Running 3DMark03 and trying to overclock my graphics card

SquirrelMan · 2004-01-19, 05:24 PM

Okey, I dont have 01 so it says it can only run one of the game tests. Tell me if I need to dowload 01 for my results to mean anything:

Nonchanged: 193 3Dmarks

Aquamark:
Nonchanged (default settings): 8171
10 MHz upgrade to both: 8167
20 MHz: 8161

that isn't good, is it. Well, i guess my graphics card is outdated then. I'll have to get a new one...

Rbstr · 2004-01-19, 05:40 PM

ok did you read tutorials on OCing at that place right? be carefull and conservative at first. A good rule of thumb is up it 5mhz untill it starts to crash or have artifacts then lower it 10.

SquirrelMan · 2004-01-19, 05:45 PM

yeah, i read the tutorial. I havent dowloaded the overclocking program yet, i'm getting AquaMark 3 to have two different "opinions" on my graphics card

Squick · 2004-01-19, 08:14 PM

Like I said, ditch 3dMark 2003, unless you think that a low end ATI card is actually 20 times faster then your nvidia card. Go to Tomshardware.com and look back a year or so about the 3dmark 2003 debacle if you wana learn more.

Rbstr · 2004-01-19, 08:59 PM

yeah 3dmark03 was in alot of hubub, mainly becasue they said Nvidia was cheating(as they were) and Nvidia said the code was unfair(and it kinda was but not realy), but the new release basicaly shows the results that one would expect bucause the truth is Nvidia's card suck at DX9.

Squick · 2004-01-20, 04:19 PM

Brand new article on the applicability of 3dMark:
http://www.overclockercafe.com/Futuremark.htm

And like I have said in the past Rbstr, no game will use those shaders that nvidia's cards suck at, they are extremely inefficient and would render a game unusable for nearly everyone. Very similar effects can be created using similar texture shaders, which both nvidia and ati could render ten times faster. To the best of my knowledge no one can do single pass pixel shaders yet.

So until a card comes out that can process 78.6 million PS2.0 pixel shaders per second, any test that tests pixel shaders is irrelevant.

In layman's terms, it would be like Car and Driver testing the 0-60 times of a car with the hood open, so that the engine stays cool; even though it adds an enormous amount of drag on the car.

Squick

Rbstr · 2004-01-20, 04:29 PM

This isn't the point of the thread i'm sorry squick, its about a benchmark to be used for testing of overclocking, and 3dmark is as good as any for that, because it is so demanding on a chip, he needs something to test the stablity and to make sure he is getting some results.

Squick · 2004-01-20, 05:17 PM

Actually he is using his score on 3dMark2003 to compare with other systems. He is likely to look at other sites and say, "Hmm... I see my 3dMark2003 score is 193, I see here that this ATI Radeon 8000 gets 2,000. That must mean that the Radeon 9000 is going to make Planetside 20 times faster then my Geforce4 Ti4600!"

The score generated from 3dMark2003 should be considered completely arbitrary, I do not even think you could use it to compute your performance gain due to overclocking since a flawed test could not be relied on to be consistent.

And if you are curious why I am pissed at 3dMark, it is because the second day 3dMark 2003 came out I used the results of 3dMark2003 to justify dumping my nVidia Geforce4 Ti4600 and getting a Radeon 9200. EVERY game I played was drastically slower. I now use the Radeon in my server, a vesa local bus video card would sufice for my server if it had the port.

Squick