gainward geforce 4 + 3dmark = why so low?

[Menardsguy]

I have the gainward geforce 4 ti4600 golden sample. Right now I am using the default heatsink (ordered waterblock) and speeds of 325/740. With 3dmark the highest score I can get is just under 10,000. Why is my score so low. With these tests I was running my cpu at about 1656.

score settings.
no aniti-aliasing (in display properties)
1024X768 @32 (in display properties)
mipmap @ best performance (in display properties)
all settings were at default in 3dmark

 

[MadMan007]

Two words: memory bandwidth. I assume this is the rig in your sig we are talking about, and it has SDR memory. I am sure it is a great computer, but 3DMark is *very* sensitive to memory bandwidth. Dont stress about the scores, if you can run games fast and with the options turned on, that is all that matters. And with 128MB onboard memory, I doubt your card will be accessing the AGP memory very much (where the mem bandwidth does matter), except in benchmarks that do that on purpose.

 

[CSaddict]

I have to agree with Mad. Its the memory holding you back.

 

[NASsoccer]

Yes, you SDR ram is dratically holding you back. Also, is there any possible way to get your FSB up? This has a very drastic affect on scores also.

I have ddr, you could send it my way ...j/k

 

[Silver]

Yes, you SDR ram is dratically holding you back. Also, is there any possible way to get your FSB up? This has a very drastic affect on scores also.

I have ddr, you could send it my way ...j/k

Darn, missed first dibs.

No, seriously I would hold off on the water block till you get the other issues resolved. You really are not going to want to play with water cooling issues at the same time that you are dealing with system/hardware issues. It won't hurt the block to sit on the shelf for a few days. Now, go out and get a epox 8k3a and some crucial (even a 128 stick if you are tight on funds). put it back together and tweak it some using the same program as well as the ram and vid. Don't tweak the vcore hard as that is what you have the water block for and it comes after you get things squared away. Don't go too hard on the ram either. I would geuss that right off this will take you to about 11,000 with good aggresive system ram settings. Now install the waterblock after you are sure all is well and I would look for something close to 11,800 maybe breaking 12,000. Sound good enough to you? You can shoot for 13,000 but you are going to get into some really radical (read risky ventures). What I mentioned above, done right, should put you in the top 8 to 10% range at Madonion. You are probably going to have to get the fsb up around 160 to 165. The 8k3a can easily handle this, keeping the T-bird cool is going to be a chore. Might consider water on the T-bird and ov to 1.7 or so at 170 fsb. Should ring up the numbers that you are looking for. An alternative would be to stay with air and get a $125 xp1900 on an 8k3a and again the numbers should ring up oc'ed.

 

[RangerJoe]

i dont see why everything is so focused on your 3dmark scores....does it make you think you have a bigger peni$ is you have over 10,000 points in a benchmarking program? the way i measure the quality of my video cards is by how well it runs in any game, or just in regular windows, or when im watching a movie....3dmark isnt very important....my voodoo3 3500 got crappy 3dmark scores, but ran the games i played fine....and i got it for free

 

[Silver]

...does it make you think you have a bigger peni$ is you have over 10,000 points in a benchmarking program?

It would seem that you could have made your point without that. Why would one need 2.0 Ghz? Not required for most games I'm aware of. This is www.overclockers.com correct? Just making sure I am at the right site.

 

[RangerJoe]

im not talking about cpu speed, im talking how people are judged here by the score they get on 3dmark 2001....having a 2 ghz computer is important, but measuring a card by how many 3dmarks it gets, is just plain dumb....why cant a card be measured here by the framerates it gets in games, cause that actually effects something....you cant participate in 3dmark, its repetitive...and boring to watch...and whether or not you get a 1, 2, 3, 4, 5, 6, or whatever digit score, it really doesnt matter...at all...framerates, and quality are the reasons gaming cards are made, not to show how many points you can get on 3dmark

 

[NASsoccer]

umm.... the higher my 3dmark score the higher my FPS. That is how it always goes for me.

 

[repo man11]

I like running 3d mark to see if new drivers or other tweaks had a positive or negative effect on my computer. It takes a lot less time than trying to see if I can see the difference in a game. But I agree that for some people it seems more important than anything. I you get a big thrill out of benchmark racing, have at it. But like most races, it goes to those with the most cash.

 

[Silver]

im not talking about cpu speed, im talking how people are judged here by the score they get on 3dmark 2001....having a 2 ghz computer is important, but measuring a card by how many 3dmarks it gets, is just plain dumb....why cant a card be measured here by the framerates it gets in games, cause that actually effects something....you cant participate in 3dmark, its repetitive...and boring to watch...and whether or not you get a 1, 2, 3, 4, 5, 6, or whatever digit score, it really doesnt matter...at all...framerates, and quality are the reasons gaming cards are made, not to show how many points you can get on 3dmark

Good response....it has gotten to be more..there is a competition amongst many different sites in an attempt to show who (what site, personnel bragging rights) knows more about their systems, etc...one way to prove this is to have a neutral test for all and 3dmark2001se is it. There are many, many involved in this and you will find some of the fastest, processors, ram and general systems posting there. It's competitive and definately a male thing. I have picked up a torch of sorts and though like you I am not terribly impressed, it does get addictive after awhile and better still, my computer has never run so fast in all areas. Heck, bios is tweaked, proc ov, ram ov, vcore ov, vmem ov, operating system tweaked, etc...it was not my intent but it is really running well in all tasks.

