id,summary,reporter,owner,description,type,status,priority,milestone,component,resolution,keywords,cc,work_issues,upstream,reviewer,author,merged,dependencies,stopgaps
9760,Possible numerical noise doctest failure in sage/matrix/matrix2.pyx on t2,mpatel,drkirkby,"Reported by John Palmieri on [http://groups.google.com/group/sage-release/browse_thread/thread/67feb1b840ff2710/90a82acffe8f10de#90a82acffe8f10de sage-release]:
{{{
On t2, I'm getting a failure:

sage -t  -long devel/sage/sage/matrix/matrix2.pyx
**********************************************************************
File ""/home/palmieri/t2/sage-4.5.3.alpha1/devel/sage-main/sage/matrix/
matrix2.pyx"", line 6406:
    sage: all(imag(e) < 1e-15 for e in eigs)
Expected:
    True
Got:
    False

This looks like numerical noise: if I replace 1e-15 by 1e-14, it passes.  Has anyone else seen this?  Could it possibly be dependent on ATLAS (e.g., the load on the system when it's tuning itself)? 
}}}
",defect,closed,blocker,sage-4.5.3,porting: Solaris,fixed,,jhpalmieri,,N/A,David Kirkby,John Palmieri,sage-4.5.3.alpha2,,