Hi,

I've just bought a new computer with the following specifications:

AMD Athlon XP 3000+ (Barton) (10.5 x 200Mhz = 2,100Mhz)
MSI K7N2M (MS-6777) mainboard (nVIDIA nForce2 chipset)
512MB PC2x00 RAM (I'm not sure if it's 2100@133Mhz or 2700@166Mhz)
MSI 8912 256MB GeForce FX 5600 (350@450)
120GB 7200rpm hard drive
Windows XP Service Pack 1
DirectX 9.0b
nVIDIA Detonator FX drivers v45.23 WHQL

3DMark 2001 SE (Build 330) with my graphic drivers on default settings i.e. no antialiasing or filtering, only gives me a score of around 8.900! Judging by some of the other posts this is slow as I've read of someone with an Athlon XP 2200+ returning scores of 12,000+.

Which chipset drivers do I use? Do I need the VIA 4-in-1 drivers? I ask this because once when I booted up my machine in Safe Mode, I noticed many different AGP drivers being loaded such as VIAAGP.sys, SiSAGP.sys etc. Surely I don't have all those chipsets so why does Windows XP load all these drivers?

Also how can I find out exactly how fast my memory is without actually opening the case?

Does anyone know why DXDiag and 3DMark 2001 report my graphic card as having only 128MB RAM when RivaTuner and the Display control panel applet show the correct amount of 256MB? Is it a bug?

In PC Gaming Hardware I read that Windows XP always assumes a L2 cache size of 256KB but as my processor has a 512KB L2 cache it recommends that I set the DWORD entry SecondLevelDataCache to 512 decimal (200 hexadecimal) in registry key HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management. Is this correct as when I tried it and then ran the nVIDIA Last Chance Gas demo it generated a STOP error which never happened before. The L2 cache is enabled in the BIOS.

Sorry for all the questions but I'd be very grateful for any help.

hehe it can be very easy ...
first NEVER INSTALL 4in1 !!!! especially not an a NF2 motherboard ....

2 things

1. have you installed the nForce 2 driver ?
if not get it here http://www.nvidia.com/object/nforce_udp_winxp_2.45

2. goto http://www.3dcenter.de/atuner/ and get atuner ... look again if AA and AF are really disbled ... then most important look for the VSync setting ... Disable it !!

does this help ?

Thanks for the reply, Mertsch. I forgot to mention that I'd installed the v2.45 nForce drivers from the nVIDIA website. I checked on the MSI website and although their drivers appear to be a later version (2.78), the driver contents are older versions than nVIDIAs! Very confusing!

but its true .... 2.45 is the newest one ..... even the 3.44 is older ...
and does vsync help ???

I downloaded aTuner and ran it making sure I'd disabled all antialiasing, filtering AND vertical synchronisation but... alas... my score is still only 8,983! Pitiful, isn't it (or is it), for a supposedly fast machine?

I'm a little confused about the memory speeds on nVIDIA's GeForce FX cards in that am I right in assuming the core clock speed on my machine is 350Mhz but, although my memory clock speed according to the overclocking tab in the Display control panel applet is 425Mhz, this is effectively running at 850Mhz because it is DDR memory?

check the AGP texture size in BIOS .... set it to the max and try again

Hey have you enable AGP 8X? Fastwrite? and make sure that you maximize your ram speed (for example, t1 command, etc?).
if you maximized your RAM speed and GFX than you will get higher score.
Don't get too excited yet, the benchmarking program that you use is 3D marks 2001 that means it depends mostly on your GFX. Geforce FX 5600 perform not that good in 3D marks 2001 as i know compare with Geforce 4 ti 4200. I own a Radeon 9600pro with suppose to be faster than the Geforce 4 ti 4200 but it stuck with lower marks in 3D marks 2001 while faster in 3D marks 2003 (because of DX 9).

I have tested both radeon 9600pro and Geforce 4 ti 4200(4x) which return following results.

Geforce 4Ti 4200(4x)= 12xxx
Radeon 9600pro(8x)=10xxx

The spec is :
AMD 2400+
MSI KT4V
512 PC2100 Virtual channel
Sblive.
WD 30GB 7200 RPM.

So don't be too upset of the marks. I get better performance i some games compare with Geforce 4 ti 4200, 3D marks is not that important after all.

