| 1 |
2006-08-06, version 1.1.12
|
| 2 |
==========================
|
| 3 |
|
| 4 |
Implementation changes
|
| 5 |
----------------------
|
| 6 |
|
| 7 |
* Fix rare assertion when printing quite large floats.
|
| 8 |
|
| 9 |
Other changes
|
| 10 |
-------------
|
| 11 |
|
| 12 |
* Compilation fixes for several platforms: *BSD, Intel Mac, and MinGW.
|
| 13 |
|
| 14 |
|
| 15 |
2005-11-23, version 1.1.11
|
| 16 |
==========================
|
| 17 |
|
| 18 |
Algorithmic changes
|
| 19 |
-------------------
|
| 20 |
|
| 21 |
* Considerably improved performance of number input.
|
| 22 |
|
| 23 |
|
| 24 |
2005-10-22, version 1.1.10
|
| 25 |
==========================
|
| 26 |
|
| 27 |
Implementation changes
|
| 28 |
----------------------
|
| 29 |
|
| 30 |
* Removed the vptr of modular integer and univariate polynomial
|
| 31 |
classes in order to fix some crashes in that sector of CLN. Code
|
| 32 |
using this needs to be recompiled.
|
| 33 |
|
| 34 |
* Many more operator signatures make the integration of CLN types with
|
| 35 |
float/double easier.
|
| 36 |
|
| 37 |
Other changes
|
| 38 |
-------------
|
| 39 |
|
| 40 |
* Several compilation fixes.
|
| 41 |
|
| 42 |
* Made it possible to cross-compile CLN.
|
| 43 |
|
| 44 |
|
| 45 |
2004-11-03, version 1.1.9
|
| 46 |
=========================
|
| 47 |
|
| 48 |
Algorithmic changes
|
| 49 |
-------------------
|
| 50 |
|
| 51 |
* Input of numbers in bases 2, 4, 8, 16 and 32 is now done in linear bit
|
| 52 |
complexity as opposed to O(N^2). Useful for all kinds of persistency.
|
| 53 |
|
| 54 |
Implementation changes
|
| 55 |
----------------------
|
| 56 |
|
| 57 |
* Fixed several bugs in the integer input and output routines that could be
|
| 58 |
blamed for random crashes in the following cases: output in base 32 for
|
| 59 |
quite large numbers, input in base 2 for fixnums and input in base 3 for
|
| 60 |
fixnums on a 64 bit target.
|
| 61 |
|
| 62 |
* Fixed crash when radix specifiers were used in input streams.
|
| 63 |
|
| 64 |
* Speed up on x86_64 and ia64 by adding some inline assembly.
|
| 65 |
|
| 66 |
Other changes
|
| 67 |
-------------
|
| 68 |
|
| 69 |
* Fixes for compilation on MacOS X and little endian Mips.
|
| 70 |
|
| 71 |
|
| 72 |
2004-07-01, version 1.1.8
|
| 73 |
=========================
|
| 74 |
|
| 75 |
Implementation changes
|
| 76 |
----------------------
|
| 77 |
|
| 78 |
* Fix bug in log(a,b) when b is 1/n.
|
| 79 |
|
| 80 |
* Fix crash in shared library on AMD64.
|
| 81 |
|
| 82 |
Other changes
|
| 83 |
-------------
|
| 84 |
|
| 85 |
* Fixes for some language conformance issues.
|
| 86 |
|
| 87 |
|
| 88 |
2004-05-02, version 1.1.7
|
| 89 |
=========================
|
| 90 |
|
| 91 |
Implementation changes
|
| 92 |
----------------------
|
| 93 |
|
| 94 |
* Fixed two corner case bugs in cl_LF multiplication and sqrt.
|
| 95 |
|
| 96 |
* Workaround GCC 3.[123] compiler bug.
|
| 97 |
|
| 98 |
* Parallel build should work reliably now.
