Vincent BARAT wrote:
> 
> In fact, I do not want to deal with *SYSTEM* results, such as I/O
> tests, but only with raw CPU speed:

OK, you have dhrystone.  That's what it is for.

> I don't beleive that bytemark raw CPU results are affected the by
> the system on which they run (are they ?), and that's why I was
> really disapointed to see that my 30 MHz 68040 is just slitly faster
> than a poor 33 MHz 386.

Did you compare *only* the dhrystone results?  If you had, you would
probably have found the 68040 was significantly faster than the 386
running at the same clock freq.

IIRC, Hauke's 68040 at 33MHz marked 14,000-15,000 dhrystone, whereas a
386DX at 33MHz should be somewhere around 8,000 (could be a little
better but not more than 10,000, I think).  SX should be even worse.

Ken