Octave gives wrong results when diagonalizing large matrices with MKL
Affects  Status  Importance  Assigned to  Milestone  

Octave 
Unknown

Unknown


octave (Ubuntu) 
New

Undecided

Unassigned 
Bug Description
Octave gives wrong results when manipulating large matrices, when the package `intelmkl` is installed, and BLAS, LAPACK are configured with `intelmkl` (from default ubuntu repositories). With MKL, Octave seems to have a 2x speedup when diagonalize certain matrices, but the results are totally wrong (when the matrix is large, say 1000x1000).
Code:
```
for a = 1:500
for b = 1:500
c(a,b) = sin(a + b^2);
endfor
endfor
g = eig(c);
max(real(g))
```
The correct result is `ans = 16.915` (I checked this before installing intelmkl, and also in Ubuntu 18.04 in the same computer, where intelmkl was not installed)
With MKL in Ubuntu 20.04, I get random numbers of order 10^5  10^6, which change every time the code is run.
lsb_release rd
Description: Ubuntu 20.04 LTS
Release: 20.04
(It is Xubuntu 20.04)
aptcache policy octave
octave:
Installed: 5.2.01
Candidate: 5.2.01
Version table:
*** 5.2.01 500
500 http://
100 /var/lib/
aptcache policy intelmkl
intelmkl:
Installed: 2020.0.1661
Candidate: 2020.0.1661
Version table:
*** 2020.0.1661 500
500 http://
100 /var/lib/
I followed https://askubuntu.com/a/1265802/124466 to install Intel MKL.