|
| 99 |
|
| 100 |
|
| 101 |
2004-01-01, version 1.1.6
|
| 102 |
=========================
|
| 103 |
|
| 104 |
Functional changes
|
| 105 |
------------------
|
| 106 |
|
| 107 |
* New function
|
| 108 |
`sintL ldegree(const cl_UP& x)'.
|
| 109 |
(Suggested by Munagala Ramanath.)
|
| 110 |
|
| 111 |
Implementation changes
|
| 112 |
----------------------
|
| 113 |
|
| 114 |
* Fixed a bug in polynomial subtraction.
|
| 115 |
(Reported by Munagala Ramanath.)
|
| 116 |
|
| 117 |
Other changes
|
| 118 |
-------------
|
| 119 |
|
| 120 |
* Provide a package control file cln.pc.
|
| 121 |
|
| 122 |
* Updated infrastructure to autoconf-2.57.
|
| 123 |
|
| 124 |
* Fixed compilation issues with GCC-3.4 prereleases and on Mac OSX.
|
| 125 |
|
| 126 |
|
| 127 |
2002-05-28, version 1.1.5
|
| 128 |
=========================
|
| 129 |
|
| 130 |
Other changes
|
| 131 |
-------------
|
| 132 |
|
| 133 |
* Fix compilation issues with GCC-3.1.
|
| 134 |
|
| 135 |
|
| 136 |
2002-01-04, version 1.1.4
|
| 137 |
=========================
|
| 138 |
|
| 139 |
Other changes
|
| 140 |
-------------
|
| 141 |
|
| 142 |
* Fix compilation issues with GMP-4.0 and on IBM s390.
|
| 143 |
|
| 144 |
* Updated infrastructure to autoconf-2.52.
|
| 145 |
|
| 146 |
|
| 147 |
2001-11-05, version 1.1.3
|
| 148 |
=========================
|
| 149 |
|
| 150 |
Implementation changes
|
| 151 |
----------------------
|
| 152 |
|
| 153 |
* Fix a computation error of sin and sinh for small arguments.
|
| 154 |
(Reported by Christopher Kennedy.)
|
| 155 |
|
| 156 |
* Module ordering works with gcc-3.0 -fuse-cxa-atexit now.
|
| 157 |
|
| 158 |
* -DNO_ASM works properly on Sparc.
|
| 159 |
|
| 160 |
|
| 161 |
2001-07-25, version 1.1.2
|
| 162 |
|
| 163 |
Implementation changes
|
| 164 |
----------------------
|
| 165 |
|
| 166 |
* Minor cleanups and portability issues.
|
| 167 |
(No need to upgrade if you have been happy so far.)
|
| 168 |
|
| 169 |
|
| 170 |
2001-05-31, version 1.1.1
|
| 171 |
=========================
|
| 172 |
|
| 173 |
Implementation changes
|
| 174 |
----------------------
|
| 175 |
|
| 176 |
* Minor cleanups for GCC 3.0 and libstdc++-v3.
|
| 177 |
|
| 178 |
* Fixes several little bugs and portability issues.
|
| 179 |
|
| 180 |
|
| 181 |
2000-12-14, version 1.1.0
|
| 182 |
=========================
|
| 183 |
|
| 184 |
Functional changes
|
| 185 |
------------------
|
| 186 |
|
| 187 |
* ISO/IEC 14882 fashion adjustments:
|
| 188 |
Put everything into namespace cln. All fundamental data types still
|
| 189 |
carry their old names. Other non-macro identifiers are now written as
|
| 190 |
cln::foo instead of cl_foo, except where the cl_ comes from a data type
|
| 191 |
(as in cl_I_to_int()). Headers are installed into a separate
|
| 192 |
directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
|
| 193 |
must be manually ported to the new style. We apologize for the
|
| 194 |
inconvenience.
|
| 195 |
|
| 196 |
Implementation changes
|
| 197 |
----------------------
|
| 198 |
|
| 199 |
* Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
|
| 200 |
Configuring with --with-gmp now relies on an installed libgmp library
|
| 201 |
version 3 or above. We recommend libgmp-3.1 or newer.
