Low 3dMark scores for p4 2.8 and oc'ed gf4 ti4200

[DubSlick]

System as follows:

p4 1.8A oc'ed to 2.8 Ghz (156 FSB)
512 MB of DDR 333 running @ DDR 390 (195 Mhz)
Leadtek GF4 ti 4200 set at 300/560 using Detonator 40.72 Drivers
Abit BG7 motherboard

When I first installed the board, I got 13000 3dMarks with this setup. Getting greedy, I flashed the BIOS for my motherboard to get a usable 3:4 memory divider, but had driver issues, so I reinstalled win2k. Newest drivers for everything.

Now, when I run the test on this clean setup, I'm only getting 10200 3dMarks. What gives? There was an issues of Direct3D being unable to create 3d Devices, so I had to change PCI Texture Memory Size from 64 MB to 2MB in order to get games to run. Would this cause the drastic drop in scores? Please help

 

[Deeko]

I don't care what your 3DMark score is. I should kick you for asking. What are your game scores? Don't come back here until you know.

 

[Uclagamer_99]

pci anything shouldn't matter if you have an agp card

 

[tenoc]

Did you re-install DirectX 8.1b after you re-installed the OS?

 

[DubSlick]

(Ducking kick from Deeko) Okay, ran UT2003 Demo benchmark, upgraded to DX 8.1b and came up with the following results:

1024x768
Flyby: 161.9
Bot: 69.4

1280x960
Flyby: 115.9
Bot: 63.5

These seem to be in line with what others are getting. Still getting 10200 3dMarks, but oh well. Worthless benchmark?

 

[Deeko]

Yes, 3DMark is extremely worthless....I'm glad to see my relentless bashing is paying off

 

[merlocka]

3dmark2001 is nearly useless these days.

It still has value as a benchmark, if you are comparing changes in the same system (new drivers, overclocking, faster timings, etc)

Unfortunatly it has become a online "bragging rights" tool, an as a result the online database is filled with scores that are unrealistic.

Overclocking past the point of stability, decreasing graphical and texture detail, and driver "tweaks" have resulted in a collection of scores which don't reflect the performance of real computers.

And everyone does it. Everyone here will overclock that extra 5MHz to get a higher 3dmark score. I hate 3dmark yet ironically I do the same dasm thing. I can't explain why.

If you want to get "cured" of the 3dmark syndrome, go hang out in the madonion forums for a day. It's horrible.

 

[Deeko]

Originally posted by: merlocka
3dmark2001 is nearly useless these days.

It still has value as a benchmark, if you are comparing changes in the same system (new drivers, overclocking, faster timings, etc)

Unfortunatly it has become a online "bragging rights" tool, an as a result the online database is filled with scores that are unrealistic.

Overclocking past the point of stability, decreasing graphical and texture detail, and driver "tweaks" have resulted in a collection of scores which don't reflect the performance of real computers.

And everyone does it. Everyone here will overclock that extra 5MHz to get a higher 3dmark score. I hate 3dmark yet ironically I do the same dasm thing. I can't explain why.

If you want to get "cured" of the 3dmark syndrome, go hang out in the madonion forums for a day. It's horrible.

Not true...I don't even look at my score. The only value 3DMark has is as a stress test...let it loop a few times to see if you're stable. That's all.

 

[merlocka]

Originally posted by: Deeko
Not true...I don't even look at my score. The only value 3DMark has is as a stress test...let it loop a few times to see if you're stable. That's all.

Not true

It is a valid benchmark, the engines used are from real games. It's not a very realistic benchmark but it is as good as any for comparisons of equalized configurations. I'd much rather see less synthetic benchmarks used in online reviews by websites, considering that it's pretty well known that vendors optimize their drivers for 3dmark scores.

Albeit I might have generalized with my statement of everyone here trying to get the highest score. I'll correct that by saying 90%

 

[Deeko]

It doesn't matter if the engine is used for games, if a setup that gets great game scores has a crappy 3DMark, why should people get worked up about?

 

[merlocka]

Originally posted by: Deeko
It doesn't matter if the engine is used for games, if a setup that gets great game scores has a crappy 3DMark, why should people get worked up about?

Perhaps because it's a potential symptom of a configuration issue on their computer? It's not like 3dmark doesn't correlate *at all* with real game performance. It's just annoying that so many people think it's the primary barometer of how fast their computer is.