Time for some Intel 486 CPU tests. The mobo is an MS-4144, PCB revision 1.4. I've got 2 sticks of single-sided FPM SIMMs in there; 32MB in total. The graphics card is an STB Trio64V+ with 2MB of RAM. The CPUs are from left to right: Intel DX2-66 (write back model), 486 DX-33, and finally on the right an SX-25. (Ignore the 2 crossed-out CPUs - I couldn't be bothered to test them.)

The board's configuration can be seen here -

http://stason.org/TULARC/pc/motherboards/M/MI … 86-MS-4144.html

DX2-66 CPU (SX955: SL enchanced & writeback)
Superscape 3Dbench 1.0 = 50.0

DX-33 CPU (not SL enchanced, and no writeback)
Superscape 3Dbench 1.0 = 27.7

SX-25 CPU (not SL enchanced, and no writeback)
Superscape 3Dbench 1.0 = 20.4

SX-25 CPU - but this time, inside the BIOS set up area, I have disabled everything speed related, including all cache, etc.
Superscape 3Dbench 1.0 = 3.5

Question: PCI performance on most, if not all PCI 486 boards is not very good. But how much does it matter? For the DX2-66 CPU, the VGA graphics test scores 50 FPS. Isn't that OK? Sure, pcpbench scores poorly, but wouldn't a game running at 640x480 be run on a Pentium?

Edit: For the last test, the one with the SX-25 with the BIOS "speed settings" set to Disabled, I did the "Ultima 6 speed test". This game has 2 introductions which both look silly if played on fast machines. However, with this SX-25 test, the introduction animations looked a bit too slow!