|
| 202 |
|
| 203 |
* Adjusted several break-even points to make better use of a faster libgmp and
|
| 204 |
better match present-day CPUs.
|
| 205 |
|
| 206 |
* Fix several errors for architectures with 64-bit wide words.
|
| 207 |
(Others than Alpha, which worked already.)
|
| 208 |
|
| 209 |
* Fix compilation errors with current GCC-3.0 snapshots.
|
| 210 |
|
| 211 |
Other changes
|
| 212 |
-------------
|
| 213 |
|
| 214 |
* Added package tools: script cln-config and autoconf macro AC_PATH_CLN
|
| 215 |
(in file cln.m4).
|
| 216 |
|
| 217 |
|
| 218 |
2000-01-13, version 1.0.3
|
| 219 |
=========================
|
| 220 |
|
| 221 |
Functional changes
|
| 222 |
------------------
|
| 223 |
|
| 224 |
* New function
|
| 225 |
`cl_I doublefactorial (uintL n)'.
|
| 226 |
|
| 227 |
Implementation changes
|
| 228 |
----------------------
|
| 229 |
|
| 230 |
* Fix several little configuration errors.
|
| 231 |
|
| 232 |
* Fix some compilation errors with gcc-2.96 prereleases.
|
| 233 |
|
| 234 |
|
| 235 |
1999-09-07, version 1.0.2
|
| 236 |
=========================
|
| 237 |
|
| 238 |
Functional changes
|
| 239 |
------------------
|
| 240 |
|
| 241 |
* New functions
|
| 242 |
`cl_I numerator (const cl_RA&)'
|
| 243 |
`cl_I denominator (const cl_RA&)'.
|
| 244 |
Suggested by Richard Kreckel and Sylvain Pion.
|
| 245 |
|
| 246 |
* New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
|
| 247 |
cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
|
| 248 |
Suggested by Richard Kreckel.
|
| 249 |
|
| 250 |
Implementation changes
|
| 251 |
----------------------
|
| 252 |
|
| 253 |
* Fix an endless loop when either of the functions
|
| 254 |
`cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
|
| 255 |
`cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
|
| 256 |
was called with an integer argument. Thanks to Richard Kreckel.
|
| 257 |
|
| 258 |
* Fix a bug in the addition and subtraction of rational numbers which could
|
| 259 |
cause unnormalized rational numbers like 3/1 to be created.
|
| 260 |
|
| 261 |
|
| 262 |
1999-06-09, version 1.0.1
|
| 263 |
=========================
|
| 264 |
|
| 265 |
Algorithmic changes
|
| 266 |
-------------------
|
| 267 |
|
| 268 |
* Speed up the functions `logand', `lognand', `logandc2'.
|
| 269 |
|
| 270 |
Implementation changes
|
| 271 |
----------------------
|
| 272 |
|
| 273 |
* Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
|
| 274 |
|
| 275 |
* Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
|
| 276 |
and SPARC with LiDIA.
|
| 277 |
|
| 278 |
* Fix compilation errors on Win32. Thanks to Mumit Khan.
|
| 279 |
|
| 280 |
|
| 281 |
1999-01-12, version 1.0
|
| 282 |
=======================
|
| 283 |
|
| 284 |
Functional changes
|
| 285 |
------------------
|
| 286 |
|
| 287 |
* New include files, containing I/O functions:
|
| 288 |
<cl_integer_io.h>
|
| 289 |
<cl_rational_io.h>
|
| 290 |
<cl_float_io.h>
|
| 291 |
<cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
|
| 292 |
<cl_real_io.h>
|
| 293 |
<cl_complex_io.h>
|
| 294 |
The file <cl_io.h> now contains only I/O functions for characters and
|
| 295 |
C integers.
|
| 296 |
|
| 297 |
* To access the base ring of a univariate polynomial ring, now use
|
| 298 |
`R->basering()' instead of `R->basering'.
|
| 299 |
|
| 300 |
* Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
|
| 301 |
cl_LF, cl_RA, cl_I.
