3DMark 2001 scores...where should I be?

[BujinZero]

Hi

I recently ran 3DMark 2001 on both of my systems, and I'm not sure where I should be. Both computers are running Win98, the latest Detonator drivers, and DX8.0. 3DMark is at its default settings.

System 1:
Athlon 1200 266fsb
ABit KT7A
Infineon 512 MB PC133 SDRAM (running at 133)
Creative Labs GeForce2 GTS 32MB
SCORE:3100

System 2:
Athlon 1333
EPoX EP-8K7A
Mushkin DDR 256MB x2
Gainward CardExpert GeForce 2 Pro Golden Sample
SCORE:3700

I was a bit concerned with these scores--they seem a bit low compared to what I'm seeing from others. Any thoughts, suggestions, or questions?

 

[OverDose]

I know I gained about 500pts switching from athlon 1.2c to athlonXP 1600+ @ PR1755

powerstrip to overclock the core and memory helps pretty good too

 

[E3Art]

and I thought my score was low i got a 7017

Athlon 1800+
Radeon 8500
512 DDR RAM
Epox 8kha+


My question is the same as above, how can i get mine higher? Is overclocking the only answer?

 

[Duvie]

System 1 sounds about right as that version I believe of the gts card has relative slow ddr ram chips...plus you are using sdram...plus you are running a kt133a mobo...plus you are spotting system 2 133mhz...

Now system 2 to me seems a bit low...I have same card and with a 1.4tbird on a sis735 I get 4960 (haven't cracked 5000 yet)...The gainward has a performance setting of 220/444 that is set with a clock of the expert tool program...This card runs the 4.5ns memory and is partly why the ram can be clocked so much higher from the default 200/400 settings...Are you sure you have vsynch off??? Make sure under direct3d tab it is also off there. Make sure antialias is off as well.

Since sis735 is a bit better performer then the amd760 or 761 chipset of the Epox, and I was using the enhanced settings boosted a bit more to a nice round 225/450...PLus I have 66mhz on you...I think you still should have gotten 4500 at least...

 

[imgod2u]

3dMark2k1 rates pixel and vertex shaders. So a Geforce 3 or Radeon 8500 would probably have significantly higher scores.

 

[damocles]

There are also tweaks that many people use to raise their scores with prgrams like nVmax. If you change the LOD, mip mapping, Vysnc etc you can really life your score. It is one of the reasons I am so dubious of 3Dmark boasters

Your scores do see a wee bot low to me though. Perhaps both being 200-300 below what I would have expected

 

[Insidious]

When I was on a GeForce2 GTS my scores were pretty much in line with yours.

You have to remember two things about benchmark boasts you see in forums like this one....

1. Many people strip their systems to a point where they will do little else but run a benchmark (both hardware and software tweaks)
So the big numbers are pretty much meaningless. er, unless all you want to do is run demo loops and post in forums.

2. Some people LIE! I don't know why some of the people can acutally believe untrue boasts make their ***** bigger, but they do


you might look at the GeForce2 advanced settings tab in display manager and make sure Vsync is Off on D3D, but I think you are fine

 

[steimm]

My rig was:
Intel PIII700E@933
ASUS P3V4X
Crucial 512MB cas2
Gf2 MX (PowerColor) 175/160 (yes 160MHz memory original was 143), a really BAAAD card!!! Driver 21.83.
WinME
DX8.0

The score I had with that videocard was round 2800 (of course 3DMark2001)... seems much better than 3100 that you get with that rig...

With the new card Leadtek Gf3 Ti200 175/470 (driver 22.80) and Win2k I get around 5400 points, but that is too low I guess, or????

/steimm

 

[BujinZero]

Well gents I toyed with the BIOS on the KT7A system and installed DX 8.1 and managed to squeak out a little over 3200. Also, either due to stupidity (most likely) or the new DX (not likely), I cannot find the option to disable vsync. My exact steps are: Display properties>Settings>Advanced>GeForce2 GTS>Additional Properties>Direct3D Settings--and from this page it looks like more SHOULD be there. All I see is something about z-buffer, fog emulation, and D3D logo. Where should I be going?

I have not yet played with the EPoX system.

Thanks for your comments!

 

[techweenie]

I guess my system is way off as well. I have the following
Shuttle mobo
1.4 GHz Amd Tbird
64mb Geforce 2 Ultra
256mb ram
and I only got a 3817... Maybe I'll put my 512 mb dimm back in, and tweak some stuff...

 

[Insidious]

OOPs!

Don't ya love it when we lie to you! Sorry 'bout that.

with my GF3 and det. 23.11, I don't have the vsync option in my D3D tab of display settings either..... Only on OpenGL tab.

If you really want to disable it you can use a program like RivaTuner (look under video utilities)

and it will let you change it.....

I use RivaTuner to enable overclocking tab in display settings, but don't change anything else because I've been too lazy to figure

out what many of the terms/functions mean.

Good Luck!

PS: I don't think you need to be worrying too much with the scores you have now

 

[steimm]

I use NVmax for tweaking the settings of the card, I like that little piece of software really much!

/steimm

 

[ConFuseMe]

Under the Direct 3d tab you guys are looking at there should be a button toward the bottom of that page that says "More Direct3D.
Thats where you can disable the vertical sync.

