Owen Green

Maybe I'm being really dim here, but is the behaviour of jit.la.mult with respect to its inputs the reverse of matlab's?

That is, in matlab if A is n-by-p and B is p-by-m then A*B gives an n*m matrix.

I'd guess I'd expected A to map to inlet 1, B to inlet 2, but it seems to be the other way round:

Presumably this is intended behaviour – is there any chance the docs could be updated to make it clearer?

Maybe I'm being really dim here, but is the behaviour of jit.la.mult with respect to its inputs the reverse of matlab's?  

That is, in matlab if A is n-by-p and B  is p-by-m then A*B gives an n*m matrix. 

Presumably this is intended behaviour – is there any chance the docs could be updated to make it clearer?


jit.la.mult input ordering