What you need is nforce driver (never try other SIS , VIA, ALI ,etc).
Detonator driver, or you can try the Omega detonator. I have try yhe Omega for my Radeon 9600pro and it is good. Customize the service in msconfig accoding to what you need and you don't.

lastly what do you mean by it display 128mb in DX? Is you Geforce FX 256Mb?

If you still encounter problems, try do a fresh install of windows XP and install all he latest drivers (not necessary the latest is the best, check the benchmarks on the net).

Good luck. :wink:

what you need to ask yourself is: "do my computer run good enough for what i need it to do?" it might have a problem with that single benchmark, and getting upset over your computers performance because a benchmark returns a number thats 'not big enough' is kinda stupid. check framerates i games and de some sisoft sandra tests. dont put all your satisfaction with your computer on stake with one benchmark.

btw sisoft sandra can also help you find where your system is running slow - if it really is.

I have the same processor as you, and close to the same mother board and memory. The only main difference is that I have a Radeon 9800 Pro.
I get a score of 17,078 in 3DMark 2001SE, and thats at stock settings,
nothing overclocked. I think you just need to tweak it a little. Try going into the bios and change the processor setting from manual to high performance.
___________________________

AMD Athlon XP 3000+
MSI K7N2 Delta-L (nForce2 Ultra)
512 X 2 DDR400
Radeon 9800 Pro (Catalyst 3.7)
Audigy Platinum
80 GB WD 7200rpm, 40 GB WD 7200rpm
Logitech Z-680 5.1 speakers
Win XP SP1

El_Coyote is right on the money. Benchmarks only demonstrate "relative" performance ... and sometimes not accurately at all. I have a buddy who has a dual AMD MP CPU system (Tyan) that in some benchmarks looks like a slug. However, in REAL-world applications it runs like a raped ape.

If you want a measure of true performance, there are a number of gaming benchmarks that are much more accurate. These benchmarks require that you have specific games already installed and they simply run scripts or invoke demo or pre-recorded matches, time how long it takes to run them and then displays the low, high and average frame rates.

There are also games that have built-in benchmarking functions like UT2K3 (Unreal Tournament 2003). A few more good examples are AquaMark (built around the AquaNox gaming engine), DroneZmark (built around the Dronez engine) and for Quake 3 there is Q3Bench. Other programs that invoke games you have installed on your machine are BenchEmAll and Seismovision 2.

Don't EVEN trust purely synthetic benchmarks. The folks who write the video drivers go out of their way to optimize their drivers for benchmarks just because the knuckle-heads who review video cards use those benchmarks in their reviews (rather than REAL games) to demonstrate 3D performance. As a result, you'll never get a TRUE indication of performance anyway.

Trust only what you see with your EYES. Play a game and if it feels nice and snappy, it is. If you really want to run benchmarks and base your purchase on those results, you'll never be satisfied with what you spend your money on. My buddy spent a fortune on his dual CPU system and if you were to look at his benchmark results you'd think it was a slug. Trust me. It's not. He could give a rat's ass what the benchmarks imply and believes his eyes rather than the synthetic numbers.

In the end it's like anything else. You get what you pay for. If you bought your system to play games or edit music or create CAD drawings ... whatever ... and that's what you intend to actually use it for, run those games and applications and judge what you see with your own eyes. If you bought your system simply to run benchmarks, you're going to have a VERY frustrating computing experience.

There is no spoon ...

Later.

Thanks for the replies. The only reason I made the post was because I'd read that someone with a much slower CPU but the same graphic card returned a 12,000+ score in 3DMark 2001 which struck me as odd.

I'm very happy with my new system as my games and applications run superbly, much smoother than on my old Athlon XP 1800+ with 128MB GeForce 3 Ti 200 card so I guess numbers don't really matter... that much! LOL

Oh and I have a 256MB GeForce FX 5600 card but some applications show it as a 128MB card for some weird reason. Has anyone else come across this 'bug'?

If you want
I can remote you and take a look?
PM Me.

V-SYNC? That is, intentionally holding back you GPU's throughput but timing its drawing speed to the refresh rate of your monitor! If it's on, DISABLE V-SYNC!

wasn't this mentioned aready ?

v-sync doesnt cause drops in scores as big as is suggested here.