System Setup

Intel Celeron 400 PPGA
Supermicro P6SBA
Abit BH6
128MB Corsair 3.3v CAS-2 unbuffered PC-100 SDRAM
nVidia reference RivaTNT - STB Velocity 4400
     Driver version 4.10.01.0152
Western Digital Caviar 26400 6.4GB EIDE HD

Benchmarks
(All tests run under a clean install of Windows 98)
Quake II 3.20 640x480x8
Ziff Davis Winbench 99 - 1024x768x16
     CPUMark32
     Business Diskmark99
Ziff Davis Winstone 99 - 1024x768x16

Special thanks go out to U-Tron Technologies for providing our test-bed components.

We'd also like to thank Corsair Microsystems for providing their high-quality Intel-certified SDRAM for our overclocking/reliability tests.

Winbench 99 CPUMark32
Slocket	Score
Abit AB RS-370	751
MSI MS6905	751
Gigabyte GA-GR7	750
Celeron 400 SEPP	750

Winbench 99 Diskmark32
Slocket	Score
Abit AB RS-370	2230
MSI MS6905	2230
Gigabyte GA-GR7	2240
Celeron 400 SEPP	2240

Winstone 99
Slocket	Score
Abit AB RS-370	20.5
MSI MS6905	20.4
Gigabyte GA-GR7	20.1
Celeron 400 SEPP	20.3

Quake 2 640x480 Software
Slocket	Score
Abit AB RS-370	30.1 fps
MSI MS6905	30.5 fps
Gigabyte GA-GR7	30.1 fps
Celeron 400 SEPP	30.2 fps

It's clear from these tests that there is no noticeable difference in performance between any of the converter cards, or between a converter and a standard Slot-1 Celeron. Each test showed virtually identical results. This is interesting due to the unique nature of each design, but if one thing is clear, it's that none of the converters here suffer from poor design.