MPFR version 2.1.0 (released on 3 November 2004)

MPFR 2.1.0 requires GMP 4.1.0 or higher.

Download

MPFR 2.1.0 can be downloaded in the following archive formats:

• mpfr-2.1.0.tar.bz2 (tar/bzip2, 481 KB),
• mpfr-2.1.0.tar.gz (tar/gzip, 605 KB),
• mpfr-2.1.0.zip (zip, 853 KB).

Note that the patches below have not been applied.

MPFR is also available as the official Debian package libmpfr-dev. Patches 1 to 13 below have been applied in libmpfr-dev 2.1.0-5. Patch 14 below have been applied in libmpfr-dev 2.1.0-6. Patches 15 and 16 below have been applied in libmpfr-dev 2.1.0-7.

Documentation

• The reference manual: [html] [dvi] [ps] [pdf]
• The reference manual translated into Japanese.

Bugs

The bugs listed below have been fixed (latest update: 2005-02-21). The following instructions are for Linux and may be similar for other operating systems. You can apply the patches in several ways:

• Apply the cumulative patch on the 2.1.0 release or a partially patched version with the command

  patch -N -Z -p1 < path_to_patches_file

  from the build directory. The -N option allows the patch command to ignore patches already applied. The -Z option sets the modification time of the patched files from time stamps given in the patch file, thus avoiding the need of some development utilities (such as autoconf).
• Or apply the individual patches below.
• Or checkout the CVS branch mpfr-2-1-branch thanks to the option -rmpfr-2-1-branch. In this case, please see the MPFR CVS page for more information.

Fixed bugs:

• Minor corrections in the manual: documentation patch.
• Better way to detect the GMP library: configure patch.
• Bugs in mpfr_get_str, in particular for huge exponents (greater than 2³⁰ on 64-bit machines): get_str patch.
• The mpfr-longlong.h file leads to a compilation problem on Opteron, when MPFR is compiled without the internal GMP files (recommended way): longlong.h patch.
• A suboptimal function declaration prevents from compiling programs including term.h then mpfr.h, due to a clash between these header files: mpfr.h patch.
• Huge performance problem in corner cases for the subtraction (mpfr_sub1 internal function): first sub1 patch.
• Bug in mpfr_sum for table size less or equal to 1: sum patch 1.
• Some conditional header inclusions were forgotten in the tests: tests patch.
• Bugs in markers (may occur only if you configured MPFR to replace alloca by something else): marker patch.
• The tests could not be compiled with a C++ compiler, such as g++ (due to implicit pointer conversions). The following patch adds the necessary casts: casts for C++ compiler patch.
• The mpfr_rint (in the rounding-to-nearest mode) and mpfr_round functions may round in the wrong direction when rounding in place (source and destination are the same pointer): mpfr_rint patch (with testcase).
• When the input value of mpfr_exp is very small, the sign of the result may be wrong (if the old destination was negative): mpfr_exp patch (with testcase).
• The above sub1 patch introduced a bug. The following patch fixes it (apply it over a version with the first sub1 patch): second sub1 patch (with testcase).
• Another bug in mpfr_sum when the destination has a different precision from the first element of the list: sum patch 2.
• The mpfr_tan function has an overflow bug, which makes the following program loop forever:

  #include <stdio.h>
  #include <mpfr.h>
  int main (void)
  {
    MPFR_DECL_INIT (x, 53);
    mpfr_set_emax (4);
    mpfr_const_pi (x, GMP_RNDD);
    mpfr_div_2ui (x, x, 1, GMP_RNDD);
    mpfr_tan (x, x, GMP_RNDD);
    mpfr_dump (x);
    return 0;
  }

  The following patch fixes this bug: mpfr_tan patch.
• The first sub1 patch introduced another bug. The following patch fixes it (apply it over a version with the first sub1 patch): third sub1 patch.
• The function mpfr_ui_sub applied to the integer 0 and the floating-point number +0 gives +0 instead of −0. This bug was introduced in MPFR 2.1.0, thus it is not present in previous MPFR versions. It is fixed by the following patch: mpfr_ui_sub signed-zero patch.

In addition to the bugs mentioned in the BUGS file, there are possible bugs with huge precisions (greater than 2³⁰) and with exponent ranges that do not allow to represent some numbers, such as 1 or 1/2.

Changes from version 2.0.3 to version 2.1.0

• Bug fixes.
• New functions mpfr_strtofr, mpfr_set_uj, mpfr_set_sj, mpfr_set_ui_2exp, mpfr_set_si_2exp, mpfr_set_uj_2exp, mpfr_set_sj_2exp, mpfr_get_uj, mpfr_get_sj, mpfr_get_z, mpfr_free_str, mpfr_si_sub, mpfr_sub_si, mpfr_mul_si, mpfr_si_div, mpfr_div_si, mpfr_sqr, mpfr_cmp_z, mpfr_cmp_q, mpfr_zero_p, mpfr_free_cache, mpfr_sum, mpfr_get_version, mpfr_get_default_rounding_mode, mpfr_get_emin_min, mpfr_get_emin_max, mpfr_get_emax_min, mpfr_get_emax_max, mpfr_inits, mpfr_inits2, mpfr_clears, mpfr_fits_intmax_p, mpfr_fits_uintmax_p, mpfr_clear_erangeflag, mpfr_erangeflag_p, mpfr_rint_round, mpfr_rint_trunc, mpfr_rint_ceil, mpfr_rint_floor.
• New macros MPFR_DECL_INIT, MPFR_VERSION, MPFR_VERSION_NUM, MPFR_VERSION_MAJOR, MPFR_VERSION_MINOR, MPFR_VERSION_PATCHLEVEL.
• Improved documentation.
• Improved configure.
• Improved portability (library and test suite).
• It handles correctly non IEEE-754 double.
• GMP internal files are no longer needed to install MPFR.
• It is faster with low-precision floating point.
• New global flag: ERANGE_FLAG.
• Binary incompatible with previous versions, but API compatible.
• mpfr_set_str doesn't allow @NAN@garbagechar and @INF@garbagechar anymore, allows base 0 (detection of the base), prefix (0x, 0b), leading whitespace.

Back to the MPFR page.