What if I base the assignment of an mpf_class variable on another mpf_class variable, like so: mpf_class a, b; a="1.123456789123456789123456789"; b=a*2; Would it still round it incorrectly? <div class="gmail_quote">On Mon, Jun 1, 2009 at 8:38 PM, Marc Glisse <<a href="mailto:marc.glisse@normalesup.org">marc.glisse@normalesup.org</a>> wrote: <blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> <div class="im">On Mon, 1 Jun 2009, Henry wrote: <blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> mpf_class x; x=1.234567891234567899052395; </blockquote> </div> This is a general C question. When the compiler sees 1.234567891234567899052395, it thinks this is a double and rounds it accordingly. I think you actually want to go though a string: x="1.23..."; In a few years, with C++0x, you may be able to write 1.234_mpf instead of mpf_class("1.234"), but that is still far away. -- Marc Glisse </blockquote></div> -- Henry