#
source:
trunk/third/gmp/NEWS
@
22254

Revision 22254, 11.1 KB checked in by ghudson, 19 years ago (diff) |
---|

Line | |
---|---|

1 | Copyright 1996, 1999, 2000, 2001, 2002, 2004 Free Software Foundation, Inc. |

2 | |

3 | This file is part of the GNU MP Library. |

4 | |

5 | The GNU MP Library is free software; you can redistribute it and/or modify |

6 | it under the terms of the GNU Lesser General Public License as published by |

7 | the Free Software Foundation; either version 2.1 of the License, or (at your |

8 | option) any later version. |

9 | |

10 | The GNU MP Library is distributed in the hope that it will be useful, but |

11 | WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY |

12 | or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public |

13 | License for more details. |

14 | |

15 | You should have received a copy of the GNU Lesser General Public License |

16 | along with the GNU MP Library; see the file COPYING.LIB. If not, write to |

17 | the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA |

18 | 02111-1307, USA. |

19 | |

20 | |

21 | Changes between GMP version 4.1.3 and 4.1.4 |

22 | |

23 | * Bug fix to FFT multiplication code (crash for huge operands). |

24 | * Bug fix to mpf_sub (miscomputation). |

25 | * Support for powerpc64-gnu-linux. |

26 | * Better support for AMD64 in 32-bit mode. |

27 | * Upwardly binary compatible with 4.1.3, 4.1.2, 4.1.1, 4.1, 4.0.1, 4.0, |

28 | and 3.x versions. |

29 | |

30 | Changes between GMP version 4.1.2 and 4.1.3 |

31 | |

32 | * Bug fix for FFT multiplication code (miscomputation). |

33 | * Bug fix to K6 assembly code for gcd. |

34 | * Bug fix to IA-64 assembly code for population count. |

35 | * Portability improvements, most notably functional AMD64 support. |

36 | * mpz_export allows NULL for countp parameter. |

37 | * Many minor bug fixes. |

38 | * mpz_export allows NULL for countp parameter. |

39 | * Upwardly binary compatible with 4.1.2, 4.1.1, 4.1, 4.0.1, 4.0, and 3.x |

40 | versions. |

41 | |

42 | Changes between GMP version 4.1.1 and 4.1.2 |

43 | |

44 | * Bug fixes. |

45 | |

46 | Changes between GMP version 4.1 and 4.1.1 |

47 | |

48 | * Bug fixes. |

49 | * New systems supported: NetBSD and OpenBSD sparc64. |

50 | |

51 | Changes between GMP version 4.0.1 and 4.1 |

52 | |

53 | * Bug fixes. |

54 | * Speed improvements. |

55 | * Upwardly binary compatible with 4.0, 4.0.1, and 3.x versions. |

56 | * Asymptotically fast conversion to/from strings (mpz, mpq, mpn levels), but |

57 | also major speed improvements for tiny operands. |

58 | * mpn_get_str parameter restrictions relaxed. |

59 | * Major speed improvments for HPPA 2.0 systems. |

60 | * Major speed improvments for UltraSPARC systems. |

61 | * Major speed improvments for IA-64 systems (but still sub-optimal code). |

62 | * Extended test suite. |

63 | * mpfr is back, with many bug fixes and portability improvements. |

64 | * New function: mpz_ui_sub. |

65 | * New functions: mpz_export, mpz_import. |

66 | * Optimization for nth root functions (mpz_root, mpz_perfect_power_p). |

67 | * Optimization for extended gcd (mpz_gcdext, mpz_invert, mpn_gcdext). |

68 | * Generalized low-level number format, reserving a `nails' part of each |

69 | limb. (Please note that this is really experimental; some functions |

70 | are likely to compute garbage when nails are enabled.) |

71 | * Nails-enabled Alpha 21264 assembly code, allowing up to 75% better |

72 | performance. (Use --enable-nails=4 to enable it.) |

73 | |

74 | Changes between GMP version 4.0 and 4.0.1 |

75 | |

76 | * Bug fixes. |

77 | |

78 | Changes between GMP version 3.1.1 and 4.0 |

79 | |

80 | * Bug fixes. |

81 | * Speed improvements. |

82 | * Upwardly binary compatible with 3.x versions. |

83 | * New CPU support: IA-64, Pentium 4. |

84 | * Improved CPU support: 21264, Cray vector systems. |

85 | * Support for all MIPS ABIs: o32, n32, 64. |

86 | * New systems supported: Darwin, SCO, Windows DLLs. |

87 | * New divide-and-conquer square root algorithm. |

88 | * New algorithms chapter in the manual. |

89 | * New malloc reentrant temporary memory method. |

90 | * New C++ class interface by Gerardo Ballabio (beta). |

91 | * Revamped configure, featuring ABI selection. |

92 | * Speed improvements for mpz_powm and mpz_powm_ui (mainly affecting small |

93 | operands). |

94 | * mpz_perfect_power_p now properly recognizes 0, 1, and negative perfect |

95 | powers. |

96 | * mpz_hamdist now supports negative operands. |

97 | * mpz_jacobi now accepts non-positive denominators. |

98 | * mpz_powm now supports negative exponents. |

99 | * mpn_mul_1 operand overlap requirements relaxed. |

100 | * Float input and output uses locale specific decimal point where available. |

