[Maxima-bugs] [ maxima-Bugs-2935631 ] bfloat(log(n)) can be different from log(bfloat(n))

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

Bugs item #2935631, was opened at 2010-01-20 09:01
Message generated for change (Settings changed) made by rtoy
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=2935631&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 - Floating point
Group: None
>Status: Closed
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: Raymond Toy (rtoy)
Assigned to: Nobody/Anonymous (nobody)
Summary: bfloat(log(n)) can be different from log(bfloat(n))

Initial Comment:
bfloat(log(n)) uses logbigfloat, which calls fplog to compute the result.

log(bfloat(n)) uses big-float-log to compute the result using the property log(f*2^m) = log(f)+m*log(2).

Since the algorithms are different, the result of bfloat(log(n)) and log(bfloat(n)) could be different.  I suggest changing logbigfloat to call big-float-log so that the results are guaranteed to be the same.

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

>Comment By: Raymond Toy (rtoy)
Date: 2010-01-20 19:48

Message:
Fixed.  logbigfloat calls big-float-log as needed.

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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=2935631&group_id=4933