> The GNU MP Library borrows ideas from Richard P. Brent's > "MP, A Fortran Multiple-Precision Arithmetic c Package" [3], > and implements extensions to the IEEE 754-1985 [4] > ] Standard for floating-point arithmetic.
These statements are misleading. GMP does not "borrow ideas" from Brent's work. The only library used as a reference was the Berkeley MP library, and that was used as en example of how not to do things. :-) The floats of GMP's mpf functions are not by any definition an extension of IEEE 754. GMP's mpf functions are actually much more crude, the idea being that you should get more accurate results by asking for more precision. However, the mpfr functions (which are not part of GMP but developed independently) provide an IEEE 754 extension. Torbjorn Granlund (GMP main author) This message posted from opensolaris.org