101 | * New gmp_printf, gmp_scanf and related functions. |

102 | * New division functions: mpz_cdiv_q_2exp, mpz_cdiv_r_2exp, mpz_divexact_ui. |

103 | * New divisibility tests: mpz_divisible_p, mpz_divisible_ui_p, |

104 | mpz_divisible_2exp_p, mpz_congruent_p, mpz_congruent_ui_p, |

105 | mpz_congruent_2exp_p. |

106 | * New Fibonacci function: mpz_fib2_ui. |

107 | * New Lucas number functions: mpz_lucnum_ui, mpz_lucnum2_ui. |

108 | * Other new integer functions: mpz_cmp_d, mpz_cmpabs_d, mpz_get_d_2exp, |

109 | mpz_init2, mpz_kronecker, mpz_lcm_ui, mpz_realloc2. |

110 | * New rational I/O: mpq_get_str, mpq_inp_str, mpq_out_str, mpq_set_str. |

111 | * Other new rational functions: mpq_abs, mpq_cmp_si, mpq_div_2exp, |

112 | mpq_mul_2exp, mpq_set_f. |

113 | * New float tests: mpf_integer_p, mpf_fits_sint_p, mpf_fits_slong_p, |

114 | mpf_fits_sshort_p, mpf_fits_uint_p, mpf_fits_ulong_p, mpf_fits_ushort_p. |

115 | * Other new float functions: mpf_cmp_d, mpf_get_default_prec, mpf_get_si, |

116 | mpf_get_ui, mpf_get_d_2exp. |

117 | * New random functions: gmp_randinit_default, gmp_randinit_lc_2exp_size. |

118 | * New demo expression string parser (see demos/expr). |

119 | * New preliminary perl interface (see demos/perl). |

120 | * Tuned algorithm thresholds for many more CPUs. |

121 | |

122 | Changes between GMP version 3.1 and 3.1.1 |

123 | |

124 | * Bug fixes for division (rare), mpf_get_str, FFT, and miscellaneous minor |

125 | things. |

126 | |

127 | Changes between GMP version 3.0 and 3.1 |

128 | |

129 | * Bug fixes. |

130 | * Improved `make check' running more tests. |

131 | * Tuned algorithm cutoff points for many machines. This will improve speed for |

132 | a lot of operations, in some cases by a large amount. |

133 | * Major speed improvments: Alpha 21264. |

134 | * Some speed improvments: Cray vector computers, AMD K6 and Athlon, Intel P5 |

135 | and Pentium Pro/II/III. |

136 | * The mpf_get_prec function now works as it did in GMP 2. |

137 | * New utilities for auto-tuning and speed measuring. |

138 | * Multiplication now optionally uses FFT for very large operands. (To enable |

139 | it, pass --enable-fft to configure.) |

140 | * Support for new systems: Solaris running on x86, FreeBSD 5, HP-UX 11, Cray |

141 | vector computers, Rhapsody, Nextstep/Openstep, MacOS. |

142 | * Support for shared libraries on 32-bit HPPA. |

143 | * New integer functions: mpz_mul_si, mpz_odd_p, mpz_even_p. |

144 | * New Kronecker symbol functions: mpz_kronecker_si, mpz_kronecker_ui, |

145 | mpz_si_kronecker, mpz_ui_kronecker. |

146 | * New rational functions: mpq_out_str, mpq_swap. |

147 | * New float functions: mpf_swap. |

148 | * New mpn functions: mpn_divexact_by3c, mpn_tdiv_qr. |

149 | * New EXPERIMENTAL function layer for accurate floating-point arithmetic, mpfr. |

150 | To try it, pass --enable-mpfr to configure. See the mpfr subdirectory for |

151 | more information; it is not documented in the main GMP manual. |

152 | |

153 | Changes between GMP version 3.0 and 3.0.1 |

154 | |

155 | * Memory leaks in gmp_randinit and mpz_probab_prime_p fixed. |

156 | * Documentation for gmp_randinit fixed. Misc documentation errors fixed. |

157 | |

158 | Changes between GMP version 2.0 and 3.0 |

159 | |

160 | * Source level compatibility with past releases (except mpn_gcd). |

161 | * Bug fixes. |

162 | * Much improved speed thanks to both host independent and host dependent |

163 | optimizations. |

164 | * Switch to autoconf/automake/libtool. |

165 | * Support for building libgmp as a shared library. |

166 | * Multiplication and squaring using 3-way Toom-Cook. |

167 | * Division using the Burnikel-Ziegler method. |

168 | * New functions computing binomial coefficients: mpz_bin_ui, mpz_bin_uiui. |

169 | * New function computing Fibonacci numbers: mpz_fib_ui. |

170 | * New random number generators: mpf_urandomb, mpz_rrandomb, mpz_urandomb, |

171 | mpz_urandomm, gmp_randclear, gmp_randinit, gmp_randinit_lc_2exp, gmp_randseed, |