|
| 302 |
|
| 303 |
* Implement `abs' for the class cl_RA.
|
| 304 |
|
| 305 |
* Replaced `read_number' with specialized functions `read_complex',
|
| 306 |
`read_real', `read_float', `read_rational', `read_integer'.
|
| 307 |
|
| 308 |
* Replaced the functions
|
| 309 |
`void fprint (cl_ostream stream, int x)'
|
| 310 |
`void fprint (cl_ostream stream, unsigned int x)'
|
| 311 |
`void fprint (cl_ostream stream, long x)'
|
| 312 |
`void fprint (cl_ostream stream, unsigned long x)'
|
| 313 |
with
|
| 314 |
`void fprintdecimal (cl_ostream stream, int x)'
|
| 315 |
`void fprintdecimal (cl_ostream stream, unsigned int x)'
|
| 316 |
`void fprintdecimal (cl_ostream stream, long x)'
|
| 317 |
`void fprintdecimal (cl_ostream stream, unsigned long x)'
|
| 318 |
|
| 319 |
Algorithmic changes
|
| 320 |
-------------------
|
| 321 |
|
| 322 |
* The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
|
| 323 |
now normalizes the resulting u and v to be of minimum absolute value.
|
| 324 |
(Suggested by the LiDIA group.)
|
| 325 |
|
| 326 |
* The conversion from string to cl_F, cl_R, cl_N now chooses the format of
|
| 327 |
the resulting float depending on the number of significant digits present
|
| 328 |
in the input string.
|
| 329 |
|
| 330 |
* Speed up the functions
|
| 331 |
`cl_R operator* (const cl_RA&, const cl_LF&)'
|
| 332 |
`cl_R operator* (const cl_LF&, const cl_RA&)'
|
| 333 |
`cl_R operator/ (const cl_RA&, const cl_LF&)'
|
| 334 |
`cl_LF operator/ (const cl_LF&, const cl_RA&)'
|
| 335 |
|
| 336 |
Implementation changes
|
| 337 |
----------------------
|
| 338 |
|
| 339 |
* Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
|
| 340 |
2^29 and 2^37.
|
| 341 |
(Reported by Thomas Papanikolaou.)
|
| 342 |
|
| 343 |
* Fixed a memory leak in the long-float truncate function.
|
| 344 |
(Reported by Keith Briggs.)
|
| 345 |
|
| 346 |
* Fixed a memory leak in the modular integer exponentiation.
|
| 347 |
|
| 348 |
* Reduced the stack space consumption. The default stack size (typically
|
| 349 |
8 MB on Unix) should now be sufficient in order to run all CLN programs.
|
| 350 |
|
| 351 |
* Portability fixes for m68k platforms.
|
| 352 |
(Reported and fixed by Roman Hodek.)
|
| 353 |
|
| 354 |
Other changes
|
| 355 |
-------------
|
| 356 |
|
| 357 |
* Restructured the sources. The subpackages
|
| 358 |
- base
|
| 359 |
- base + integer
|
| 360 |
- base + integer + rational
|
| 361 |
- base + integer + rational + float
|
| 362 |
- base + integer + rational + float + real
|
| 363 |
- base + integer + rational + float + real + complex
|
| 364 |
- base + integer + modinteger
|
| 365 |
are now self-contained.
|
| 366 |
|
| 367 |
* The configuration script can be called with option `--enable-shared',
|
| 368 |
to build CLN as a shared library.
|
| 369 |
|
| 370 |
* Updated the LiDIA interface.
|
| 371 |
|
| 372 |
|
| 373 |
1998-05-07, version 0.98
|
| 374 |
========================
|
| 375 |
|
| 376 |
Implementation changes
|
| 377 |
----------------------
|
| 378 |
|
| 379 |
* Portability fixes for 64-bit platforms like DEC Alpha.
|
| 380 |
(Reported by John Cremona and Thomas Papanikolaou.)
