Re: Octave/C++ matrix Inv() comparison

[CdeMills]

I computed the 20x20 version of  your matrix; as I suspected it is very badly
scaled and is nearly impossible to invert in 64bits precision. Whatever the
method you'll get wrong results.

>From your script, I tried the multi-precision routine:

Rm = mp(R);
Sm = inv(Rm);
Ierr = Sm*Rm-eye(size(R));
In=double(max(max(abs(Ierr))))
In =  1.0335e-59

I enclose the Sm matrix Values are comprised betwwen 1e9 and 1e13.

Regards

Pascal

ionone.txt <http://octave.1599824.n4.nabble.com/file/n4655529/ionone.txt>  






--
View this message in context: 
http://octave.1599824.n4.nabble.com/Octave-C-matrix-Inv-comparison-tp4655291p4655529.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.