Matlab data type implementation & speed

[uzil24]

Hi all,
I have a great deal of Matlab code which I'm required to implement in JAVA.
Since the code is complicated and far from being OO I tried to create a Matlab data type in JAVA. after doing so I simply translated the code using the data type.
The problem is run time.
The same code takes a fraction of a second in Matlab and almost 2 minutes(!!!) in JAVA. I used the "ojalgo" package with my implementation, so most of the code should be efficient.
why is this happening? Am I too naive to try this?
Does Matlab have some special "trick" to hasten it's running time?
Any insights will be highly appreciated.

[r035198x]

It depends on the functions you are talking and how you have implemented them.

[uzil24]

It depends on the functions you are talking and how you have implemented them.

How do you propose I do it?
Would you do something like that or try a different approach?

[r035198x]

Why do the functions need to be reimplemented?
I would take a look at apache's commons-net project if I were you.