|
| 381 |
|
| 382 |
|
| 383 |
1998-02-27
|
| 384 |
==========
|
| 385 |
|
| 386 |
Other changes
|
| 387 |
-------------
|
| 388 |
|
| 389 |
* Portability fixes for new ISO/ANSI C++ compilers.
|
| 390 |
|
| 391 |
* Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
|
| 392 |
bugs.
|
| 393 |
|
| 394 |
|
| 395 |
1997-09-06
|
| 396 |
==========
|
| 397 |
|
| 398 |
Implementation changes
|
| 399 |
----------------------
|
| 400 |
|
| 401 |
* Portability fix for i686 platforms.
|
| 402 |
(Reported by Michael Pfeifer.)
|
| 403 |
|
| 404 |
|
| 405 |
1997-09-01
|
| 406 |
==========
|
| 407 |
|
| 408 |
Functional changes
|
| 409 |
------------------
|
| 410 |
|
| 411 |
* New include files:
|
| 412 |
<cl_ring.h>, declaring general rings,
|
| 413 |
<cl_GV.h>, <cl_SV.h>, defining vectors,
|
| 414 |
<cl_univpoly.h>, defining univariate polynomial rings.
|
| 415 |
|
| 416 |
Algorithmic changes
|
| 417 |
-------------------
|
| 418 |
|
| 419 |
* Speed up the internal computation of ln(2) and ln(10) by about 20%.
|
| 420 |
|
| 421 |
* Speed up the computation of exp (for 1000 digits or more) by about 10%.
|
| 422 |
|
| 423 |
Implementation changes
|
| 424 |
----------------------
|
| 425 |
|
| 426 |
* Portability fix for 64-bit platforms like DEC Alpha:
|
| 427 |
Fixed a bug: endless loop in `sqrt' for large precision.
|
| 428 |
(Reported by Blake Jones.)
|
| 429 |
|
| 430 |
* Portability fix for RS/6000 platforms.
|
| 431 |
|
| 432 |
|
| 433 |
1997-02-21
|
| 434 |
==========
|
| 435 |
|
| 436 |
Algorithmic changes
|
| 437 |
-------------------
|
| 438 |
|
| 439 |
* Speed up the computation of the elementary transcendental functions:
|
| 440 |
exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
|
| 441 |
log (for 100 digits) by about 25%,
|
| 442 |
sin, cos (for 1000 or 10000 digits) by about 20%,
|
| 443 |
asin, acos, atan (for 1000 digits) by about 25%.
|
| 444 |
|
| 445 |
Implementation changes
|
| 446 |
----------------------
|
| 447 |
|
| 448 |
* Portability fix for SunOS 4 and other BSD platforms.
|
| 449 |
|
| 450 |
|
| 451 |
1997-01-31
|
| 452 |
==========
|
| 453 |
|
| 454 |
Algorithmic changes
|
| 455 |
-------------------
|
| 456 |
|
| 457 |
* Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
|
| 458 |
|
| 459 |
* Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
|
| 460 |
for moderate precision (< 11370 digits) were correct.
|
| 461 |
(Reported by Thomas Papanikolaou.)
|
| 462 |
|
| 463 |
Implementation changes
|
| 464 |
----------------------
|
| 465 |
|
| 466 |
* Fix static initialization order bug. Requires g++ 2.7.0 or newer
|
| 467 |
(g++ 2.8.0 or newer on HPPA).
|
| 468 |
|
| 469 |
* New method `debug_print()', for pretty printing of CLN objects, intended
|
| 470 |
to be called from the debugger.
|
| 471 |
|
| 472 |
|
| 473 |
1997-01-07
|
| 474 |
==========
|
| 475 |
|
| 476 |
Functional changes
|
| 477 |
------------------
|
| 478 |
|
| 479 |
* New functions `float cl_float_approx (const cl_R&)',
|
| 480 |
`double cl_double_approx (const cl_R&)'
|
| 481 |
for converting a real number to a C `float' or `double'.