 

[steimm]

you can't always get that from clicking on "More Direct3D"...
You have to activate coolbits (or use NVmax) to make that (vertical sync information) appear!

/steimm

 

[chizow]

<< and I thought my score was low i got a 7017

Athlon 1800+
Radeon 8500
512 DDR RAM
Epox 8kha+


My question is the same as above, how can i get mine higher? Is overclocking the only answer? >>



Actually, that does seem a bit low. That cpu/mobo combo with a Radeon 8500 usually yields around 8000 3dmarks stock outta the box. You may want to try checking your AGP settings and play with the fast writes. I'm using 4x AGP with 128MB AGP aperture with fast writes enabled with the same RAM, mobo and CPU, but a GF3 Ti200. I get around 7000 stock, but after OCing I'm sitting around 8700

I could probably go higher but I just haven't had the time to update my BIOS and check the cooling on the card. When I was running stock and comparing to other systems with same CPU and mobo, the Radeons I saw were consistently scoring 8K 3dmarks stock, and 9K or so when they OC'd their system and their card. I haven't messed with too many of the options in nVMax, leaving LOD, mipmapping, AA, and V-Sync on "Auto". I did enable some of the other features such as win2k/Via fix, Via 4x fast write and sidebanding and of course a healthy 240/510 overclock

Chiz

 

[BujinZero]

I think I'll download nVMax or RivaTuner and have a go at it. I don't think I'll worry too much about the scores, though. This is just an observation, but it seems like the 3DMark scores increase exponentially depending on someone's video card. For example, use of a GeForce 3 or Radeon 8500 seems to double 3DMark scores when compared to a GeForce 2. That would clearly explain the 8000s you guys get with your GF3s/R8500s in computers so similar to my EPoX set up. That observation leads me to ask you GF3/R8500 owners: Are those cards twice as good as your old GF2s? Is a difference really that noticable, or are these just flashy numbers we're showing each other?

My GF2 GTS/Pro seems to play games very well. If GF3s and R8500s are THAT good, you Anandtechers out there must be the only computer users who can really make use of it. Man, do I love technology! It improves itself before your order even ships. Things are very interesting for us geeks. I've encountered many people who are sold a souped up Pentium 4/GF3 system (ok, hyped up for the P4) and use it for email and the internet. C'mon, its like giving grandma a Ferarri, sure she can brag about it and drive it, but she'd never notice if that extra 350 horsepower was even there. Sorry, I'm off on another rant/tangent again!!!

Thanks for the help everyone, these forums seem like a good place to ask.

 

[anime]

Chiz--he was saying he got that score stock out of the box. Not OCed or anything--he was asking how...
For radeon 8500- I'm using powerstrip to OC it. I got 8400 ish scores more or less after some tweaking. Base score is about 7400 ish though--but My system is an OCed system to start with.

 

[bjc112]

<< 3dMark2k1 rates pixel and vertex shaders. So a Geforce 3 or Radeon 8500 would probably have significantly higher scores. >>



I think only the games are counted..

 

[chizow]

<<

<< 3dMark2k1 rates pixel and vertex shaders. So a Geforce 3 or Radeon 8500 would probably have significantly higher scores. >>



I think only the games are counted.. >>



Yes, only the games are counted, but the extra features of the 8500 and GF3 push those extra "theoretical" tests which are translated into the "real world" performance seen in the game tests (particularly the nature test). Btw, there is a significant difference in the nature test between tweaks and cards. I got a slide show @ 3fps with my V5, my GF3 runs it @50fps and @30fps before I tweaked it.

anime you're right, I shoulda clarified what I meant. I consider "tweaking" running your system out-of-spec. I don't really consider software/bios optimizations as tweaking, I consider them stock. For instance, I don't run my system if it is unstable at turbo ram timings/cas2/4-way enabled. I won't go to cas 2.5 or 4-way off to bump up my FSB a few points. I also don't consider driver or AGP settings to be tweaks. Yes, the average user might not ever upgrade or set their AGP settings to where they *should* be set, but to me, these are normal settings. If you have a 4x AGP card and running 2x AGP, you're not running stock specs, you're running under-spec

BujinZero I haven't owned a GF2, only Voodoos prior to this GF3 ti200, but I must say I am extremely impressed with this cards speed. That's not saying a whole lot b/c the V5 struggled at resolutions above 32-bit and 1024 in fps games. I never tried running any of the newer gen. games, but I'm sure its struggles would have been magnified. Having seen the difference, I'm thinking the difference between a GF2 and a GF3 would be noticeable at higher resolutions in 32-bit color as well as lower resolutions in 32-bit color with AA on. You also might see differences in newer games that take advantage of the more robust GF3 and 8500 T&L engines. Basically, the GF3 and the 8500 make high resolution and Anti-Aliased gaming a more viable option. No, you won't see a big difference in Quake3 in 16 bit at 1024 on a GF3 or a GF2, but the difference (in FPS at least) might be significant at 32-bit and 1280. Average fps is deceiving as well, b/c I'm a firm believe anything over 70fps isn't very noticeable at all. However, in a scene where your frames start to stutter at 30fps, it becomes very noticeable. So even though a GF3 at 120 avg fps doesn't seem that noticeable over a GF2 at 70 avg fps, you need to take into consideration the lowest amount of frames in that average, b/c slowdowns will be very noticeable.

Chiz