FS#2213 - [octave 2.1.64-1] basic computations fail

Attached to Project: Arch Linux
Opened by jan willem (jw) - Tuesday, 15 February 2005, 19:21 GMT
Last edited by Damir Perisa (damir.perisa) - Wednesday, 16 February 2005, 00:41 GMT
Task Type Bug Report
Category Packages: Extra
Status Closed
Assigned To Damir Perisa (damir.perisa)
Architecture not specified
Severity High
Priority Normal
Reported Version 0.7 Wombat
Due in Version Undecided
Due Date Undecided
Percent Complete 0%
Votes 0
Private No

Details

Problem: Example code does not finish computation in any reasonable amount of time

> a = rand(2,1); b=rand(3,2);
> c = b*a;
> d = b\c;

The calculation finishes just fine if aborted with ctrl+c and tried again afterwards. However, this bug does break scripts, since the first calculation of this type will never finish.

The problem is known on the octave mailing lists. Have a look at these two threads:

http://www.octave.org/octave-lists/archive/bug-octave.2005/msg00105.html
http://www.octave.org/octave-lists/archive/octave-maintainers.2004/msg00817.html

The solution is to compile octave using g77 with FFLAGS="-O -ffloat-store". More specifically, I am told that the "-ffloat-store" needs to be enabled for the following four files:

dcabs1.f
dlamch.f
slamch.f
dgeev.f

This requires recompilation of gcc, however, so I will leave it up to you how to deal with this.

I would like to stress that the severity of this bug is high, if not critical, as it causes octave to lock up under very basic (and common) computations.

Thanks in advance for looking at this,
Jan-Willem van de Meent
Closed by  Jan de Groot (JGC)
Wednesday, 23 February 2005, 11:55 GMT
Reason for closing:  Duplicate
Additional comments about closing:  Task 2060 is almost the same, closing dupe (fixed too :P)
Comment by Damir Perisa (damir.perisa) - Tuesday, 15 February 2005, 20:56 GMT
this would mean that we need to have gcc-g77 in the repos

Loading...