Hi, I wrote some code using the Boost matrix library and CLN. But I get 'floating point underflow' errors every time. I thought by using CLN, I would not have problems with very small and very large numbers. Is there anything I can do.. I really need this to work! Sincerely, Elvis Dieguez