;; @module gmp.lsp
;; @description GNU MP Bignum Library interface
;; @version 1.52 updated for new install locations and GMP website
;; @version 1.53 took out redefinition of normal int operations to floats
;; @version 1.6 added library path for OpenBSD and tested for 64-bit newLISP
;; @version 1.7 doc changes
;; @version 1.8 doc changes
;; @version 1.9 fixes for 64-bit in <,>,>=,<=,factor didn't sign extend 32->64
;; @version 2.0 more fixes for 64-bit, allocated space for handles was to small
;; @version 2.1 changed deprecated name to term
;; @version 2.2 out-comment 64-bit lib on Mac OSX, newLISP is shipped as 32-bit
;; @author Lutz Mueller, 2007-2011
;;
The GNU MP Bignum Library
;; This modules interfaces to libgmp which can be obtained from @link http://gmplib.org/ http://gmplib.org/ .
;;
;; Source code for libgmp.so (UNIX) or lbgmp.dll (Win32) or libgmp.dylib (Mac OS X)
;; is available at this site. After installing the library the correct path-name should
;; be added in the (set 'files ... ) statement around line 130 in this module.
;;
;; When compiling libgmp for Mac OS X on Intel CPUs and for 32-bit newLISP use:
;;
;; ./configure CFLAGS="-m32" ABI=32
;; make
;; sudo make install
;;
;; This interface module presets the maximum precision to 1024.
;; The precision can be changed to any other value by changing the
;; definition of 'MAX_PRECISION' in the source of this module.
;;
;; All arguments to the GMP functions in this module must be given as strings,
;; an error will be thrown otherwise. When starting with a 0 the number is
;; assumed to be in octal format when starting with 0x in hexadecimal format.
;;
;; This file only imports a few functions from the many available in GNU GMP.
;; See the GMP manual for more functions.
;;
;; Note, that since version 8.9.7 newLISP does all integer arithmetik with
;; 64 bits giving up to 19 digits of precision. For precisions less or equal
;; 19 digits newLISP's built-in 64-bit integer arithmetik is much faster.
;;
;; Usage
;; At the beginning of the programfile include a 'load' statement for the module:
;;
;; (load "/usr/share/newlisp/modules/gmp.lsp")
;; ; or shorter
;; (module "gmp.lsp")
;;
;; @example
;; (GMP:+ "123456789012345678901234567890" "123456789012345678901234567890")
;; => "246913578024691357802469135780"
;; Adding more functions to the library
;; When adding functions be aware that inside the GMP context
;; +,-,*,/,=,<,>,<=,>= are overwritten for multiple precision and the
;; original operators would have would have to be prefixed with MAIN when used,
;; inside the 'gmp.lsp' module.
;;
;; Integer arithmetik
;; @syntax (GMP:+ )
;; add two integers in and
;; @syntax (GMP:- )
;; subtract from
;; @syntax (GMP:* )
;; multiply by
;; @syntax (GMP:/ )
;; divide by , round towards '0' (zero)
;; @syntax (GMP:% )
;; calc rest of division /
;; @syntax (GMP:** )
;; calc power(, )
;; @syntax (GMP:= )
;; test for equal
;; @syntax (GMP:< )
;; test for smaller
;; @syntax (GMP:> )
;; test for bigger
;; @syntax (GMP:<= )]
;; test for smaller or equal
;; @syntax (GMP:>= )
;; test for bigger or equal
;;
;; Bit operations
;; @syntax (GMP:& )
;; bitwise of ,
;; @syntax (GMP:| )
;; bitwise inclusive of ,
;; @syntax (GMP:^ )
;; bitwise exclusive of ,
;; @syntax (GMP:~ )
;; bitwise complement of
;;
;; Number theory
;; @syntax (GMP:prime? )
;; check if is prime
;; @syntax (GMP:next-prime )
;; calc closes prime greater than
;; @syntax (GMP:factor )
;; calc a list of prime factors for
;; @syntax (GMP:gcd )
;; greatest common divisor of and
;; @syntax (GMP:bin )
;; calc binomial ( )
;; @syntax (GMP:fac )
;; arg! factorial()
;; @syntax (GMP:fib )
;; fibonacci(arg)
;;
;; Random numbers
;; @syntax (GMP:seed )
;; seed the random generator
;; @syntax (GMP:rand )
;; generate random numbers between 0 and arg - 1
;;
;(constant '+ add '- sub '* mul '/ div)
(if (< (sys-info -2) 10111)
(constant (global 'term) name))
(context 'GMP)
; convert from int32 to int64 with sign extension
; in both newLISP 32-bit and 64-bit integers are 64-bit
(define (int64 x) (get-int (pack "ld" x)))
; maximum digits, can be set to any value higher if required
; when choosing a different number the functions GMP:fac and GMP:fib
; have to be re-calibrated for their maximum value not to overflow
; the result space
;
(define MAX_PRECISION 1024)
(define gmp-type-error "String expected in GMP module.")
