#20803: Wrong kernel computation over finite field on OSX
------------------------------+----------------------------
Reporter: ehlen | Owner:
Type: defect | Status: new
Priority: major | Milestone: sage-7.3
Component: linear algebra | Keywords:
Merged in: | Authors:
Reviewers: | Report Upstream: N/A
Work issues: | Branch:
Commit: | Dependencies:
Stopgaps: |
------------------------------+----------------------------
I ran into a strange problem and it took me very long to find out that
it's probably not a bug in my code but maybe in sage.
I compute a large matrix over a finite field using the same code with sage
7.1 on OSX 10.11.5 and on Linux.
I attached pickles of the two matrices I obtain. They seem identical (I
have made sure that the kernel computations below are not cached in the
pickles).
On Linux, the right kernel of the matrix is correctly computed (and has
dimension 20) whereas on OSX, the result only has dimension 16. It is a
proper subspace of the kernel!
I have no idea what's going on. Can someone help me to debug it?
Here are some facts, it's really a bit confusing.
I ran the following code on OSX:
{{{
sage: U=load('U.sobj')
sage: UX=load('U-osx.sobj')
sage: U == UX
True
sage: V = U.right_kernel(); V
Vector space of degree 613 and dimension 16 over Finite Field of size 101
Basis matrix:
16 x 613 dense matrix over Finite Field of size 101
sage: VX = UX.right_kernel(); VX
Vector space of degree 613 and dimension 16 over Finite Field of size 101
Basis matrix:
16 x 613 dense matrix over Finite Field of size 101
sage: U.right_nullity()
20
sage: UX.right_nullity()
20
}}}
Now, If I take the same objects and run the code on Linux, both of them
give the correct result, a space of dimension 20. Also, If I cache the
result on linux, pickle it and load it in OSX, I get (unsurprisingly) the
correct answer (from cache) but also, I found out that the 16-dimensional
space I'm getting is really a proper subspace of the kernel (and not some
random stuff).
I will also recompile sage from source and see what happens but maybe
someone with OSX can try to reproduce it and maybe
--
Ticket URL: <http://trac.sagemath.org/ticket/20803>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/sage-trac.
For more options, visit https://groups.google.com/d/optout.
