3D Game BenchMarks
Home 3D Game BenchMarks Search Page

3D perfomance of the PC (x86) systems can be characterized hardly by a single number, because there are so many ways to make 3D graphics. It may happen that a certain system turns out to be fast when tested by one method and slow when tested by another. We can use 3D Bench to obtain preliminary impression about the speed of older (classical) 3D games like the "Wolfenstein 3D", "Spear Of Destiny", "Blake Stone", "DOOM", "DOOM II", "Heretic", "HeXeN", etc. by ID Software and others. The common feature of these games is the relatively low (or even missing) FPU usage.

Recent sVGA 3D games utilize the FPU of the processor rather intensively. Therefore we need new 3D benchmark programs (FPU users) like, for example, the Chris' 3D Benchmark. Unfortunately, even this program uses less intensively the FPU of the processor than many top quality 3D games like "Quake", "Quake II" and "HeXeN II. Therefore, we have to use the real games if we really want to know the 3D-FPU speed of a given system. You can download the most popular "test games" from here: Quake v1.06, Quake II v3.13, Turok D3D/OGL.

There are many excellent 3D games out there which are very playable in sVGA (640x480) resolution even without a strong FPU or 3D accelerator card, like "Duke Nukem 3D", "BLOOD", "Tomb Raider", "Tomb Raider 2", "Manic Karts", "Need for Speed", "Need for Speed 2".



How to Test by the Games ?

Quake v1.06 test: The most popular test game. It is very FPU dependent. If you want to run the test, you need a VESA VBE 2.0 complient 2D sVGA card, or a 3D accelerator. If the BIOS of your 2D card is just VESA 1.1 or 1.2 complient, you can use the SciTech Display Doctor to obtain VBE 2.0 (or higher) compatibility. To run the test, choose 640x480 resolution, full screen, without CD audio (start Quake with the parameter "-nocdaudio"). To start the benchmark, type "timedemo demo2" in the console, than press the consol button " ` " (next to the "!" key in the upper left corner of your keyboard) again immediately. The result is given in fps.
There is also an Open GL version of the game especially for 3D accelerators (3Dfx), the so called GLQuake v0.93, (Quake v1.08). (You can read more about GLQuake here. You may also find quite useful Chris' Guide about how to increase the 3Dfx fps in GLQuake, and the GLQPlus starter which can also help optimize the performance.)

Quake II v3.13 test: It is very FPU dependent. After you have started Quake II, choose the appropriate graphics driver under the VIDEO/driver settings ("softvare", "default OpenGL", "3Dfx OpenGL", "PowerVR OpenGL") and choose also the appropriate Video Mode (512x384, 640x480, 800x600, ...). When you are in the game, anytime you may test the refreshing time by typing in the consol the "timerefresh" command. If you want to run the timedemo, first have to switch to the demo mode by typing in the consol the "timedemo 1" command. You also have to enter a second command, corresponding to the demo file you want to run: "demomap demo1.dm2"

Turok D3D/OGL: You have to start the game with the parameter "-benchmark". The current and also the avarage frame rate will be continuously displayed. To see the "end result" ("TMarks in fps") you must see the game for quite long time (:-))).



Interesting Links:

Drivers for 3D Accelerators
OverClocking Guide

3Dfx Interactive
3Dfx Links
3Dfx Drivers
3D Test results of KeyLabs
How to optmimize Cyrix/IBM 6x86MX systems for 3D Games?
ATI's Graphics Products
Björn's 3D World
Tom's 3D Accelerator Benchmarks
NVIDIA's RIVA 128ZX at Tom's Harware Guide
The Voodoo2 3D Chip from 3Dfx at Tom's Harware Guide
About 3D Winbench 98 at Tom's Harware Guide
Rendition Verite 3D
3D Benchmarks



Links to MOTHERBOARD - PROCESSOR configurations:

6x86L-P200+ in VX PRo+
6x86L-P200+ in TX PRo
6x86MX-PR200 in TX PRo
P55C-200MHz in TX PRo
6x86L-P200+ in GA586HX512
6x86MX-PR200 in GA586HX512
P55C-200MHz in GA586HX512



Color Codes: UNDER-Clock NORMAL-Clock OVER-Clock MISSING-Clock Notice



Interesting facts for QUAKE gamers:

  • An over-clocked P55C-250 (=3x83) MHz with S3 VirgeDX can be slower (16.3 fps), than P55C-200 (=3x66) MHz with ET6100 (16.7 fps)!
  • It is possible that certain systems perform better at 3x50(=150) MHz (7.1 fps) than at 2x83(=166) MHz (6.8 fps) despite the slower core and external frequency!
  • It is possible, that a motherboard (like the GA-586HX512) performs better at 3x60(=180) MHz (12.3 fps), than another motherboard (like the TX Pro, 512) at 2.5x75(=188) MHz (11.9 fps), despite the slower CPU (the 6x86L instead of 6x86MX) and the slower core and external frequency!



