home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mitech!gjc
- From: gjc@mitech.com (George J. Carrette)
- Newsgroups: comp.lang.scheme
- Subject: Re: Are interpreters now as fast as compiled code used to be?
- Message-ID: <4066@mitech.com>
- Date: 15 Dec 92 10:40:39 GMT
- References: <4051@mitech.com> <BzA0Mo.II@cs.psu.edu>
- Organization: Mitech Corporation, Concord MA
- Lines: 28
-
- In article <BzA0Mo.II@cs.psu.edu>, schwartz@roke.cs.psu.edu (Scott Schwartz) writes:
- > With Bartlett's Scheme->C on a sparcstation 2, I get:
- > 0.2 real 0.0 user 0.1 sys
- >
- > Of course, nothing that runs for that short a time will be a sensible
- > yardstick.
-
- Indeed. Of course when I first ran the thing (fib 20) was a reasonable
- chunk of time! A Sun 3/180 took 17.5 seconds. The first Sun 4's took 4.2
- seconds. (My 0.42 second time figure on ALPHA/VMS is accurate, by the way).
-
- Obviously on an operating-system/machine with a lousy clock we are going
- to have to up the time figures back into the range of seconds.
-
- How about (standard-fib 25) ;; 4.85 seconds on Alpha.
- or (sfib 25) ;; 3.98 seconds on Alpha.
-
- Where sfib has built-in procedures pre-bound:
-
- (define sfib
- (eval `(lambda (x)
- (,if (,< x 2)
- x
- (,+ (sfib (,- x 1))
- (sfib (,- x 2)))))))
-
-
-
-