NAME

``   Math::GMPf - perl interface to the GMP library's floating point (mpf) functions.``

DEPENDENCIES

``````   This module needs the GMP C library - available from:
http://gmplib.org``````

DESCRIPTION

``````   A bigfloat module utilising the Gnu MP (GMP) library.
Basically this module simply wraps all of the 'mpf'
floating point functions provided by that library.
The documentation below extensively plagiarises the
GMP documentation at http://gmplib.org .
See the Math::GMPf test suite for some examples
of usage.``````

SYNOPSIS

``````   use Math::GMPf qw(:mpf);

my \$string = '.123542@2'; # mantissa = (.)12345
# exponent = 2
# my \$string = '12.354'; # alternative string format

my \$base = 10;

# Set the default precision to at least 80 bits.
Rmpf_set_default_prec(80);

# Create the Math::GMPf object
my \$bn1 = Rmpf_init_set_str(\$string, \$base);

# Create another Math::GMPf object that holds
# an initial value of zero, but with at least
# 131 bits of precision.
my \$bn2 = Rmpf_init2(131);

# Create another Math::GMPf object that holds
# an initial value of zero, with default precision.
my \$bn3 = Rmpf_init();

# Or just use the new() function:
my \$bn4 = Math::GMPf->new(116.8129);

# Perform some operations ... see 'FUNCTIONS' below.

.
.

# print out the value held by \$bn1 (in octal):
print Rmpf_get_str(\$bn1, 8, 0), "\n";

# print out the value held by \$bn1 (in decimal):
print Rmpf_get_str(\$bn1, 10, 0);

# print out the value held by \$bn1 (in base 29)
# using the (alternative) TRmpf_out_str()
# function. (This function doesn't print a newline.)
TRmpf_out_str(*STDOUT, 29, 0, \$bn1);``````

MEMORY MANAGEMENT

``````   Objects created with Rmpf_init* functions have been
blessed into package Math::GMPf. They will
therefore be automatically cleaned up by the
DESTROY() function whenever they go out of scope.

For each Rmpf_init* fnction there is a corresponding
Rmpf_init*_nobless function. If you wish you can
create unblessed objects using these functions.
It will then be up to you to clean up the memory
associated with these objects by calling
Rmpf_clear(\$op), for each object. Alternatively the objects
will be cleaned up when the script ends. I don't know
why you would want to create unblessed objects. The
point is that you can if you want to.  ``````

FUNCTIONS

``````   See the GMP documentation at http://gmplib.org

These next 3 functions are demonstrated above:
\$rop   = Rmpf_init_set_str(\$str, \$base); # 1 < abs(\$base) < 63
\$rop   = Rmpf_init2(\$bits); # \$bits > 0
\$str = Rmpf_get_str(\$r, \$base, \$digits); # 1 < abs(\$base) < 63
The third argument to Rmpf_get_str() specifies the number
of digits required to be output. Up to \$digits digits
will be generated.  Trailing zeros are not returned.  No
more digits than can be accurately represented by OP are
ever generated.  If \$digits is 0 then that accurate
maximum number of digits are generated.

The following functions are simply wrappers around a GMP
function of the same name. eg. Rmpf_swap() is a wrapper around
mpf_swap() which is fully documented in the GMP manual at
http://gmplib.org.

"\$rop", "\$op1", "\$op2", etc. are simply Math::GMPf objects
- the return value of one of the Rmpf_init* functions
(or their '_nobless' counterpart).
They are in fact references to GMP structures.
The "\$rop" argument(s) contain the result(s) of the calculation
being done, the "\$op" argument(s) being the input(s) into that
calculation.
Generally, \$rop, \$op1, \$op2, etc. can be the same perl variable,
though usually they will be distinct perl variables referencing
distinct GMP structures.
Eg. something like Rmpf_add(\$r1, \$r1, \$r1),
where \$r1 *is* the same reference to the same GMP structure,
would add \$r1 to itself and store the result in \$r1. Think of it
as \$r1 += \$r1. Otoh, Rmpf_add(\$r1, \$r2, \$r3), where each of the
arguments is a different reference to a different GMP structure
would add \$r2 to \$r3 and store the result in \$r1. Think of it as
\$r1 = \$r2 + \$r3. Mostly, the first argument is the argument that
stores the result and subsequent arguments provide the input values.
