lu_factor regression

Computer Algebra System written in Common Lisp

Brought to you by: kjak, l_butler, macrakis, robert_dodier, and 2 others

#2602 lu_factor regression

Milestone: None

Status: closed

Owner: nobody

Labels: regression (1)

Priority: 7

Updated: 2013-07-03

Created: 2013-07-01

Creator: roucarb

Private: No

The following file is read ok with maxima 5.27.0 but fail with 5.30.0 with:
lu:lu_factor(y0,crering)$

Warning:
WITH-COMPILATION-UNIT is being redefined.
Maxima encountered a Lisp error:

Error in MFUNCALL [or a callee]: 1 is not of type LIST.

1 Attachments

test2.max

Discussion

Rupert Swarbrick - 2013-07-01

status: open --> closed
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Rupert Swarbrick - 2013-07-01

This was caused by a a function called some-bfloatp that was recursing wrongly on rational expressions. I've just pushed a fix for that bug as commit [24090c]. With your test-case, the lu_factor call now just takes an awfully long time (I didn't wait for it to finish). I don't know whether that's a regression or whether it was always a big computation. If the former, then maybe you want to open a new bug for that: performance timings would probably be very helpful.

Related

Commit: [24090c]

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

roucarb - 2013-07-03

Hi,

Thanks for the patch.

Previously it take about 20 minutes to run this file on a xeon at 3Ghz.

If I do not use crering, it run for a month without result

Bastien

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

roucarb - 2013-07-03

I will test ASAP

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Log in to post a comment.