[CLN-list] Floating point underflow.

[Bruno Haible]

Hi,

> I wrote some code using the Boost matrix library and CLN. But I get
> 'floating point underflow' errors every time.

The documentation section Customizing / Floating-point underflow
answers this question.

But since floating point underflow means a 100% rounding error, you
might afterwards run the computation with larger precision, to see if
these rounding errors are harmful (produce really wrong results) or not.

Bruno