Viewing file:  FAQ.html (17.74 KB) -rw-r--r--
Select action/file-type:
 (+) | (+) |  (+) | Code (+) | Session (+) |  (+) | SDB (+) |  (+) |  (+) |  (+) |  (+) |  (+) |
Frequently Asked Questions about GNU MPFR
Frequently Asked Questions about GNU MPFR
Important notice: Problems with a particular version of
MPFR are discussed in the corresponding
bugs page.
The latest version of this FAQ is available at
http://www.mpfr.org/faq.html.
Please look at this version if possible.
- What are the differences between
MPF from GMP
and MPFR?
- How to convert my program written using
MPF to
MPFR?
- At configure time, I get the error:
libgmp not found or uses a different ABI.
- I get undefined reference to
__gmp_get_memory_functions.
- When I link my program with
MPFR, I get undefined reference
to
__gmpXXXX.
- My program crashes with high precisions.
- Though I have increased the precision, the results
are not more accurate.
- How can I detect MPFR
installation using autoconf or pkg-config?
- How to cite MPFR in a
scientific publication?
- When I build MPFR, I get
an error asking me to recompile with -fPIC.
- 1. What are the differences between
MPF from GMP
and MPFR?
The main differences are:
The precision of a MPFR variable
is the exact number of bits used for its mantissa, whereas in
MPF, the precision requested by the user
is a minimum value (MPF generally uses a
higher precision). With the additional difference below, this implies that
the MPFR results do not depend on the
number of bits (16, 32, 64 or more) of the underlying architecture. As a consequence, MPFR uses a
base-2 exponent, whereas in MPF, this
is a base-232 or base-264 exponent, depending on
the limb size. For this reason (and other internal ones), the maximum
exponent range in MPFR is different
(and smaller, if the exponent is represented by the same type as in
MPF). MPFR provides an additional rounding
mode argument to its functions; furthermore, it is guaranteed that the
result of any operation is the nearest possible floating-point value from
the exact result (considering the input variables as exact values), taking
into account the precision of the destination variable and the rounding
mode. MPFR also says whether the rounded
result is above or below the exact result. MPFR supports much more functions
(in particular transcendental functions such as exponentials, logarithms,
trigonometric functions and so on) and special values: signed zeros,
infinities, not-a-number (NaN). - 2. How to convert my program written using
MPF to
MPFR?
You need to add r
to the function names, and to
specify the rounding mode (MPFR_RNDN for rounding to nearest,
MPFR_RNDZ for rounding toward zero, MPFR_RNDU
for rounding toward plus infinity, MPFR_RNDD for rounding
toward minus infinity). You can also define macros as follows:
#define mpf_add(a, b, c) mpfr_add(a, b, c, MPFR_RNDN)
The header file mpf2mpfr.h from the
MPFR distribution automatically
redefines all MPF functions in this
way, using the default MPFR rounding
mode. Thus you simply need to add the following line in all your files
using MPF functions:
#include <mpf2mpfr.h>
just after the gmp.h and mpfr.h
header files. If the program uses MPF
internals (such as direct access to __mpf_struct members),
additional changes will be needed. - 3. At configure time, I get the error:
libgmp not found or uses a different ABI.
This test (checking for __gmpz_init in -lgmp) comes
after the gmp.h detection. The failure occurs either because
the GMP library could not be found
(as it is not in the provided library search paths) or because the
GMP library that was found does not have
the expected ABI
(e.g. 32-bit vs 64-bit). The former problem can be
due to the fact that a static build of MPFR
was requested while only a shared GMP library
is installed (or the opposite, but another error can also show up in this
case, see the question about -fPIC). The
latter problem can have several causes:
- A wrong libgmp library has been picked up. This can occur if you have
several GMP versions installed on the
machine and something is wrong with the provided library search paths.
- Wrong compiler options (CFLAGS) were given. In general, the
presence or absence of the -m64 compiler option must match the
library ABI.
- A wrong gmp.h file has been picked up (if you have several
GMP versions installed). Indeed, by default,
MPFR gets the compiler options from the
gmp.h file (with GMP 4.2.3
or later); this is needed because GMP does
not necessarily use the default ABI. The consequence is
that if the gmp.h file is associated with a library using a
different ABI, the ABI-related options
will be incorrect. Hence the failure.
Note: The config.log output gives more information
than the error message. In particular, see the output of the test:
checking for CC and CFLAGS in gmp.h; it should give you
the default compiler options (from gmp.h).
See also the answer to the next question. - 4. I get undefined reference to
__gmp_get_memory_functions.
Note: this was mainly a problem when upgrading from
GMP 4.1.4 to a later version,
but information given below may still be useful in other cases,
when several GMP libraries are
installed on the same machine.
