3d Mark 2000 Score

[WillyCAD]

Asus P3V4X (Bios 1003)
PIII 600e@800mHz
256mb Crucial PC133
TNT II 32mb 4X AGP
SoundBlaster Live! (X-Gamer)
Western Digital 20.5GB
Maxtor 10.2GB

I ran the standard benchmark (my desktop settings were at 1280x1024, vsync was off).

Score = 1941

When the prices drop on any NVIDIA GeForce GTS 2 card, I will most likely buy one. I think this will make a good deal of difference on Quake III, Ground Control, and Daikatana. Any comments???

What scores did anyone get???

 

[wonder mike]

the geforce 2 mx cards will be coming out soon. it has performance on par with the geforce ddr and will be out for about $120, wait for that, by the time the geforce 2 gts drops in price quake4 will be around.

 

[Fireball]

Ooof. That can't be right, can it? 1941? I got 5431 on my system, which is listed in my sig. There can't be that much difference, could there?

 

[Sparky Anderson]

That score does seem low...I got 2450 or so with:

P3 450
Voodoo 3 3000
Monster MX300
Abit BE6
128 MB PC100

 

[coke]

Yeah it's quite low
I got 3239 with P3-650@650 256k (doh!) * 128MB Kingmax TinyBGA PC133 CAS2 * ASUS P3B-F * 26.0GB Quantum LCT (argh!) * Viper V770 ultra @ 150:183

 

[Prodigy^]

that sounds a bit low......i get around 2950 with PIII 750, v3 3k....

 

[sov05]

Please list the rez u tested at. He used 1280, that is a reason why.


At 640 I get 6850


PIII 800
256 RAM
GeForce SDR from ELSA 145/180
40 GB Raid 0

 

[sov05]

Please say the resolution and driver version. The resolution makes a huge difference in score. He tested in 1280x1024 resolution.
Also, if you oc, the oc speed. Further, the optimization u use, T&L or PIII SSE or Athlon, or 3dNOW

I get 6850 using 640x480 T&L and Detonator 5.32.

Geforce SDR running 145/180.
PIII 800
256 RAM

 

[WillyCAD]

O.K.

So am I doing something wrong???

or.... is it my settings. I have my desktop at 1280x1024......

 

[xtreme2k]

well me

P3/700E at 818MHz
256MB PC100
Abit BX6/2
Creative TNT2ultra 166/200

~3700 3dmark2000 1024/16

 

[Fireball]

He said his desktop was at 1280x1024, but 3dmark doesn't use that as the res. Anyways, with that system, he should still be faster. I still hit well over 4k at 1280x1024. Something definately is not right. I'd check the drivers first.

 

[sov05]

try reinstalling the drivers maybe

 

[[Frosty]]

3Dmark is video-biased. If you have a PIII 1GHz with a V3, it'll probably score lower than a PIII 600 with a GTS.

 

[MisterM]

3D Mark 2k is basically worthless as a benchmark, it completely favors the geForce based vid cards, try using an game you play very often as an benchmark instead.

 

[WillyCAD]

Does everyone agree with this???

Is this benchmark more based on video cards, not processor?

My NVIDIA TNT II is the M64 version. I know it is not the best video card, but I refuse to pay $400+ for a video card right when it comes out..... I will get a GTS 2 card when they hit $100 to $150.

 

[Electric Amish]

3dMark is biased toward NVidia cards....always has been.

My G400 and an Athlon 700@805 gets about 2400 on the default bench.

My friend with an Athlon 700 not overclocked and an Elsa Gladiac GeForce2 GTS gets over 6000 on the default bench.


amish

 

[bluemax]

I agree, 3Dmark is biased. Not only that, but my old i740 gave me 1000 marks on 1.1, my new V3-3k gives ~1550. I can *see* more than 50% improvement!!!! Better visual quality and VASTLY better speed. Those scores don't mean much... it *is* a good demo of a video card's performance, though. If your card flies though it silky smooth, you have a good setup!
If you get a weird coloured slideshow, you don't.