Exceptions to this can be found in some of the functions that
actually return a value.
Like I say, see the GMP manual for details. I hope it's
intuitively obvious or quickly becomes so. Also see the test
suite that comes with the distro for some examples of usage.

"\$ui" means any integer that will fit into a C 'unsigned long int'.

"\$si" means any integer that will fit into a C 'signed long int'.

"\$double" means any number (not necessarily integer) that will fit
into a C 'double

"\$bool" means a value (usually a 'signed long int') in which
the only interest is whether it's true or false.

"\$str" simply means a string of symbols that represent a number,
eg "1234567890987654321234567@7" which might be a base 10 number,
or "zsa34760sdfgq123r5@11" which would have to represent a base 36
number (because "z" is a valid digit only in base 36). Valid
bases for GMP numbers are 2 to 62 (inclusive).

########################

INITIALIZATION FUNCTIONS

Normally, a variable should be initialized once only or at least be
cleared, using `Rmpf_clear', between initializations.
'DESTROY' (which calls 'Rmpf_clear') is automatically called on
blessed objects whenever they go out of scope.

First read the section 'MEMORY MANAGEMENT' (above).

\$bits = Rmpf_get_default_prec();
Return the current default default precision.

Rmpf_set_default_prec(\$bits);
Set the default precision to be *at least* \$bits bits.  All
subsequent calls to `Rmpf_init' will use this precision, but
previously initialized variables are unaffected.

\$rop = Math::GMPf::new();
\$rop = Math::GMPf->new();
\$rop = new Math::GMPf();
\$rop = Rmpf_init();
\$rop = Rmpf_init_nobless();
Initialize \$rop to 0. The precision of \$rop is undefined
unless a default precision has already been established by
a call to `Rmpf_set_default_prec'.

\$rop = Rmpf_init2(\$bits);
\$rop = Rmpf_init2_nobless(\$bits);
Initialize \$rop to 0 and set its precision to be
*at least* \$bits bits.

\$bits = Rmpf_get_prec(\$op);
Return the current precision of \$op, in bits.

Rmpf_set_prec(\$rop, \$bits);
Set the precision of \$rop to be *at least* \$bits bits.
The value in \$rop will be truncated to the new precision.
This function requires internal reallocation of memory,
and so should not be used in a tight loop.

Rmpf_set_prec_raw(\$rop, \$bits);
Set the precision of \$rop to be *at least* \$bits bits, without
changing the memory allocated. \$bits must be no more than the
allocated precision for \$rop, that being the precision when \$rop
was initialized, or in the most recent `Rmpf_set_prec'.
The value in \$rop is unchanged, and in particular if it had a
higher precision than \$bits it will retain that higher precision
New values written to \$rop will use the new value \$bits.
Before calling `Rmpf_clear' (which will happen when a blessed
Math::GMPf object goes out of scope) or the full `Rmpf_set_prec',
another `Rmpf_set_prec_raw' call must be made to restore \$rop to
its original allocated precision.  Failing to do so will have
unpredictable results.
`Rmpf_get_prec' can be used before `Rmpf_set_prec_raw' to get the
original allocated precision.  After `Rmpf_set_prec_raw' it
reflects the \$bits value set.
`Rmpf_set_prec_raw' is an efficient way to use a Math::GMPf
object at different precisions during a calculation, perhaps to
gradually increase precision in an iteration, or just to use
various different precisions for different purposes during a
calculation.

####################

ASSIGNMENT FUNCTIONS

Rmpf_set(\$rop, \$op);
Rmpf_set_ui(\$rop, \$ui);
Rmpf_set_si(\$rop, \$si);
Rmpf_set_d(\$rop, \$double);
Rmpf_set_z(\$rop, \$z); # \$z is a Math::GMPz object.
Rmpf_set_q(\$rop, \$q); # \$q is a Math::GMPq object.
Set the value of \$rop from the 2nd arg.

Rmpf_set_str(\$rop, \$str, \$base);
Set the value of \$rop from the string in \$str. The string is of
the form `M@N' or, if the base is 10 or less, alternatively
`MeN'. `M' is the mantissa and `N' is the exponent. The mantissa
is always in the specified base. The exponent is either in the
specified base or, if base is negative, in decimal.
The argument \$base may be in the ranges 2 to 62, or -62 to -2.
Negative values are used to specify that the exponent is in
decimal. For bases up to 36, case is ignored; upper-case and
lower-case letters have the same value; for bases 37 to 62,
upper-case letter represent the usual 10..35 while lower-case
letter represent 36..61.
Unlike the corresponding mpz function, the base will not be
determined from the leading characters of the string if base is 0.
This is so that numbers like `0.23' are not interpreted as octal.
This function croaks if the entire string is not a valid number
in base \$base.

Rmpf_swap(\$rop1, \$rop2);
Swap \$rop1 and \$rop2. Both the values and the
precisions of the two variables are swapped.

######################################

COMBINED INITIALIZATION AND ASSIGNMENT

NOTE: Do NOT use these functions if \$rop has already
been initialised. Instead use the Rmpz_set* functions
in 'Assignment Functions' (above)

First read the section 'MEMORY MANAGEMENT' (above).

\$rop = Math::GMPf->new(\$arg);
\$rop = Math::GMPf::new(\$arg);
\$rop = new Math::GMPf(\$arg);
Returns a Math::GMPf object with the value of \$arg, with default
precision. \$arg can be either a number (signed integer, unsigned
integer, signed fraction or unsigned fraction) or a string that
represents a numeric value. If \$arg is a string, an optional
additional argument that specifies the base of the number can
be supplied to new(). If \$arg is a string and no additional
argument is supplied, base 10 is assumed.

\$rop = Rmpf_init_set(\$op);
\$rop = Rmpf_init_set_nobless(\$op);
\$rop = Rmpf_init_set_ui(\$ui);
\$rop = Rmpf_init_set_ui_nobless(\$ui);
\$rop = Rmpf_init_set_si(\$si);
\$rop = Rmpf_init_set_si_nobless(\$si);
\$rop = Rmpf_init_set_d(\$double);
\$rop = Rmpf_init_set_d_nobless(\$double);
Initialise \$rop and assign to it the value held by
the functions argument. See the 'Rmpf_set*'
functions above.

\$rop = Rmpf_init_set_str(\$str, \$base);
\$rop = Rmpf_init_set_str_nobless(\$str, \$base);
Initialise \$rop and assign to it the base \$base
value represented by \$str. See the 'Rmpf_set_str'
documentation above for details.

####################

CONVERSION FUNCTIONS

\$double = Rmpf_get_d(\$op);
Convert \$op to a 'double'.

\$si = Rmpf_get_si(\$op);
\$ui = Rmpf_get_ui(\$op);
Convert \$op to a `signed long' or `unsigned long',
truncating any fraction part.  If \$op is too big for
the return type, the result is undefined.

(\$double, \$exp) = Rmpf_get_d_2exp(\$op);
Find \$double and \$exp such that \$double * (2 ** \$exp),
with 0.5<=abs(\$double)<1, is a good approximation to \$op.
This is similar to the standard C function `frexp'.

\$str = Rmpf_get_str(\$op, \$base, \$digits);
Convert \$op to a string of digits in base \$base. \$base can
be 2 to 62.  Up to \$digits digits will be generated.
Trailing zeros are not returned.  No more digits than can be
accurately represented by \$op are ever generated.  If \$digits
is 0 then that accurate maximum number of digits are generated.

(\$man, \$exp) = Rmpf_deref2(\$op, \$base, \$digits);
Returns the mantissa to \$man (as a string of digits, prefixed with
a minus sign if \$op is negative), and returns the exponent to \$exp.
There's an implicit decimal point to the left of the first digit in
\$man. The third argument to Rmpf_deref2() specifies the number of
digits required to be output in the mantissa. No more digits than
can be accurately represented by \$op are ever generated. If \$digits
is 0 then that accurate maximum number of digits are generated

####################

ARITHMETIC FUNCTIONS

\$rop = 2nd arg + 3rd arg.

Rmpf_sub(\$rop, \$op1, \$op2);
Rmpf_sub_ui(\$rop, \$op, \$ui);
Rmpf_ui_sub(\$rop, \$ui, \$op);
\$rop = 2nd arg - 3rd arg.

Rmpf_mul(\$rop, \$op1, \$op2);
Rmpf_mul_ui(\$rop, \$op, \$ui);
\$rop = 2nd arg * 3rd arg.

Rmpf_div(\$rop, \$op1, \$op2);
Rmpf_ui_div(\$rop, \$ui, \$op);
Rmpf_div_ui(\$rop, \$op, \$ui);
\$rop = 2nd arg / 3rd arg.

Rmpf_sqrt(\$rop, \$op);
Rmpf_sqrt_ui(\$rop, \$ui);
\$rop = 2nd arg ** 0.5.

Rmpf_pow_ui(\$rop, \$op, \$ui);
\$ROP = \$OP ** \$ui.

Rmpf_neg(\$rop, \$op);
\$rop = -\$op.

Rmpf_abs(\$rop, \$op);
\$rop = abs(\$op).

Rmpf_mul_2exp(\$rop, \$op, \$ui);
\$rop = \$op * (2 ** \$ui).

Rmpf_div_2exp(\$rop, \$op, \$ui);
\$rop = \$op / (2 ** \$ui).

####################

COMPARISON FUNCTIONS

\$si = Rmpf_cmp(\$op1, \$op2);
\$si = Rmpf_cmp_ui(\$op, \$ui);
\$si = Rmpf_cmp_si(\$op, \$si);
\$si = Rmpf_cmp_d(\$op, \$double);
Compare 1st arg and 2nd arg.  Return a positive value if
1st arg >  2nd arg, zero if 1st arg = 2nd arg, and a
negative value if 1st arg < 2nd arg.

Rmpf_eq(\$op1, \$op2, \$bits);
Return non-zero if the first \$bits bits of \$op1 and \$op2
are equal, zero otherwise.  I.e., test if \$op1 and \$op2
are approximately equal.
Caution: Currently only whole limbs are compared, and only in an
exact fashion.

Rmpf_reldiff(\$rop, \$op1, \$op2);
\$rop = abs(\$op1 - \$op2) / \$op1.

\$si = Rmpf_sgn(\$op);
Returns either +1 or -1 (or 0 if \$op is zero).

##########################

INPUT AND OUTPUT FUNCTIONS

\$bytes_read = Rmpf_inp_str(\$rop, \$base);
BEST TO USE TRmpf_inp_str INSTEAD.
Read a string in base \$base from STDIN, and put the read
float in \$rop. The string is of the form `M@N' or, if
\$base is 10 or less, alternatively `MeN'.  `M' is the
mantissa and `N' is the exponent.  The mantissa is always
in the specified base. The exponent is either in the
specified base or, if \$base is negative,in decimal. The
decimal point expected is taken from the current locale,
on systems providing `localeconv'. The argument \$base may
be in the ranges 2 to 36, or -36 to -2. Negative values
are used to specify that the exponent is in decimal.
Unlike the corresponding `Math::GMPz' function, the
base will not be determined from the leading characters
of the string if \$base is 0. This is so that numbers
like `0.23' are not interpreted as octal.

\$bytes_read = TRmpf_inp_str(\$rop, \$stream, \$base);
As for Rmpf_inp_str, except that there's the capability to read
from somewhere other than STDIN.
To read from STDIN:
TRmpf_inp_str(\$rop, *stdin, \$base);
To read from an open filehandle (let's call it FH):
TRmpf_inp_str(\$rop, \*FH, \$base);

\$bytes_written = Rmpf_out_str([\$prefix,] \$op, \$base, \$digits  [, \$suffix]);
BEST TO USE TRmpf_out_str INSTEAD.
Print \$op to stream, as a string of digits. Return the number of
bytes written, or if an error occurred, return 0. The mantissa is
prefixed with an `0.' and is in the given base, which may vary
from 2 to 62 or from -2 to -36. An exponent is then printed,
separated by an `e', or if the base is greater than 10 then by an
`@'. The exponent is always in decimal. The decimal point follows
the current locale, on systems providing localeconv. For bases in
the range 2..36, digits and lower-case letters are used; for
-2..-36, digits and upper-case letters are used; for 37..62, digits,
upper-case letters, and lower-case letters (in that significance
order) are used. Up to \$digits will be printed from the mantissa,
except that no more digits than are accurately representable by \$op
will be printed. \$digits can be 0 to select that accurate maximum.
The optional last argument (\$suffix) is a string (eg "\n") that
will be appended to the output. The optional first argument
(\$prefix) is a string that will be prepended to the output. Note
that either none, one, or both, of \$prefix and \$suffix may be
supplied. (\$bytes_written does not include the number of bytes in
\$suffix and \$prefix.)

\$bytes_written = TRmpf_out_str([\$prefix,] \$stream, \$base, \$digits, \$op, [, \$suffix]);
As for Rmpf_out_str, except that there's the capability to print
to somewhere other than STDOUT. Note that the order of the args
is different (to match the order of the mpf_out_str args).
To print to STDERR:
TRmpf_out_str(*stderr, \$base, \$digits, \$op);
To print to an open filehandle (let's call it FH):
TRmpf_out_str(\*FH, \$base, \$digits, \$op);

#######################

MISCELLANEOUS FUNCTIONS

Rmpf_ceil(\$rop, \$op);
Rmpf_floor(\$rop, \$op);
Rmpf_trunc(\$rop, \$op);
Set \$rop to \$op rounded to an integer.  `Rmpf_ceil' rounds to the
next higher integer, `mpf_floor' to the next lower, and
`Rmpf_trunc' to the integer towards zero.

\$bool = Rmpf_integer_p(\$op);
Return non-zero if \$op is an integer.

\$bool = Rmpf_fits_ulong_p(\$op);
\$bool = Rmpf_fits_slong_p(\$op);
\$bool = Rmpf_fits_uint_p(\$op);
\$bool = Rmpf_fits_sint_p(\$op);
\$bool = Rmpf_fits_ushort_p(\$op);
\$bool = Rmpf_fits_sshort_p(\$op);
Return non-zero if OP would fit in the respective C data
type, when truncated to an integer.

In Rmpf_urandomb() (below), @r is an array of
Math::GMPf objects (one for each random number that is
required). \$how_many is the number of random numbers you
want and must be equal to scalar(@r). \$bits is simply the
number of random bits required. Before calling
Rmpf_urandomb(), you first initialise state by calling
Math::GMPz::rand_init(). When you've finished with
Rmpf_urandomb, call Math::GMPz::rand_clear().
With Rmpf_random2() there is no need to call rand_init()
and rand_clear().

\$state = Math::GMPz::rand_init(\$z);
\$z is the seed - a Math::GMPz object.

Rmpf_urandomb(@r, \$state, \$bits, \$how_many);
Generate uniformly distributed random floats, all
between 0 and 1, with \$bits significant bits in the mantissa.

Math::GMPz::rand_clear(\$state);
Rmpf_random2(@r, \$limbs, \$exp, \$how_many);
Generate random floats of at most \$limbs limbs, with long
strings of zeros and ones in the binary representation.
The exponent of the number is in the interval -\$exp to \$exp.
This function is useful for testing functions and algorithms,
since this kind of random numbers have proven to be more
likely to trigger corner-case bugs.  Negative random
numbers are generated when \$limbs is negative.

####################

Overloading works with numbers, strings (base 10 only)
and Math::GMPf objects. Strings are coerced into
Math::GMPf objects (with default precision).

The following operators are overloaded:
+ - * / ** sqrt (Return values have default precision)
+= -= *= /= **= (Precision remains unchanged)
< <= > >= == != <=>
! not
abs (Return value has default precision)
int (on perl 5.8 only, NA on perl 5.6.
Return value has default precision.)
= (The copy that gets modified will have default precision.
The other copy retains the precision of the original)
""

Atempting to use the overloaded operators with objects that
have been blessed into some package other than 'Math::GMPf'
will not work.

In those situations where the overload subroutine operates on 2
perl variables, then obviously one of those perl variables is
a Math::GMPf object. To determine the value of the other variable
the subroutine works through the following steps (in order),
using the first value it finds, or croaking if it gets
to step 6:

1. If the variable is an unsigned long then that value is used.
The variable is considered to be an unsigned long if
(perl 5.8) the UOK flag is set or if (perl 5.6) SvIsUV()
returns true.

2. If the variable is a signed long int, then that value is used.
The variable is considered to be a signed long int if the
IOK flag is set. (In the case of perls built with
-Duse64bitint, the variable is treated as a signed long long
int if the IOK flag is set.)

3. If the variable is a double, then that value is used. The
variable is considered to be a double if the NOK flag is set.

4. If the variable is a string (ie the POK flag is set) then the
base 10 value of that string is used. If the POK flag is set,
but the string is not a valid base 10 number, the subroutine
croaks with an appropriate error message.

5. If the variable is a Math::GMPf object then the value of that
object is used.

6. If none of the above is true, then the second variable is
deemed to be of an invalid type. The subroutine croaks with
an appropriate error message.

#####

OTHER

\$GMP_version = Math::GMPf::gmp_v;
Returns the version of the GMP library (eg 4.1.3). The function
is not exportable.

\$GMP_cc = Math::GMPf::__GMP_CC;
\$GMP_cflags = Math::GMPf::__GMP_CFLAGS;
Returns respectively the CC and CFLAGS settings that were used
to compile the gmp library. (Not exportable.)

\$major = Math::GMPf::__GNU_MP_VERSION;
\$minor = Math::GMPf::__GNU_MP_VERSION_MINOR;
\$patchlevel = Math::GMPf::__GNU_MP_VERSION_PATCHLEVEL;
Returns respectively the major, minor, and patchlevel numbers
for the GMP library version used by Math::GMPf. (These
functions are in @EXPORT_OK and are therefore exportable by
request.)

################

FORMATTED OUTPUT

NOTE: The format specification can be found at:
http://gmplib.org/manual/Formatted-Output-Strings.html#Formatted-Output-Strings
However, the use of '*' to take an extra variable for width and
precision is not allowed in this implementation. Instead, it is
necessary to interpolate the variable into the format string - ie,
Rmpf_printf("%*Zd\n", \$width, \$mpz);
we need:
Rmpf_printf("%\${width}Zd\n", \$mpz);

\$si = Rmpf_printf(\$format_string, \$var);

This function changed with the release of Math-GMPz-0.27.
Now (unlike the GMP counterpart), it is limited to taking 2
arguments - the format string, and the variable to be formatted.
That is, you can format only one variable at a time.
Returns the number of characters written, or -1 if an error
occurred.

\$si = Rmpf_fprintf(\$fh, \$format_string, \$var);

This function (unlike the GMP counterpart) is limited to taking
3 arguments - the filehandle, the format string, and the variable
to be formatted. That is, you can format only one variable at a time.
Other than that, the rules outlined above wrt Rmpf_printf apply.
Returns the number of characters written, or -1 if an error
occurred.

\$si = Rmpf_sprintf(\$buffer, \$format_string, \$var);

This function (unlike the GMP counterpart) is limited to taking
3 arguments - the buffer, the format string, and the variable
to be formatted. \$buffer must be large enough to accommodate the
formatted string, and is truncated to the length of that formatted
string. If you prefer to have the resultant string returned (rather
than stored in \$buffer), use Rmpz_sprintf_ret instead - which will
also leave the length of \$buffer unaltered.
Returns the number of characters written, or -1 if an error
occurred.

\$string = Rmpf_sprintf_ret(\$buffer, \$format_string, \$var);

As for Rmpf_sprintf, but returns the formatted string, as well as
storing it in \$buffer. \$buffer needs to be large enough to
accommodate the formatted string. The length of \$buffer will be
unaltered.

###############################
###############################``````

BUGS

``````   You can get segfaults if you pass the wrong type of
argument to the functions - so if you get a segfault, the
first thing to do is to check that the argument types
you have supplied are appropriate.``````

``````   This program is free software; you may redistribute it and/or
``   Sisyphus <sisyphus at(@) cpan dot (.) org>``