🚀 go-pugleaf

To settle the question once and for all I ran the DBcc instruction
which takes exactly 10 clocks to branch in the following loop:

	    move.w  $99,D0
   inloop   move.w  $10000,D1
	    dbra    D1,*
	    dbra    D0,inloop

which should produce very close to 10 million cycles of execution time
(10,003,422 cycles to be exact.)
The result ran in 1.87 seconds which indicates an effective clock speed
of 5.349 Megahertz for programs running out of ram on the Mac.

Remember that the vertical retrace exception handler was intermittently
gaining control of the MPU during the test.  Therefore, the 6 Megahertz
claims may very well be valid if you include all processing done in ram.

However, the effective rate a program can obtain is limited by what's
left over and if your running disk activity, the serial ports, DA's which
hook themselves into timing exceptions or are moving the mouse, your effective
rate may be lower than the 5.349 Figure I got.

Dave Trissel  {ihnp4,seismo}!ut-sally!oakhill!davet