[Maxima-bugs] [ maxima-Bugs-1480338 ] ratsimp(exp(constant)) very slow

[SourceForge.net <no...@so...>]

Bugs item #1480338, was opened at 2006-05-02 07:59
Message generated for change (Comment added) made by rtoy
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=1480338&group_id=4933

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Lisp Core
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Barton Willis (willisbl)
Assigned to: Nobody/Anonymous (nobody)
Summary: ratsimp(exp(constant)) very slow

Initial Comment:
ratsimp, when applied to 

%e^((sqrt(852469675641479773175661572149741)
-15762598695796738)/2251799813685248);

is very slow. But can ratsimp simplify the argument
to exp quickly. 

This expression happended towards the end of a
matrixexp calculation. I've tried variations of this
expression (delete hunks of various numbers) --
sometimes ratsimp 
is very fast, other times it seems to hang.

(%i18) build_info();
Maxima version: 5.9.3
Maxima build date: 0:52 3/20/2006
host type: i686-pc-mingw32
lisp-implementation-type: GNU Common Lisp (GCL)
lisp-implementation-version: GCL 2.6.7

Barton

----------------------------------------------------------------------

>Comment By: Raymond Toy (rtoy)
Date: 2006-05-08 10:47

Message:
Logged In: YES 
user_id=28849

Interesting.  Barton's simplified ratsimp example causes 2
errors with CMUCL because pcetimes1 does f+ with a bignum
and palgsimp does f< with a bignum.  Replacing these with +
and < fixes the issue.

However, the ratsimp example causes a warning from cmucl
because it wants to compute 1^1932473987149817.  This is a
cmucl problem, but depending on how smart the Lisp is, it
might actually try to compute that power.

Finally, for the complicated ratsimp(exp((sqrt ...)))
example, I think the real issue is that maxima is trying to
simplify the sqrt.  If you just enter

sqrt(852469675641479773175661572149741);

there's a delay before maxima prints out the warning about
not finding the factors.

Also, factor(852469675641479773175661572149741) spends some
time and returns 2 factors: 34130646845983 24976663333933005427.

Not sure why there's a warning at all.

So I think it used to be fast because sqrt(852...) factor or
somebody gave up early and returned sqrt(852...).  ifactor
works harder before giving up.

----------------------------------------------------------------------

Comment By: Robert Dodier (robert_dodier)
Date: 2006-05-06 23:03

Message:
Logged In: YES 
user_id=501686

Not sure what's going on here, but maybe the problem is in
the factoring code.

Maxima 5.9.3 / clisp 2.34:
ratsimp (%e^((sqrt(852469675641479773175661572149741)
-15762598695796738)/2251799813685248)); 
 => fast

Maxima 5.9.3 / clisp 2.38 (after factor code was revised):
 => slow, and it complains 
"WARNING: could not find factors of composite:
852469675641479773175661572149741"



----------------------------------------------------------------------

Comment By: Barton Willis (willisbl)
Date: 2006-05-06 07:40

Message:
Logged In: YES 
user_id=895922

A simpler example:

   ratsimp(exp((1932473987149817)/589347638476187146))

Barton

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=1480338&group_id=4933