(define gmp-size-error "Argument too big in GMP module")
(define gmp-divzero-error "Division by zero in GMP module")
(set 'files '(
; "/usr/local/lib/libgmp.dylib" ; Mac OSX for 64-bit newLISP
"/usr/lib/libgmp.dylib" ;Mac OSX
"/usr/lib/libgmp.3.dylib" ;Mac OSX
"/opt/local/lib/libgmp.3.dylib" ;Mac OSX
"/opt/local/lib/libgmp.dylib" ;Mac OSX
"/usr/local/lib/libgmp.so.8.0" ; OpenBSD 4.6
"/usr/lib/libgmp.so.3" ; Linux, BSDs
"/usr/local/lib/libgmp.so" ; Linux, BSDs
"/WINDOWS/system32/libgmp-3.dll" ; Win32 DLL path on current drive
"libgmp-3.dll" ; Win32 in path for current directory
))
(set 'library (files (or
(find true (map file? files))
(throw-error "cannot find GMP library"))))
; integer arithmetik
(import library "__gmpz_init")
(import library "__gmpz_add")
(import library "__gmpz_add_ui")
(import library "__gmpz_sub")
(import library "__gmpz_mul")
(import library "__gmpz_tdiv_q")
(import library "__gmpz_tdiv_r")
(import library "__gmpz_cmp")
(import library "__gmpz_cmp_si")
(import library "__gmpz_set_si")
(import library "__gmpz_divisible_p")
(import library "__gmpz_pow_ui")
(import library "__gmpz_clear")
; bit operators
(import library "__gmpz_and")
(import library "__gmpz_ior")
(import library "__gmpz_xor")
(import library "__gmpz_com")
; number theory
(import library "__gmpz_probab_prime_p")
(import library "__gmpz_nextprime")
(import library "__gmpz_gcd")
(import library "__gmpz_bin_ui")
(import library "__gmpz_fac_ui")
(import library "__gmpz_fib_ui")
; random numbers
(import library "__gmpz_urandomm")
(import library "__gmp_randseed")
; conversion functions
(import library "__gmpz_set_str")
(import library "__gmpz_get_str")
; auxiliary functions
(import library "__gmpz_sizeinbase")
(import library "__gmp_randinit_default")
(import library "__gmp_randseed_ui")
; reserve handles
(define op1 (dup "\000" 16)) ; 12 on 32-bit 16 on 64-bit
(define op2 (dup "\000" 16))
(define rop (dup "\000" 16))
(define randstate (dup "\000" 32)) ; 20 on 32-bit 32 on 64-bit
; init handles
(__gmpz_init op1)
(__gmpz_init op2)
(__gmpz_init rop)
; handles to speed up factor
(define mp-n (dup "\000" 16))
(define mp-d (dup "\000" 16))
(define mp-k (dup "\000" 16))
(__gmpz_init mp-n)
(__gmpz_init mp-d)
(__gmpz_init mp-k)
; init randstate
(__gmp_randinit_default randstate)
(__gmp_randseed_ui randstate (time-of-day))
; init / reserve result string
(define rops (dup "\000" (MAIN:+ 2 MAX_PRECISION)))
; add two integers
;
(define (GMP:+ p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(__gmpz_add rop op1 op2)
(if (MAIN:> (__gmpz_sizeinbase rop 10) MAX_PRECISION)
(throw-error gmp-size-error))
(get-string (__gmpz_get_str rops 10 rop))
)
; subtract two integers
;
(define (GMP:- p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(__gmpz_sub rop op1 op2)
(if (MAIN:> (__gmpz_sizeinbase rop 10) MAX_PRECISION)
(throw-error gmp-size-error))
(get-string (__gmpz_get_str rops 10 rop))
)
; multiply two integers
;
(define (GMP:* p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(__gmpz_mul rop op1 op2)
(if (MAIN:> (__gmpz_sizeinbase rop 10) MAX_PRECISION)
(throw-error gmp-size-error))
(get-string (__gmpz_get_str rops 10 rop))
)
; divide two integers
; return floor value of result
(define (GMP:/ p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(if (MAIN:= (__gmpz_cmp_si op2 0) 0) (throw-error gmp-divzero-error))
(__gmpz_tdiv_q rop op1 op2)
(get-string (__gmpz_get_str rops 10 rop))
)
; modulo two integers
; return rest value of division result
(define (GMP:% p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(if (MAIN:= (__gmpz_cmp_si op2 0) 0) (throw-error gmp-divzero-error))
(__gmpz_tdiv_r rop op1 op2)
(get-string (__gmpz_get_str rops 10 rop))
)
; exponentiation function
; return power(p1 p2)
;
(define (GMP:** p1 p2 , pexp)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(set 'pexp (int p2 0))
(__gmpz_pow_ui rop op1 pexp)
(if (MAIN:> (__gmpz_sizeinbase rop 10) MAX_PRECISION)
(throw-error gmp-size-error))
(get-string (__gmpz_get_str rops 10 rop))
)
; test of two integers are equal, return true if equal
; otherwise nil
;
(define (GMP:= p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(MAIN:= (__gmpz_cmp op1 op2) 0)
)
; test is p1 is smaller than p2
; return true or nil
;
(define (GMP:< p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(MAIN:< (int64 (__gmpz_cmp op1 op2)) 0)
)
; test is p1 is smaller than p2
; return true or nil
;
(define (GMP:> p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(MAIN:> (int64 (__gmpz_cmp op1 op2)) 0)
)
; test is p1 is smaller or eaual than p2
; return true or nil
;
(define (GMP:<= p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(MAIN:<= (int64 (__gmpz_cmp op1 op2)) 0)
)
; test is p1 is bigger or equal than p2
; return true or nil
;
(define (GMP:>= p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(MAIN:>= (int64 (__gmpz_cmp op1 op2)) 0)
)
; bitwise and two integers
;
(define (GMP:& p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(__gmpz_and rop op1 op2)
(get-string (__gmpz_get_str rops 10 rop))
)
; bitwise inclusive or two integers
;
(define (GMP:| p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(__gmpz_ior rop op1 op2)
(get-string (__gmpz_get_str rops 10 rop))
)
; bitwise exclusive or two integers
;
(define (GMP:^ p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(__gmpz_xor rop op1 op2)
(get-string (__gmpz_get_str rops 10 rop))
)
; bitwise complement of one integer
;
(define (GMP:~ p1)
(if (not (string? p1)) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_com rop op1)
(get-string (__gmpz_get_str rops 10 rop))
)
; check if a prime
; returns true if a prime
; returns nil if not a primes or probably not a prime
;
(define (GMP:prime? p1 , r)
(if (not (string? p1)) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(set 'r (__gmpz_probab_prime_p op1 10))
(if
(MAIN:= r 0) nil
(MAIN:= r 1) (MAIN:= (next-prime (- p1 "1")) p1)
(MAIN:= r 2) true)
)
; get the next prime higher then arg
;
(define (GMP:next-prime p1)
(if (not (string? p1)) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_nextprime rop op1)
(get-string (__gmpz_get_str rops 10 rop))
)
; factorize n
;
(define (GMP:factor n)
(if (not (string? n)) (throw-error gmp-type-error))
(set 'factors nil)
(set 'prevfact nil)
(__gmpz_set_str mp-n n 0)
(if (MAIN:> (int64 (__gmpz_cmp_si mp-n 2)) 0)
(begin
(__gmpz_set_si mp-d 2)
(__gmpz_set_si mp-k 0)
(while (MAIN:!= 0 (__gmpz_divisible_p mp-n mp-d))
(__gmpz_tdiv_q mp-n mp-n mp-d)
(__gmpz_add_ui mp-k mp-k 1)
)
(if (MAIN:> (int64 (__gmpz_cmp_si mp-k 0)) 0)
(push-factor mp-d mp-k))
(__gmpz_set_si mp-d 3)
(__gmpz_mul op1 mp-d mp-d)
(while (MAIN:<= (int64 (__gmpz_cmp op1 mp-n)) 0)
(__gmpz_set_si mp-k 0)
(while (MAIN:!= 0 (__gmpz_divisible_p mp-n mp-d))
(__gmpz_tdiv_q mp-n mp-n mp-d)
(__gmpz_add_ui mp-k mp-k 1) )
(if (MAIN:> (int64 (__gmpz_cmp_si mp-k 0)) 0) (push-factor mp-d mp-k))
(__gmpz_add_ui mp-d mp-d 2)
(__gmpz_mul op1 mp-d mp-d) )
)
)
(if (MAIN:> (int64 (__gmpz_cmp_si mp-n 1)))
(if prevfact
(begin
(___gmpz_set_si op1 1)
(push-factor mp-n op1))
(begin
(set 'n (get-string (__gmpz_get_str rops 10 mp-n)))
(push n factors -1))))
factors ; not necessary starting v.9.9.5 because push returns list
)
(define (push-factor f k)
(set 'f (get-string (__gmpz_get_str rops 10 f)))
(set 'k (get-string (__gmpz_get_str rops 10 k)))
(if (not prevfact)
(begin
(push f factors -1)
(set 'k (- k "1"))))
(while (> k "0")
(push f factors -1)
(set 'k (- k "1")) )
)
; get the greates common divisor
;
(define (GMP:gcd p1 p2)
(if (or (not (string? p1)) (not (string? p2))) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_set_str op2 p2 0)
(__gmpz_gcd rop op1 op2)
(get-string (__gmpz_get_str rops 10 rop))
)
; get binomial of
;
(define (GMP:bin n k)
(if (not (string? n)) (throw-error gmp-type-error))
(__gmpz_set_str op1 n 0)
(set 'k (int k 0))
(__gmpz_bin_ui rop op1 k)
(if (MAIN:> (__gmpz_sizeinbase rop 10) MAX_PRECISION)
(throw-error gmp-size-error))
(get-string (__gmpz_get_str rops 10 rop))
)
; get factorial of arg
; args may not be bigger than 458
; for 1024 digits precision in the result
;
(define (GMP:fac p1 , n)
(if (not (string? p1)) (throw-error gmp-type-error))
(set 'n (int p1))
(__gmpz_fac_ui rop n)
(if (MAIN:> (__gmpz_sizeinbase rop 10) MAX_PRECISION)
(throw-error gmp-size-error))
(get-string (__gmpz_get_str rops 10 rop))
)
; get fibonacci of arg
; arg may not be bigger than for a 1024 digits result
;
(define (GMP:fib p1 , n)
(if (not (string? p1)) (throw-error gmp-type-error))
(set 'n (int p1))
(__gmpz_fib_ui rop n)
(if (MAIN:> (__gmpz_sizeinbase rop 10) MAX_PRECISION)
(throw-error gmp-size-error))
(get-string (__gmpz_get_str rops 10 rop))
)
; get a random number between 0 and arg - 1
;
(define (GMP:rand p1)
(if (not (string? p1)) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmpz_urandomm rop randstate op1)
(get-string (__gmpz_get_str rops 10 rop))
)
; seed the random generator
;
(define (GMP:seed p1)
(if (not (string? p1)) (throw-error gmp-type-error))
(__gmpz_set_str op1 p1 0)
(__gmp_randseed randstate op1)
true
)
(define (check func result)
(if (MAIN:= (apply func (list )) result)
(println "GMP:" (term func) "\t-> Ok")
(println "Problem in GMP:" (term func)))
)
(context 'MAIN)
; this test assumes that libgmp itself is correct, only the
; newLISP implementation is tested
;
(define (test-GMP)
; INTEGER ARITHMETIK
(GMP:check 'GMP:+ "123" "456" "579")
(GMP:check 'GMP:- "100" "99" "1")
(GMP:check 'GMP:* "123" "456" "56088")
(GMP:check 'GMP:/ "56088" "456" "123")
(GMP:check 'GMP:% "56088" "456" "0")
(GMP:check 'GMP:** "2" "10" "1024")
(GMP:check 'GMP:= "999999" "999999" true)
(GMP:check 'GMP:< "999999" "1000000" true)
(GMP:check 'GMP:> "999999" "1000000" nil)
(GMP:check 'GMP:<= "999999" "1000000" true)
(GMP:check 'GMP:>= "999999" "1000000" nil)
; BIT OPERATTIONS
(GMP:check 'GMP:& "0xAAAA" "0xFF00" "43520")
(GMP:check 'GMP:| "0xAAAA" "0x5555" "65535")
(GMP:check 'GMP:^ "0xAAAA" "0xAAAA" "0")
(GMP:check 'GMP:~ "0xFFFF" nil "-65536")
; NUMBER THEORY
(GMP:check 'GMP:prime? "127" nil true)
(GMP:check 'GMP:next-prime "127" nil "131")
(GMP:check 'GMP:factor "123" nil '("3" "41"))
(GMP:check 'GMP:gcd "20" "8" "4")
(GMP:check 'GMP:bin "10" "2" "45")
(GMP:check 'GMP:fac "5" nil "120")
(GMP:check 'GMP:fib "30" nil "832040")
; RANDOM NUMBERS
(GMP:check 'GMP:seed "12345" nil true)
;(GMP:check 'GMP:rand "1000000" nil "18235") ; check manually
)
; eof