I believe I managed to reduce the amount of memory needed for A⌹B
considerably. SVN 791.
On 08/28/2016 12:05 PM, Juergen
thanks, fixed in SVN 790.
You will now get a WS FULL because the
computation of A⌹A needs
a 99999×99999 matrix internally which is about 80 GB large for
160 GB for complex matrices.
On 08/28/2016 04:08 AM, Kacper
Matrix division returns no value for large-shaped arguments.
There doesn't seem to be any fixed threshold above which it fails,
it differs from case to case.