Fruitless 1800XP upgrade?

[FooDog]

I recently upgraded a rig and the results were not exactly what I expected. After having swithed to an 1800XP and 256MB of DDR RAM, the systems 3Dmark score only jumped by a measly 100 points. Can somebody explain this to me? I know 3Dmark is more dependent on the video card than anything else, but I thought the upgrade would at least have some tangible impact on performance.

Here the old rig's spec's:
Abit BX6 rev 1.0 w/ Asus SlotKet
Celeron 900 @1009
256 MB Kingston PC133 SDRAM
Veisiontek Xstacy 5632 GF2 GTS-V 32 MB DDR RAM
WIN 98 SE


Here's what's new:
Gigabyte GA-7DX rev 4.3
Athlon 1800XP no OC yet [need a better HSF...]
256MB Crucial PC2100 DDR RAM

 

[pressureEze]

You should have noticed an enormous difference there. I went from a tbird 1400 on sdram to a xp 1600+ on ddr ram and went from 6600 to 8600. What scores are you getting with the new setup?

 

[BuckNaked]

I have to agree, you should see a much larger increase than that...

Dave

 

[flexy]

an xp1800 with an decent 266fsb/ ddr ram mobo is about THREE times as fast as an athlon 900mhz on an 133FSB board.

(See SETI....like 10 hours at my machine at work.....3.40 hours on my machine at home [xp1800])

And you had an celeron !

Something must be terrible wrong with your setup...bios settings....whatever...


the xp1800 PC (especially if you manage to OC your FSB to 144/145 with the crucial ram) should be an incredible fast machine !

OUPS....maybe yiur gfx card is a bottleneck ????

I'd test each component (Sisoft Sandra 2002, eg. memory, processor etc.....and THEN look at 3dmark for the graphics performance !)

 

[Valinos]

My advice is not to worry about it.

3dmark2k1, Sandra, all these bench utilities are useless. It is all theoretical, and they aren't always accurate. I've seen 3dmark give out some really retarded scores with extremely good hardware, and I just conclude it is just plain wrong. The minute I try to play a game such as Wolfenstein on the same system it runs great at 1024, with all effects on max, at a steady 70-100 frames. At the same time, 3dmark2k1 gives my system 3000 3dmarks. I don't care, because it isn't real. It is theoretical. I have the same exact setup as many other people and they get 5000 3dmarks. I think its useless.

If it works for you in your games and you feel the extra performance, then don't worry about it. Just go by how your games play, not what some theoretical benchmarking utility says you have.

 

[Kerle]

That's basically right - worry about your real game/windows performance, not some benchmark. If the games are running slow, though, there are some things to check.

Make sure that your FSB and memory settings are correct, and that your processor is at the right multiplier. Start with that, though there are certainly more things to go through.

 

[pressureEze]

<< My advice is not to worry about it.

3dmark2k1, Sandra, all these bench utilities are useless. It is all theoretical, and they aren't always accurate. I've seen 3dmark give out some really retarded scores with extremely good hardware, and I just conclude it is just plain wrong. The minute I try to play a game such as Wolfenstein on the same system it runs great at 1024, with all effects on max, at a steady 70-100 frames. At the same time, 3dmark2k1 gives my system 3000 3dmarks. I don't care, because it isn't real. It is theoretical. I have the same exact setup as many other people and they get 5000 3dmarks. I think its useless.

If it works for you in your games and you feel the extra performance, then don't worry about it. Just go by how your games play, not what some theoretical benchmarking utility says you have. >>



Sorry, I gotta disagree. Benchmarks are a very useful way for us to measure improvements that might otherwise be too small to see just by running a game, etc. For instance, if I set up a new system, I might use something like 3dmark 2001 to determine at what point my ram timings, FSB, CPU and video overclocks all come together to yield the best results. I'll usually make small changes here and there in an attempt to squeeze as much performance out of my system as possible. While I might not notice an extra 2fps while I'm gaming, the benchmarks let me know I'm either making progress or going backwards. Individually the changes might not amount to much, but overall they can make a huge difference. [sp]

 

[formulav8]

His video card is a Visiontek GF2 GTS-V 32 MB DDR. Thats the card that has the slow 7ns ram. Only running at like 286 mhz. His video card it definitely the bottle neck. You overclock your video card and you will see improvements in your 3dmark score. That video card is barely better then a gefore2 mx 400. Overclock that card or get a better one.


Jason

 

[Budman]

<< Fruitless 1800XP upgrade? >>



Fruitless,do you mean to tell me that all these years when I upgraded I was supposed to get fruit with my hardware?:|

I must talk to my supplier about these missing fruits.

 

[Agamar]

Yep...It is the GF2 GTS-V...I have one with a 1700+ and 256 Crucial DDR...My 3Dmark scores did get a lot better than my P-III 667 with 512 PC133 on a Be-6....I had 2411 on 3DMark2001...went to 3381 with the XP and Soyo Dragon +.... 970 points upgrade. I am overclocking my card to 180 / 300 now though...I donno what difference it makes...but makes me feel better

 

[Budman]

FooDog,

Since you have an AMD Chipset did you install the AMD AGP driver ?





AMD AGP Miniport Driver

 

[ScrapSilicon]

<< Yep...It is the GF2 GTS-V...I have one with a 1700+ and 256 Crucial DDR...My 3Dmark scores did get a lot better than my P-III 667 with 512 PC133 on a Be-6....I had 2411 on 3DMark2001...went to 3381 with the XP and Soyo Dragon +.... 970 points upgrade. I am overclocking my card to 180 / 300 now though...I donno what difference it makes...but makes me feel better >>

hehe

 

[mschell]

The video card itself is not the problem. I was running a GeForce 2 MX/ Intel I815, 825MHz Coppermine and got 1200 more points when upgraded to a 1.4GHz Athlon/Iwill KK266. The Athlon system ran dog slow without the 4 in 1 drivers installed however. Tweaking the AGP settings in the BIOS helped also. Be sure to run the 21.xx or later series of Nvidia drivers as the later ones are optimized for Athlon systems

 

[mithrandir2001]

FooDog, try this: overclock your GTS-V's memory from 286MHz to 333MHz and see what happens to your 3DM2K1 score. If it goes up by at least 5%, then you know your video card is the bottleneck (which I'm certain is the case).

 

[Maggotry]

<< My advice is not to worry about it.

3dmark2k1, Sandra, all these bench utilities are useless. It is all theoretical, and they aren't always accurate. I've seen 3dmark give out some really retarded scores with extremely good hardware, and I just conclude it is just plain wrong. The minute I try to play a game such as Wolfenstein on the same system it runs great at 1024, with all effects on max, at a steady 70-100 frames. At the same time, 3dmark2k1 gives my system 3000 3dmarks. I don't care, because it isn't real. It is theoretical. I have the same exact setup as many other people and they get 5000 3dmarks. I think its useless.

If it works for you in your games and you feel the extra performance, then don't worry about it. Just go by how your games play, not what some theoretical benchmarking utility says you have. >>



That's not exactly true. I too have seen some rediculous 3DMark scores, but benchmarks, by their nature, are not theoretical. Theoretical scores are when you look at system specs and say, "This is the score I think I should receive based on this information". 3DMark, SANDRA, HDTach, etc., test your system. There's nothing theoretical about it. The utilities run tests and display results based on testing.

I do agree that people put too much stock in 3DMark. If your game works for you, then enjoy it. But benchmarking is a useful tool to measure system performance.