OK, so I've found the AGP X1900 I've been looking for and I was going to do some comparative benchmarks with my X800XT. "First of all" - I thought - "I'll do the Quake III with timedemo". I disabled the vsync (easily spotted by constant tearing of the game menus) and run the test. Much to my dismay, the FPS counter has shown only 59,6 FPS, and exactly the same number the next time. I was shocked, as I remember more like 90 FPS when running the same test the last time (in the MorphOS 3.9 times). It looks like the engine is ignoring the disabled vsync and is dropping the framerate to match with the screen. Is that possible?

Ah, OK. I've found the

Code:

console switch and now the game runs at 103,5 FPS with all the bells and whistles in 1600x1200. Now it's time to switch the gfx cards and check if it works. Keep your fingers crossed!

OK, I can see a clear progress. The last time I've tested two X1900's Pro AGP versions and none worked. This time none worked, BUT in one case I was able to see my desktop... for about 10 seconds and then everything hang. Then I couldn't reproduce my success any more, no matter how much I've tried. What is more interesting - that (partial) success happened when I was running a gfx card that PC failed to initialize. The one that worked like a charm in a PC didn't show me anything. Even the fans didn't spin to a higher speed (like when the system boots).

Moreover, when I put back my X800XT, computer started to misbehave. The mouse coursor refused to move, CPU temperature upped to 95 degrees Celcius... Had to remove it, unplug everything, clean the AGP slot with some air and now it looks like it works again. I'm a bit scared to do another try today, so let's call it quits, OK? I'll go and find some better powering cable than from a DVD molex and then... maybe... I'll dare to do it again. Hopefully.