Are you currently reading using csvread(), or import(), or load() ?
If each file has exactly the same structure (same number of fields, some number of headers, same fields are character vs numeric) then you might be able to do better by using textscan()
If you are currently using xlsread() on MS Windows then you can almost certainly do better than you are doing now (but textscan() might be better still). xlsread() opens and closes Excel each time; Image Analyst has posted code several times showing how to open Excel just once and do several operations with it.