  • goto SYS Bench VXPro+ 4ISA/4PCI, 512kB, 16MB/60ns EDO, 6x86L-P200+
    TEST_Type 3 x 50 3 x 55 3 x 60 2 x 66 2 x 75 2 x 83 Notice
    3D Bench S3VirgeDX
    3D Bench 200 200 250 200 166.6 166.6 ET6100
    Quake v1.06 S3VirgeDX
    Quake v1.06 7.1 7.9 8.8 8.2 6.1 6.8 ET6100



    goto SYS Bench Acer TXPro 3ISA/5PCI, 512kB, 16MB/60ns EDO, 6x86L-P200+
    TEST_Type 3 x 50 3 x 55 3 x 60 2 x 66 2 x 75 2 x 83 Notice
    3D Bench 250 200 200 250 S3VirgeDX
    3D Bench 250 200 250 250 ET6100
    Quake v1.06 9.5 8.4 9.1 10.4 S3VirgeDX
    Quake v1.06 9.6 8.4 9.1 10.4 ET6100
    GLQuake v1.09 25.1(30.2) 23.1(25.3) 24.6(27.4) 25.7 3Dfx/55MHz
    TurokMarks 25.2(28.9) 22.6(23.3) 24.7(24.7) 25.8(26.2) 3Dfx/55MHz
    Q2 timerefresh 9.6 9.4 9.4 ET6100
    Q2 timerefresh 26.8(35) 25.6(31.6) 25.7(33.7) 3Dfx/55MHz



    goto SYS Bench Acer TXPro 3ISA/5PCI, 512kB, 16MB/60ns EDO, 6x86MX-PR200
    TEST_Type 3 x 60 3 x 66 2.5 x 66 2.5 x 75 2.5 x 83 2 x 83 Notice
    3D Bench 250 200 250 250 S3VirgeDX
    3D Bench 250 250 250 250 ET6100
    Quake v1.06 10.0 10.1 11.6 10.8 S3VirgeDX
    Quake v1.06 10.7 10.7 11.9 11.0 ET6100
    GLQuake v1.09 3Dfx/55MHz
    TurokMarks 3Dfx/55MHz
    Q2 timerefresh ET6100
    Q2 timerefresh 3Dfx/55MHz



    goto SYS Bench Acer TXPro 3ISA/5PCI, 512kB, 16MB/60ns EDO, P55C-200 MHz
    TEST_Type 3 x 50 3 x 55 3 x 60 3 x 66 3 x 75 3 x 83 Notice
    3D Bench 166.6 200 200 200 S3VirgeDX
    3D Bench 166.6 200 200 250 ET6100
    Quake v1.06 13.2 14.7 14.7 16.3 S3VirgeDX
    Quake v1.06 14.6 16.7 17.2 18.6 ET6100
    GLQuake v1.09 28.2(35.8) 28.2(35.8) 28.2(35.8) 3Dfx/55MHz
    TurokMarks 33.4(41.0) 33.9(42.5) 34.5(45.9) 3Dfx/55MHz
    Q2 timerefresh 15.1 16.6 18.7 ET6100
    Q2 timerefresh 30.1(43.4) 30.1(43.3) 30.1(43.3) 3Dfx/55MHz



    goto SYS Bench GA-586HX512 4ISA/4PCI, 512kB, 16MB/60ns EDO, 6x86L-P200+
    TEST_Type 3 x 50 3 x 55 3 x 60 3 x 66 2.5 x 66 3.5 x 60 Notice
    3D Bench S3VirgeDX
    3D Bench 200 250 250 ET6100
    Quake v1.06 S3VirgeDX
    Quake v1.06 10.6 11.3 12.3 ET6100



    goto SYS Bench GA-586HX512 4ISA/4PCI, 512kB, 64MB/60ns EDO, 6x86MX-PR200
    TEST_Type 3 x 50 3 x 55 3 x 60 3 x 66 2.5 x 66 3.5 x 60 Notice
    3D Bench S3VirgeDX
    3D Bench 250 ET6100
    Quake v1.06 S3VirgeDX
    Quake v1.06 10.8 ET6100
    GLQuake v1.09 26.7(30.8) 3Dfx/55MHz
    TurokMarks (30.8) 3Dfx/55MHz
    Q2 timerefresh
    Q2 time demo1
    12
    7.4
    ET6100
    Q2 timerefresh
    Q2 time demo1
    28.7(33.8)
    15.4(15.5)
    3Dfx/55MHz



    goto SYS Bench GA-586HX512 4ISA/4PCI, 512kB, 16MB/60ns EDO, P55C-200MHz
    TEST_Type 3 x 50 3 x 55 3 x 60 3 x 66 2.5 x 66 3.5 x 60 Notice
    3D Bench S3VirgeDX
    3D Bench 166.6 ET6100
    Quake v1.06 S3VirgeDX
    Quake v1.06 ET6100
    GLQuake v1.09 (35.8) 3Dfx/55MHz
    TurokMarks (43.4) 3Dfx/55MHz
    Q2 timerefresh
    Q2 time demo1
    14.6
    8.9
    ET6100
    Q2 timerefresh
    Q2 time demo1
    3Dfx/55MHz

    E-mail to DcsabaS E-mail # Last Mod: Thu, 22 May 2014 23:13:08 GMT Hungary Balaton Budapest iNDEX MTA KFKI MFA Alag3