 

[vaclav]

Yeah, 3dMark 2000 is biased! Just look what it did to my office PC:

Dell XPS T500 and ATI Xpert98D with 256MB PC100 RAM:

CPU Intel Pentium III 500 Mhz
3D Accelerator ATI Xpert98D AGP 2X
Chipset ATI 3D RAGE PRO TURBO
Resolution 768*1024
Z-Buffer Depth 16-bit
Frame Buffer Triple
Refresh Rate 83 Hz
CPU Optimization Intel(R) Pentium(R) III

3DMark Result 484 3D marks
CPU Speed 78 CPU 3D marks
Game 1 - Helicopter - Low Detail 9.3 FPS
Game 1 - Helicopter - Medium Detail 7.0 FPS
Game 1 - Helicopter - High Detail 3.1 FPS
Game 2 - Adventure - Low Detail 7.7 FPS
Game 2 - Adventure - Medium Detail 7.0 FPS
Game 2 - Adventure - High Detail 6.2 FPS
Fill Rate (Single-Texturing) 33.3 MTexels/s
Fill Rate (Multi-Texturing) 33.3 MTexels/s
High Polygon Count (1 Light) 544 KTriangles/s
High Polygon Count (4 Lights) 544 KTriangles/s
High Polygon Count (8 Lights) 548 KTriangles/s
8MB Texture Rendering Speed 25.4 FPS
16MB Texture Rendering Speed 21.9 FPS
32MB Texture Rendering Speed 13.0 FPS
64MB Texture Rendering Speed Not enough AGP memory
Bump Mapping (Emboss, 3-pass) Not supported
Bump Mapping (Emboss, 2-pass) Not supported
Bump Mapping (Emboss, 1-pass) Not supported
Bump Mapping (Environment) Not supported

 

[Quickfingerz]

Yeah those scores do seem slow. Under default benchmark, the card should be only slightly slower than a voodoo3 3k. I use a voodoo3 3k and i score around 3150 with my Athlon @900.

 

[Eric_D]

Soyo 6BA+IV
P3 700e cB0
128 MB infineon
3D Prophet II GTS 64MB


700e @

700= 5844

933= 7109

980= 7199

1008= 7345


At default test 1024*768 16bit

 

[SuperFreaky]

My PIII 450 and Velocity 100 score a whopping 600!!!!!!!! yet I can still play UT and Q3 fine......

 

[Eug]

Depends on what you use your video card for.

3DMark2000 is only really useful for testing your own computer when tweaking drivers, etc. It's not very useful to compare against other cards, esp. since it strongly seems to favour nVidia cards. Perhaps nVidia cards are often faster in general or esp. in Quake 3, but similarly priced lower-end nVidia cards are no match for 3dfx cards in Glide-based games.

I get around 3070 with my V3 2000 Celeron 840 at 1024x768 16-bit. (I can get over 3200 if I up the V3 and Celeron speeds though.) A TNT2Ultra will score higher, but that doesn't mean it's better in UT.

 

[blade47]

My score is around 3100, I can't remember the exact number from the last time I ran it though.

Abit BX6-rev2
P3-500e@620mhz
128mb PNY Ram Cas3 (124mhzfsb)
Voodoo3 2000@167mhz (83mhz agp bus)
WD 10.1gig ide ATA33
Monster MX-300
Archtek USR V.90

1024*768 16bit color
v-sync off
P3 optimization
Triple frame buffer
16 bit Z-buffer depth

It doesn't matter what resolution the desktop is set at 3dmark defaults back to 1024*768 unless you change the settings in 3dmark itself.

WillyCAD
3dMark 2000 is a benchmark measuring 3d performance so it only makes sense that the video card is going to affect the scores more than anything else.

Keep in mind though that it is measuring the performance in D3D applications. If you want to compare openGl performance use Q3. One card may be faster in D3D but slower in openGl.

Another thing to consider is glide. A Voodoo card may get beat by another card in D3D or openGl but if the game also supports glide the Voodoo running in glide may be much faster than the other card running in D3D or openGl.

As far as 3dMark being biased I can't say from personal experience since my Voodoo3 is the only card i've run 3dMark on.

 

[sov05]

I use 3dmark for 1 main reasonm

to see if my overclocking went too far, if it passes 3dmark I believe it is stable.

I also, use Quake III TimeDemo 001 as another "stabilitity test"

 

[TurtleCrusher]

38320 with desktop in sig.

http://www.futuremark.com/benchmarks/legacy