> gvb announce that his vampire is faster than X5000.

...in memory bandwidth performance (LONG access specifically), measured by running an m68k program. Having to go through Trance/Petunia seems to degrade the results of this benchmark to a fraction of the results a PPC-native program would reach, so this is surely misleading.
We also know that compared to X1000 or PowerMac G5, the X5000 is a slouch in memory bandwidth performance, but in this test the Vampire V2 also beats the X1000 and the 2.7 GHz PowerMac G5.
To my mind, these results are not meaningful, because MorphOS and OS4 users primarily run PPC-native software, and with such, the PowerMac G5, X1000 and even X5000 destroy the Vampire V2 in absolute memory bandwidth performance.

Btw, on PowerMac G5 2.3 GHz, 'bustest size=8m fast' gives 350 MB/s for the 'readl' test, which is faster than the X1000.

> Amiga X500

What?