🚀 go-pugleaf

Earlier this year Semantic Microsystems  ported a  Scheme (Lisp) byte
code interpreter from  my Sage  II to  a Macintosh.   The interpreter
kernel and  heap image  were ported  in binary  form, so  there is no
doubt  that  the  Macintosh  was  executing exactly  the same machine
instructions  as  the  Sage.    On every  one of  our standard Scheme
benchmarks the Macintosh was 49-51% slower than the Sage.   Since the
Sage has an 8 MHz clock with  zero wait  states, I  conclude that the
effective clock rate of the Macintosh lies between 5.3 and 5.4 MHz.  

In other words, the  video refresh  interrupt (and  whatever code was
associated with it by the stock Finder that we were using)  eats up a
third of the processor.

				William Clinger
				willc@tekchips