|
| 482 |
|
| 483 |
Algorithmic changes
|
| 484 |
-------------------
|
| 485 |
|
| 486 |
* Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
|
| 487 |
|
| 488 |
Implementation changes
|
| 489 |
----------------------
|
| 490 |
|
| 491 |
* New function `cl_print', for pretty printing of CLN objects, intended to be
|
| 492 |
called from the debugger.
|
| 493 |
|
| 494 |
* Portability fixes for 64-bit platforms like DEC Alpha.
|
| 495 |
|
| 496 |
|
| 497 |
1996-11-29
|
| 498 |
==========
|
| 499 |
|
| 500 |
Functional changes
|
| 501 |
------------------
|
| 502 |
|
| 503 |
* New include files:
|
| 504 |
<cl_string.h>, defining strings,
|
| 505 |
<cl_symbol.h>, defining symbols,
|
| 506 |
<cl_proplist.h>, defining property lists,
|
| 507 |
<cl_condition.h>, defining conditions/exceptions.
|
| 508 |
|
| 509 |
* New functions `cl_F cl_catalanconst ()',
|
| 510 |
`cl_F cl_catalanconst (const cl_F&)',
|
| 511 |
`cl_F cl_catalanconst (cl_float_format_t)'
|
| 512 |
which return Catalan's constant.
|
| 513 |
(Suggested by Thomas Papanikolaou.)
|
| 514 |
|
| 515 |
* New functions `cl_F cl_zeta (int s)',
|
| 516 |
`cl_F cl_zeta (int s, const cl_F&)',
|
| 517 |
`cl_F cl_zeta (int s, cl_float_format_t)'
|
| 518 |
which return the Riemann zeta function at an integral point s>1.
|
| 519 |
|
| 520 |
* New functions `cl_F cl_exp1 ()',
|
| 521 |
`cl_F cl_exp1 (const cl_F&)',
|
| 522 |
`cl_F cl_exp1 (cl_float_format_t)'
|
| 523 |
which return e = exp(1).
|
| 524 |
|
| 525 |
* New function `cl_I binomial (uintL n, uintL k)' which returns the
|
| 526 |
binomial coefficient (n choose k).
|
| 527 |
|
| 528 |
* New functions `int cl_I_to_int (const cl_I&)',
|
| 529 |
`unsigned int cl_I_to_uint (const cl_I&)',
|
| 530 |
`long cl_I_to_long (const cl_I&)',
|
| 531 |
`unsigned long cl_I_to_ulong (const cl_I&)'
|
| 532 |
for converting an integer to a C `int' or `long'.
|
| 533 |
|
| 534 |
* New functions `float cl_float_approx (const cl_I&)',
|
| 535 |
`float cl_float_approx (const cl_RA&)',
|
| 536 |
`double cl_double_approx (const cl_I&)',
|
| 537 |
`double cl_double_approx (const cl_RA&)'
|
| 538 |
for converting a rational number to a C `float' or `double'.
|
| 539 |
|
| 540 |
Implementation changes
|
| 541 |
----------------------
|
| 542 |
|
| 543 |
* Moved the sources to subdirectories.
|
| 544 |
(Suggested by Jörg Arndt.)
|
| 545 |
|
| 546 |
Other changes
|
| 547 |
-------------
|
| 548 |
|
| 549 |
* New benchmark for better comparing LiDIA, Pari and CLN.
|
| 550 |
|
| 551 |
* Added an interface to LiDIA, allows using CLN as a replacement of libI.
|
| 552 |
(Suggested by Thomas Papanikolaou.)
|
| 553 |
|
| 554 |
* Added an ILOG Talk interface, for interactive use of the library.
|
| 555 |
|
| 556 |
|
| 557 |
1996-10-13
|
| 558 |
==========
|
| 559 |
|
| 560 |
Functional changes
|
| 561 |
------------------
|
| 562 |
|
| 563 |
* New include file <cl_modinteger.h>, defining modular integer rings.
|
| 564 |
|
| 565 |
* New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
|
| 566 |
`cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
|
| 567 |
|
| 568 |
Algorithmic changes
|
| 569 |
-------------------
|
| 570 |
|
| 571 |
* Speed up square root of large numbers by use of Newton's algorithm.
