dwighth at ipa.net
Mon Feb 22 19:24:32 UTC 1999
Carl Watts wrote:
> As you know, Smalltalk should spend most of its time in primitives (if I remember correctly we targeted it at over 70% at ParcPlace). And, of course, primitives can run at the speed of optimized C. If this is true, then the need to speed up the other ~20% is lessened.
I have some reservations about this 70% figure -- I could only see this
as a valid figure in reference to a particular benchmark suite. Much
would depend on how your code is written and what you are doing. If your
code is structured with abstractions layered on top of abstractions etc.
then I would expect the method send/execution time to become a much more
significant portion of the execution time, and therefore a much more
attractive optimization target. Basically, I would want to minimize the
"penalty" for using extensive abstraction and higher level modeling.
(BTW, what happened to the other 10%? <g>)
More information about the Squeak-dev