I wrote:
> I kind of suspect that Wirt's programs are more CPU-intensive than most,
> since I think the 30% is at least on the order of the maximum possible
> improvement resulting from an OCTCOMP.

A quick test of the good old Prime Number Sieve benchmark in BASIC/V shows
that it runs 3.75 times faster after OCTCOMP, which is faster than I
expected (but still almost within an order of magnitude of the 30% value
hence my hedge in the previous statement :-).

This value of 3.75x faster is probably pretty close to the upper limit on
the possible improvement for any section of CM code.  Wirt's reported 30%
overall improvement is quite believable assuming that his program spends a
lot of time in its own code.

Again, your mileage *will* vary.

G.

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *