N801-AR GeForce 8800GT 512MB 256-bit GDDR3 PCI Express 2.0 x16
- Arctic Cooling Freezer 7 Pro 92mm CPU Cooler

Additional devices:
* 1x Samsung Spinpoint T Series HD501LJ 500 GB SATA 3.0 Gb/s HDD
* 6x Samsung Spinpoint F1 HD753LJ 750 GB SATA 3.0 Gb/s HDD
* Samsung SH-S202J 20x IDE DVD burner
* Promise SATA300 TX4 PCI SATA II controller
* Intel EXPI9300PT 10/100/1000Mbps PCI-Express Gigabit NIC
* MSI PC60G 32bit PCI2.2 Turbo G Wlan adapter

3DMark06 Revision 1 Build 0 (free version, can only run standard 1280 x 1024 test):

3DMark Score = 5020 3DMarks
SM 2.0 Score = 2220
SM 3.0 Score = 3119
CPU Score = 835

[Graphics Tests]
1 - Return to Proxycon = 18.67 FPS
2 - Firefly Forest = 18.33 FPS

[CPU Tests]
CPU1 - Red Valley = 0.27 FPS
CPU2 - Red Valley = 0.42 FPS

[HDR Tests]
1 - Canyon Flight (SM 3.0) = 36.39 FPS
2 - Deep Freeze (SM 3.0) = 25.99 FPS

As said above I'm fine with the CPU scores since I'm waiting on the better one to arrive. The graphics and HDR test results is what is worrying me...

How do these results look to you all? If something is wrong what do you think it might be? Graphics card? Mainboard?

I've tried to be thorough with this post (I know everyone's time is valuable). If you do need more info let me know and I'll add it. Thanks a lot for any help!

Best,

Rob
 
That processor is holding back that video card big time, upgrade it asap. (EDIT, see you getting a E8400! whew)

Other then that, could be some video card driver problem (lol don't ask me, make sure you completely remove the old drivers).

CPU effects not just the video card, but a large portion of 3dmarks is made by the CPU.
 
3dmark sucks because CPU is weighted way to heavily in the GPU tests, hence whey all the top scores are run using heavily overclocked CPUs, even though that won't do crap in games.

Such a waste of a program.
 
Thanks for the quick replies! I'll sleep a bit better until that new cpu comes in. I guess I was expecting it to influence the 3DMarks, but not the actual FPS of the tests. I guess they must tax the cpu quite heavily during the graphics tests as well then?
 
It's the Celeron D obviously. There's a reason there is so much vitriol against 3Dmark06. It places like 33% of it's score on the CPU. If your CPU sucks, your 3DMark score sucks whereas most games are more GPU reliant.
 
My E6750 holds my score back with my 8800GT SSC, so I'd say your CPU is REALLY holding you back.
 
I guess you learn something new every day :) Is there a better program to use to make sure the graphics card is working as it should? (performance wise etc.)
 
visuar said:
I guess you learn something new every day :) Is there a better program to use to make sure the graphics card is working as it should? (performance wise etc.)
Click to expand...

They're called games. I play them sometimes.

Honestly, unless it seems like something is wrong... chances are, nothing's wrong. Funny how that works.
 
Yeah, sounds like it should all be good.... time to finish the machine and have some fun!
 
Your cpu score is way low.. expect your 3dmark score to more then double after installing the E8400.. especially after an OC.
 
yamahaSHO said:
My E6750 holds my score back with my 8800GT SSC, so I'd say your CPU is REALLY holding you back.
Click to expand...

Is this true? I have a 6750 also oc'ed to 3.45, and I have a 8800 GTS, how do you know that the cpu is holding your back?

Is it because quads score better or the 6750 speed too low? comparatively is the 3dmark score inflated by the cpu score or is the gpu too much for the 6750.

Let me know really curious,

thanks
 
Ndel said:
Is this true? I have a 6750 also oc'ed to 3.45, and I have a 8800 GTS, how do you know that the cpu is holding your back?

Is it because quads score better or the 6750 speed too low? comparatively is the 3dmark score inflated by the cpu score or is the gpu too much for the 6750.

Let me know really curious,

thanks
Click to expand...

The faster the CPU, the more inflated your 3DMark06 score becomes.

Real world gaming: Going from 2.4GHz Dual Core to 3.4GHz Quad Core does NOTHING for a single 8800 GTX at high resolutions. Literally less than 0.1% framerate increase. (A few exceptions for older games which have become CPU-bound as their frame rates have topped 100, and CPU bound simulators/strategy games.)

But in 3DMark06, from my understanding, your CPU score will increase linearly forever, with no penalty for extremely high clocks that will do nothing in gaming. If they wanted to be honest they'd cap the CPU score or implement some serious diminishing returns on CPU score. But then that would alienate those "extreme benchmark" kids with their LN2.
 
Ndel said:
Is this true? I have a 6750 also oc'ed to 3.45, and I have a 8800 GTS, how do you know that the cpu is holding your back?

Is it because quads score better or the 6750 speed too low? comparatively is the 3dmark score inflated by the cpu score or is the gpu too much for the 6750.

Let me know really curious,

thanks
Click to expand...



I know because as I overclock it higher (currently at 3.1Ghz) the score goes up accordingly.
 
NKDietrich said:
The faster the CPU, the more inflated your 3DMark06 score becomes.

Real world gaming: Going from 2.4GHz Dual Core to 3.4GHz Quad Core does NOTHING for a single 8800 GTX at high resolutions. Literally less than 0.1% framerate increase. (A few exceptions for older games which have become CPU-bound as their frame rates have topped 100, and CPU bound simulators/strategy games.)

But in 3DMark06, from my understanding, your CPU score will increase linearly forever, with no penalty for extremely high clocks that will do nothing in gaming. If they wanted to be honest they'd cap the CPU score or implement some serious diminishing returns on CPU score. But then that would alienate those "extreme benchmark" kids with their LN2.
Click to expand...

Agreed... The highest scores I've seen consist of a failry normal - higher end video card (I've even seen a X2XXX) with an extremely overclocked CPU (mainly quad cores, which most games do not benefit from).
 
You know a portion of the 3dmark score is based on your CPU. That doesn't mean your CPU is bottlenecking your video card... and it isnt.
 
colinstu said:
You know a portion of the 3dmark score is based on your CPU. That doesn't mean your CPU is bottlenecking your video card... and it isnt.
Click to expand...

Thanks for the clarification, that's what I thought that higher frequencie cpu's score better because it obtains a better cpu score overall, not that there is bottlenecking. Although I was worried and about to OC some more if that were the case.

All my games run pretty good for now, no need to OC some more, although I havent tried Crysis yet. Honestly im waiting for cards to improve to make it playable at 1920x1200 at high settings before I make the purchase.
 
colinstu said:
You know a portion of the 3dmark score is based on your CPU. That doesn't mean your CPU is bottlenecking your video card... and it isnt.
Click to expand...

I just said it was holding my score back, not my video card. Either way, the overall score relies too much on the CPU considering the graphics card does most of the work in games.
 
You must log in or register to reply here.
Back
Top