slow matrix multiply

Henry Cejtin henry@sourcelight.com
Tue, 10 Jul 2001 20:37:56 -0500


In general, and it really depends, the 'a array implementation should be
faster: it is trading a multiply for an indirection.  And in normal mode,
the indirection is going to cost you a check on the subscript.  Also it
costs you space.  This can be a very high price if the array is long and
narrow.