Allegro CL Floating Point

[berger%tartarus@gargoyle.uchicago.edu (jeff berger)]

I tried running the programs in the e-mail message from
maida@cs.psu.edu in our sun4 environment and got similar results. The
C program:


tartarus 25 % time ~/allegro/a.out
user=8.2 sec, system=0.0 sec, elapsed=0:14 min, cpu use=55%, 
         0 swaps, 0 disk reads, 0 disk writes.  Data+stack+text pages: 
         144 avg., 9 max.
tartarus 26 % 

The lisp version of same:

<cl> (time (speed 500))
cpu time (non-gc) 30716 msec user, 333 msec system
cpu time (gc)     7217 msec user, 17 msec system
cpu time (total)  37933 msec user, 350 msec system
real time  79440 msec

NIL 

Our CL version is Allegro CL 3.0.1.beta.sun4.1 [sun4] (8/4/89 11:31).

BTW, the argument to cos in the lisp code is .51 in the code contained
in the mail message rather than -.51 as in the C program. The above CL
run used -.51. There was no substantial difference in time between the
two arguments.

-Jeff Berger
berger@gargoyle.UChicago.edu