Websites on Increasing Precision: Double-double precision Fortran 90 code by Claire Moreau-Finot for double-double precision arithmetic. That is about a hundred bits of precision with a floating point notation. The library computes the exponential function radix 2 and e, the cosine and the sine. FMLIB Multiple precision package David Smith's package for multi-precision arithmetic. Unlike most other packages, the components of these large numbers are stored as REALs. This is usually much more efficient than storing them as integers. This is TOMS algorithm 786. High-Precision Software Directory David Bailey's package for multiple precision, and packages for double-double (~32 dec. digits) and quad-double (~64 dec. digits) precision. The last two are in C++ with Fortran front ends. MPFUN Multiple precision arithmetic. Multiple precision arithmetic This is Richard Brent's classic MP package, which was published as TOMS algorithm 524. Contains some features such as the Bernoulli numbers, which are not available in either Smith or Bailey's packages. Dates from 1981. Quadruple precision for the NAS FortranPlus compiler. Using the 10-byte format available with Intel processors, this module gives about 38 decimal digits accuracy.
|
|