stk and memory leaks

From: Andrew Dorrell <andrewd_at_ee.uts.edu.au>
Date: Wed, 3 Jul 1996 10:19:25 +1000 (EST)

I have noticed that stk grows as well. It seems to me as though this
may be a bug with the heap allocation algorithm in stk itself. For
example:

(define a (vector->list (make-vector (* 512 512) 10.1)))
;; cells used 268348/300000
;; # of used heaps 12
;; # of GC calls 12 (time spent in GC 6283.33ms)

(define a (map (lambda(x) (+ x 1)) a))
;; cells used 530491/850000
;; # of used heaps 34
;; # of GC calls 35 (time spent in GC 34383.33ms)

(define a (map (lambda(x) (+ x 1)) a))
;; cells used 530491/1100000
;; # of used heaps 44
;; # of GC calls 46 (time spent in GC 58600.00ms)

(define a (map (lambda(x) (+ x 1)) a))
;; cells used 530491/1200000
;; # of used heaps 48
;; # of GC calls 52 (time spent in GC 74133.33ms)

(define a (map (lambda(x) (+ x 1)) a))
;; cells used 530491/1225000
;; # of used heaps 49
;; # of GC calls 57 (time spent in GC 88350.00ms)


Notice that while the number of cells being used is constant, the number
of heaps increases continually. Could this be the problem?

The above results were generated using stk-3.0b2 under sunOS4.1.3.
The same thing happens under solaris 2.4.

Regards,

-- 
------------------------------------------------------------------------------
Mr Andrew Dorrell
School of Electrical Engineering                 *
University of Technology, Sydney                     *
PO Box 123                                     *
Broadway NSW 2007                                   .  
AUSTRALIA
                                                    *       /---\  Whoo?
Phone:   61 2 330 2395                                      (o o) /
Fax:     61 2 330 2435                                      ( : )  
email:   andrewd_at_ee.uts.edu.au                               ^ ^     
    OR   dorrell_at_ihf.uts.edu.au
------------------------------------------------------------------------------

Received on Wed Jul 03 1996 - 02:21:04 CEST