Skip to content
Prev 7028 / 63421 Next

Matrix identification bug (PR#1361)

On Tue, 5 Mar 2002 hyu@stats.uwo.ca wrote:

            
[...]
Nothing in this report suggests a bug in solve().  Can you please provide
some evidence for your claim (which is not `clear' to me), and
function (a, b, ...)
UseMethod("solve")

seems hard to get wrong!  (It matters: solve is part of R: solve.Matrix is
a contributed addon.)

Incidentally, these sort of timings are crucially dependent on the BLAS
library used, and in my experience that is a far larger factor that the OS
used.  However, that was not part of the report ....