Hello,
I tried one of the tests proposed by Prof. Fateman in 'Speeding up
Lisp-based Symbolic Mahematics' - computing sum(a[i],i,1,N) for
different values of N. The timings on my PII-233 PC are ([N,time])
[[100,
0.17],
[200, 0.43],
[300, 1.87],
[400, 2.53],
[500, 3.45],
[600, 4.82],
[700, 6.58],
[1000, 14.1],
[1200, 22.02],
[1500, 32.18],
[2000, 64.37],
[2200, 83.27]],
I think it's still O(n^2). build_info() is
Maxima version: 5.9.0
Maxima build date: 19:10 2/9/2003
host type: i686-pc-mingw32
lisp-implementation-type: Kyoto Common Lisp
lisp-implementation-version: GCL-2-5.0
Then I tried a[i]=f(a[i-1]) stuff, literally
(C39) a[0]:x$
(C40) for i:1 thru 3000 do a[i]:f(a[i-1])$
and got
(C41) a[345]
;
Evaluation took 0.00 seconds (0.00 elapsed)
(D41)
f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f
(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f(f
Error: Bind stack overflow.
Fast links are on: do (si::use-fast-links nil) for debugging
Error signalled by CATCH.
Broken at MACSYMA-TOP-LEVEL. Type :H for Help.
MAXIMA>>:q
Computing sum(a[i],i,1,2200)$ didn't stop in few minutes.
So, my question is: are the changes proposed in the above article added
to the current Maxima?
--
Andrei Zorine