 

[Silver]

I like running 3d mark to see if new drivers or other tweaks had a positive or negative effect on my computer. It takes a lot less time than trying to see if I can see the difference in a game. But I agree that for some people it seems more important than anything. I you get a big thrill out of benchmark racing, have at it. But like most races, it goes to those with the most cash.

Definately ain't this guy...make it or make do. That pretty much covers it. Thus a ti4600 is out of the question as is the xp2100. The xp1600 has to run at xp2100 speed and the ti200 is trying to approach ti4600 speed. So, I have to make it or make do. Alot of fun though and cheaper than drinking, chasing girls (wife is really down on that one) and a number of other so called hobbies. To get out of the house, I go Striper fishing....from shore.

 

[repo man11]

Definitely not me either. My best score is 4,880 I think. And I've had to back down my overclock because it is warming up! I'd like to buy an AX-7, but I'm unemployed right now, so unless I get on the stick, I won't even be able to keep my dial up connection! No goodies for the 'puter for the foreseeable future. (heavy sigh)

 

[RangerJoe]

i am mainly talking about Menardsguy, and people just like him, they buy a $300 + card, and then, they decide that the card doesnt live up to what other people get on a benchmarking program, so they are not happy....i will admit, i use 3dmark 2001 SOMETIMES, but i dont use it excessively, i used it when i got my new computer, and when i got my new video card, to see what the difference was, but i dont tweak my system to get a higher score....my system is PLENTY fast for anything i ever need to do...runs all the games i play fast enough....i need more ram just because 256 isnt enough to play alot of games....but other than that, my system is great, i dont need to spend $300+ on a video card just because its the newest thing out....id rather get the 2nd to best model, because it will do anything the gf4 can do, and will run it plenty fast....there is just no use for 128 megs of ram yet...AT ALL...i just got back to the computer, and i lost my train of thought...so ill finish this later

 

[MadMan007]

i dont see why everything is so focused on your 3dmark scores....does it make you think you have a bigger peni$ is you have over 10,000 points in a benchmarking program? the way i measure the quality of my video cards is by how well it runs in any game, or just in regular windows, or when im watching a movie....3dmark isnt very important....my voodoo3 3500 got crappy 3dmark scores, but ran the games i played fine....and i got it for free

E_X_A_C_T_L_Y

It is a good yardstick, but I just don't understand people who tweak their sstem *purely* for 3DMark. Oh well, to each his own...

 

[Ugmore Baggage]

I think you're missing the point. Menardsguy seems to be saying, "Is there something I should be doing differently to make my expensive card worth the money, or is 3DMark inaccurate?"

I don't see any elitism at all in Menardguy's post. It's a humble request for information. Are you jealous RangerJoe? There will always be people with more or less worldly possessions than each other. Differences don't make people better or worse unless they consider it so (those are the worse ones). Sure, Menardsguy may have "too much" money but he still has a right to our help and respect.

My take on the matter at hand: the 4600 isn't worth the extra money except for a very few professionals, or people for who don't need personal budgets (the very rich). Tom's Hardware has an article about the 4200 that benchmarks several games and 3DMark. The 3DMark scores are the least impressive of the tests they did.

 

[RangerJoe]

hahaha, yea im jealous, thats it....my gf3 ti200 runs wonderfully, never had a problem, and paid $160 for it....my point of selecting menardsguy for my example is he came on here wondering if anything was wrong because of the low scores on 3dmark...i wouldnt have said ANYTHING if he came on here wondering why he was getting 25 fps in half life (you get the point) he selected 3dmark as his question base, simply because he knows that many of the people around here are sooo focused on if a card is good or bad based on its scores in 3dmark....as i have said before, 3dmark doesnt matter, but fps is games does

 

[deez]

10,000 is not a low score on 3DMark and every game should run perfectly smooth. Try running it with a TNT2 or even a Geforce2 and see the difference.

 

[RoskO (_!_)]

and...

3dmark does test the FPS, among a lot of other things. So what the hell do you think 3dmark does? The point of the program is to test your video card/system and give you a score to compare to others, and see how well your system performs compared to others with similar...or maybe not so similar setups. Just because my system runs good doesn't mean it's configured correctly, and maybe that is what he was asking. Why are his scores lower than others with his card? He just wanted to know if maybe he wasn't doing something right.....wrong drivers...maybe a bios update someone knew of......(get the point). Just because a system runs great doesn't mean it's performing how it could....or should. If I set my FSB to 100mhz on my athlon and ran games it would still kick ***.....but does it kick as much *** as if it were 133mhz? no. Not trying to be a jerk, just trying to shed some light.

Regards.

Jeremy

 

[Silver]

I have the gainward geforce 4 ti4600 golden sample. Right now I am using the default heatsink (ordered waterblock) and speeds of 325/740. With 3dmark the highest score I can get is just under 10,000. Why is my score so low. With these tests I was running my cpu at about 1656.

score settings.
no aniti-aliasing (in display properties)
1024X768 @32 (in display properties)
mipmap @ best performance (in display properties)
all settings were at default in 3dmark

Anyway, back to the thread starter. Menardsguy, you have several options to attain your goal. Pick an avenue and many here will most likely jump in and help you to achieve it.

Guys, we did kind of kill his thread wherein he asked for assistance in achieving better performance as measured by 3dmark2001. Much of the input was of no assistance in helping him to achieve it.