172 | gmp_randseed_ui. |

173 | * New function for quickly extracting limbs: mpz_getlimbn. |

174 | * New functions performing integer size tests: mpz_fits_sint_p, |

175 | mpz_fits_slong_p, mpz_fits_sshort_p, mpz_fits_uint_p, mpz_fits_ulong_p, |

176 | mpz_fits_ushort_p. |

177 | * New mpf functions: mpf_ceil, mpf_floor, mpf_pow_ui, mpf_trunc. |

178 | * New mpq function: mpq_set_d. |

179 | * New mpz functions: mpz_addmul_ui, mpz_cmpabs, mpz_cmpabs_ui, mpz_lcm, |

180 | mpz_nextprime, mpz_perfect_power_p, mpz_remove, mpz_root, mpz_swap, |

181 | mpz_tdiv_ui, mpz_tstbit, mpz_xor. |

182 | * New mpn function: mpn_divexact_by3. |

183 | * New CPU support: DEC Alpha 21264, AMD K6 and Athlon, HPPA 2.0 and 64, |

184 | Intel Pentium Pro and Pentium-II/III, Sparc 64, PowerPC 64. |

185 | * Almost 10 times faster mpz_invert and mpn_gcdext. |

186 | * The interface of mpn_gcd has changed. |

187 | * Better support for MIPS R4x000 and R5000 under Irix 6. |

188 | * Improved support for SPARCv8 and SPARCv9 processors. |

189 | |

190 | Changes between GMP version 2.0 and 2.0.2 |

191 | |

192 | * Many bug fixes. |

193 | |

194 | Changes between GMP version 1.3.2 and 2.0 |

195 | |

196 | * Division routines in the mpz class have changed. There are three classes of |

197 | functions, that rounds the quotient to -infinity, 0, and +infinity, |

198 | respectively. The first class of functions have names that begin with |

199 | mpz_fdiv (f is short for floor), the second class' names begin with mpz_tdiv |

200 | (t is short for trunc), and the third class' names begin with mpz_cdiv (c is |

201 | short for ceil). |

202 | |

203 | The old division routines beginning with mpz_m are similar to the new |

204 | mpz_fdiv, with the exception that some of the new functions return useful |

205 | values. |

206 | |

207 | The old function names can still be used. All the old functions names will |

208 | now do floor division, not trunc division as some of them used to. This was |

209 | changed to make the functions more compatible with common mathematical |

210 | practice. |

211 | |

212 | The mpz_mod and mpz_mod_ui functions now compute the mathematical mod |

213 | function. I.e., the sign of the 2nd argument is ignored. |

214 | |

215 | * The mpq assignment functions do not canonicalize their results. A new |

216 | function, mpq_canonicalize must be called by the user if the result is not |

217 | known to be canonical. |

218 | * The mpn functions are now documented. These functions are intended for |

219 | very time critical applications, or applications that need full control over |

220 | memory allocation. Note that the mpn interface is irregular and hard to |

221 | use. |

222 | * New functions for arbitrary precision floating point arithmetic. Names |

223 | begin with `mpf_'. Associated type mpf_t. |

224 | * New and improved mpz functions, including much faster GCD, fast exact |

225 | division (mpz_divexact), bit scan (mpz_scan0 and mpz_scan1), and number |

226 | theoretical functions like Jacobi (mpz_jacobi) and multiplicative inverse |

227 | (mpz_invert). |

228 | * New variable types (mpz_t and mpq_t) are available that makes syntax of |

229 | mpz and mpq calls nicer (no need for & before variables). The MP_INT and |

230 | MP_RAT types are still available for compatibility. |

231 | * Uses GNU configure. This makes it possible to choose target architecture |

232 | and CPU variant, and to compile into a separate object directory. |

233 | * Carefully optimized assembly for important inner loops. Support for DEC |

234 | Alpha, Amd 29000, HPPA 1.0 and 1.1, Intel Pentium and generic x86, Intel |

235 | i960, Motorola MC68000, MC68020, MC88100, and MC88110, Motorola/IBM |

236 | PowerPC, National NS32000, IBM POWER, MIPS R3000, R4000, SPARCv7, |

237 | SuperSPARC, generic SPARCv8, and DEC VAX. Some support also for ARM, |

238 | Clipper, IBM ROMP (RT), and Pyramid AP/XP. |

239 | * Faster. Thanks to the assembler code, new algorithms, and general tuning. |

240 | In particular, the speed on machines without GCC is improved. |

241 | * Support for machines without alloca. |

242 | * Now under the LGPL. |

243 | |

244 | INCOMPATIBILITIES BETWEEN GMP 1 AND GMP 2 |

245 | |

246 | * mpq assignment functions do not canonicalize their results. |

247 | * mpz division functions round differently. |

248 | * mpz mod functions now really compute mod. |

249 | * mpz_powm and mpz_powm_ui now really use mod for reduction. |

250 | |

251 | |

252 | |

253 | ---------------- |

254 | Local variables: |

255 | mode: text |

256 | fill-column: 76 |

257 | End: |

**Note:**See TracBrowser for help on using the repository browser.