Aziz,
So I keep some of those little combinators in separate libraries.
Here's the time for a version which imports 'bi' and 'uni' from a
separate library:
~ # ikarus --r6rs-script /scratch/_fib-fp-dharmalab-a.scm
running stats for (fib 34):
no collections
1376 ms elapsed cpu time, including 0 ms collecting
1389 ms elapsed real time, including 0 ms collecting
0 bytes allocated
~ #
vs the version when they're in the same library:
~ # ikarus --r6rs-script /scratch/_fib-fp-a.scm
running stats for (fib 34):
no collections
336 ms elapsed cpu time, including 0 ms collecting
358 ms elapsed real time, including 0 ms collecting
0 bytes allocated
Any more tricks up your sleeve? :-)
The actual 'dharmalab' libraries are at:
http://github.com/dharmatech/dharmalab
By the way, perhaps 'time' should show the 'cp0-size-limit' and
'optimize-level' so that you don't have to guess (or ask) users when
they submit timing information.
Ed
----------------------------------------------------------------------
(import (rnrs)
(dharmalab combinators concatenative)
(ikarus))
(define (sub x)
(lambda (y)
(- y x)))
(define (constant x)
(lambda (y)
x))
(define (ifte f g h)
(lambda (x)
(if (f x)
(g x)
(h x))))
(define (less-than= x)
(lambda (y)
(<= y x)))
(define fib
(ifte (less-than= 1)
(constant 1)
(bi (uni (sub 1) (lambda (x) (fib x)))
(uni (sub 2) (lambda (x) (fib x))) +)))
(time (fib 34))
----------------------------------------------------------------------