If you get such an error, in particular when running
make check, then this probably means that you are using
the header file from GMP 4.2.x but the
GMP 4.1.4 library. This can happen if
several GMP versions are installed on
your machine (e.g., one provided by the system in
/usr/{include,lib} and a new one installed by the owner or
administrator of the machine in /usr/local/{include,lib})
and your include and library search paths are inconsistent. On various
GNU/Linux machines, this is unfortunately the case
by default (/usr/local/include is in the default include
search path, but /usr/local/lib is not in the
default library search path). Typical errors are:
undefined reference to `__gmp_get_memory_functions'
in make check. The best solution is to add
/usr/local/include to your C_INCLUDE_PATH
environment variable and to add /usr/local/lib to your
LIBRARY_PATH and LD_LIBRARY_PATH
environment variables (and/or LD_RUN_PATH).
Alternatively, you can use --with-gmp* configure options,
e.g. --with-gmp=/usr/local, but this is
not guaranteed to work (in particular with gcc and
system directories such as /usr or /usr/local),
and other software that uses GMP and/or
MPFR will need correct paths too;
environment variables allow you to set them in a global way.
Other information can be given in the INSTALL file and
ld manual. Please look at them for more details. See also
the next question. - 5. When I link my program with
MPFR, I get undefined reference
to
__gmpXXXX.
Link your program with GMP. Assuming
that your program is foo.c, you should link it using:
cc link.c -lmpfr -lgmp
MPFR library reference (-lmpfr)
should be before GMP's one
(-lgmp). Another solution is, with GNU
ld, to give all the libraries inside a group:
gcc link.c -Wl,--start-group libgmp.a libmpfr.a -Wl,--end-group
See INSTALL file and ld manual for more
details.
If you used correct link options, but still get an error, this may mean
that your include and library search paths are inconsistent. Please see the
previous question. - 6. My program crashes with high precisions.
Your stack size limit may be too small; indeed, by default,
GMP 4.1.4 and below allocates all
temporary results on the stack, and in very high precisions, this
limit may be reached. You can solve this problem in different ways:
You can upgrade to GMP 4.2 (or above),
which now makes temporary allocations on the stack only when they are
small. You can increase the stack size limit with the limit,
unlimit or ulimit command, depending on your
shell. This may fail on some systems, where the maximum stack size cannot
be increased above some value. You can rebuild both GMP and
MPFR to use another allocation method. - 7. Though I have increased the precision, the results
are not more accurate.
The reason may be the use of C floating-point numbers. If you want
to store a floating-point constant to a mpfr_t, you should use
mpfr_set_str (or one of the MPFR
constant functions, such as mpfr_const_pi for π) instead
of mpfr_set_d or mpfr_set_ld. Otherwise the
floating-point constant will be first converted into a reduced-precision
(e.g., 53-bit) binary number before
MPFR can work with it. This is the case
in particular for most exact decimal numbers, such as 0.17, which are
not exactly representable in binary.
Also remember that MPFR does not track
the accuracy of the results: copying a value x to y
with mpfr_set (y, x, MPFR_RNDN) where the variable y
is more precise than the variable x will not make it more
accurate; the (binary) value will remain unchanged. - 8. How can I detect MPFR
installation using autoconf or pkg-config?
The MPFR team does not currently
recommend any autoconf code, but a section will later
be added to the MPFR manual. The
MPFR team does not wish to support
pkg-config yet. - 9. How to cite MPFR in a
scientific publication?
To properly cite MPFR in a scientific
publication, please cite the
ACM
TOMS
paper
(BibTeX)
and/or the library web page
http://www.mpfr.org. If your publication
is related to a particular release of MPFR,
for example if you report timings, please also indicate the release number
for future reference. - 10. When I build MPFR, I get
an error asking me to recompile with -fPIC.
A typical error looks like:
/usr/bin/ld: /path/to/libgmp.a(realloc.o): relocation
R_X86_64_32 against `.rodata.str1.1' can not be used when making a
shared object; recompile with -fPIC
/path/to/libgmp.a: could not read symbols: Bad value
collect2: ld returned 1 exit status
The probable reason is that you tried to build
MPFR with the shared library enabled (this
is the default), while only a static GMP
library could be found. To solve this problem, either rebuild and reinstall
GMP without the --disable-shared
configure option, or configure MPFR with
--disable-shared. If you did this and still get the above
error, the cause may be conflicting GMP
versions installed on your system; please check that your search path
settings are correct.
Additional note about the last sentence: Under GNU/Linux
(for instance), the linker takes the first library found in the library search
path, whether it is dynamic or static. The default behavior under darwin is
different, but MPFR will change it.
|