These are some results I've collected. These results are not to be taken
to seriously, cause some of them could be wrong. I myself forgot to set
the frame rate to 25.6 when I first tested :o). Motherboards could make
a difference, digital or analog steering devices could make a
difference.
I've also gotten posts on two Processor Occupancy sites which I
recommend in the following order:
1. http://www.racesimcentral.net/
2. http://www.racesimcentral.net/
Because in the second you replay a presaved hotlap and check the
occupancy. I'm not sure that this will give a correct reading.
The results:
The line is built up by:
processor,
motherboard cache size,
primary memory,
graphics card memory,
graphics card and
highest occupancy in SVGA with all graphics on at the start of a quick
race at Monaco from 13:th grid pos during the signal before the engine
starts.
486 120MHz, 256kB, 32MB, 2MB, Stealth 64 VLB, 790%
P90 (100MHz), 512kB, 32MB, 1MB, Avance Logic 2302/1300 PCI, 600%
P133, 512kB, 16MB, 1MB, ET4000, 390%
P133, 512kB, 32Mb, 2Mb, Stealth 64 Video 2201 PCI card, 390%
P133, 256KB, 16MB, 2MB, ATI Graphic Expression PCI, 390%
P133, 512kB, 16MB, 1MB, Cirrus, 380%
P133, 512kB, 16MB, 2MB, ATI 3D, 380%
P133, 512kB, 16MB, 2MB, Matrox Millenium, 350%
P133, 256kB, 32MB, 8MB, ATI 3D turbo pro pc2tv, 390%
P133, 256kB, 32MB, 4mb, Matrox Millenium, 340%
P150, 256kB, 32MB, 2MB DRAM, Daimond Stealth64 2001, 380%
P166; 512kB; 32MB, 2Mb; ATI Mach64; 300%
P166, 256kB, 32MB, 2MB, S3TrioV+, 330%
P166, 256kB, 32MB, 2MB, Genoa Videobliz PCI, 320%
P166, 512kB, 48MB, 4MB, 3dblaster, 310%
P166, 256kB, 16MB, 2MB, Matrox Millium, 340%
P200, 512KB, 32MB, 2MB, Matrox Millenium, 260%
P200 (225), 512kB, 32 MB, 4MB, Millenium, 240%
P200 (BUS 75MHz), 512kB, 96MB, 4MB, Matrox Millenium, 260%
P166MMX, 512kB, 32MB, 2MB, ET6000, 300%
P166MMX, 512kB, 32MB, 4MB, ATI 3D Xpression, 260%
P166MMX, 512kB, 48MB, 4MB, Diamond Stealth 3D 2000 PCI, 290%
P166MMX, 512kB, 32MB, -, Matrox Millenium, 260%
P200MMX, 512Kb, 96MB, 2MB, PCI Matrox Millenium, 230%
P200MMX, 512kB, 32MB, 4MB, ATI 3D, 240%
Pentium II 266, 512kb, 32MB, 4MB, Diamond Stealth 3D 2000, 220%
P200 Pro, 256k, 64MB, 4MB wram, Matrox MGA Millennium, 300%
Cyrix 150+, 256kB, 16MB, 1MB, S3+ PCI, 430%
Cyrix 166+, 256kB, 32Mb, 2Mb Hercules Dynamite 128, 300%
Cyrix 166+ (200+), 512kb, 32MB, 4MB, Diamond Stealth 3D 2000 using
S3spdup, 300%
Cyrix 166+; 512kB; 96MB; 2Mb; s3; 300%
Cyrix 166+, 256kB, 64MB, 4MB, Intergraph Intense 3D 100 PCI, 340%
Cyrix 166+, 512kB, 64mB, 4mB, EDO S3Virge3D 6x86opt.exe and S3vbe20 run,
330%
K5 100, 256kB, 32MB, 4MB, Matrox Mystique, 510%
K6 233(at225/75), 512kB, 32MB, 4MB, mystique, 220%
K6 233(at250/83), 512kB, 32MB, 4MB, mystique, 190%
The data in between () are overclocking or underclocking. I'm guessing
that the two overclocked P200 are overclocked to 225MHZ and has a bus
speed of 75MHz.
Goy Larsen sent me the data on the K6 233. He tried it at 225MHz and
250MHz. He also said he tested to overclock the P200MMX to 250MHz, but
it was slower than the K6 at 250MHz, unfortenately he didn't give me the
data on that one.
From the posts it was clear that GP2 should be run in dos, NOT from
Win95, so I hope all you guys restart your computer's in dos mode :o).
As for the results, you draw the conclusions :o).
/Christer, will probably go for the K6 :o)