|
| 572 |
|
| 573 |
* Speed up multiplication and division of large long-floats by small integers.
|
| 574 |
|
| 575 |
* Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
|
| 576 |
asin, acos, atan, asinh, acosh, atanh.
|
| 577 |
All of them are now computed with asymptotically fast algorithms, of
|
| 578 |
bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
|
| 579 |
|
| 580 |
* Fixed several bugs in the transcendental functions routines. In particular,
|
| 581 |
the `log' function went into an endless loop.
|
| 582 |
|
| 583 |
* Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
|
| 584 |
correctly for negative integers.
|
| 585 |
|
| 586 |
|
| 587 |
1996-09-08
|
| 588 |
==========
|
| 589 |
|
| 590 |
Functional changes
|
| 591 |
------------------
|
| 592 |
|
| 593 |
* New include file <cl_numtheory.h>, containing number theoretic functions,
|
| 594 |
for now only the Jacobi symbol.
|
| 595 |
|
| 596 |
Algorithmic changes
|
| 597 |
-------------------
|
| 598 |
|
| 599 |
* Speed up squaring of large numbers by 30%.
|
| 600 |
|
| 601 |
* Speed up division of large numbers by use of Newton's algorithm. The
|
| 602 |
improvement is noticeable for numbers with at least about 1000 words =
|
| 603 |
32000 bits = 10000 decimal digits.
|
| 604 |
|
| 605 |
* Speed up the binary-to-decimal conversion of large numbers by use of
|
| 606 |
a divide-and-conquer method. The improvement is noticeable for numbers
|
| 607 |
of at least 1250 words = 40000 bits = 12000 decimal digits.
|
| 608 |
|
| 609 |
* Fixed a bug: The functions `cl_boolean minusp (const type&)' and
|
| 610 |
`cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
|
| 611 |
|
| 612 |
* Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
|
| 613 |
didn't work correctly if both arguments were bignums.
|
| 614 |
|
| 615 |
* Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
|
| 616 |
as `mod(-n*m,m)' (n>0), returned m instead of 0.
|
| 617 |
|
| 618 |
* Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
|
| 619 |
values for arguments x = 2^n, n >= 29.
|
| 620 |
|
| 621 |
Implementation changes
|
| 622 |
----------------------
|
| 623 |
|
| 624 |
* Speed up CLN by using the low-level addition/subtraction, shift and
|
| 625 |
multiplication loops from GNU gmp-2.0.2.
|
| 626 |
|
| 627 |
* Fixed a severe reference counting bug in the constructors
|
| 628 |
`type::type (const long)' and `type::type (const unsigned long)'
|
| 629 |
for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
|
| 630 |
caused a core dump.
|
| 631 |
|
| 632 |
* CLN's word sequences can be chosen to be stored big-endian in memory
|
| 633 |
(like it was up to now) or little-endian (needed for interfacing to GMP).
|
| 634 |
|
| 635 |
|
| 636 |
1996-05-20
|
| 637 |
==========
|
| 638 |
|
| 639 |
Functional changes
|
| 640 |
------------------
|
| 641 |
|
| 642 |
* New include file <cl_timing.h>, contains timing facilities.
|
| 643 |
|
| 644 |
Algorithmic changes
|
| 645 |
-------------------
|
| 646 |
|
| 647 |
* Speed up the multiplication of large numbers by use of an FFT based
|
| 648 |
algorithm (Schönhage-Strassen). The improvement is noticeable when both
|
| 649 |
factors have at least 1250 words = 40000 bits = 12000 decimal digits.
|
| 650 |
|
| 651 |
* Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
|
| 652 |
`cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
|
| 653 |
of 1.5. Simple trick.
|
| 654 |
|
| 655 |
* Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
|
| 656 |
trick from Henri Cohen.
|
| 657 |
|
| 658 |
* Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
|
| 659 |
which occurred when one of the arguments is zero.
|
| 660 |
|
| 661 |
* Fixed an endless loop in the function `cl_I factorial (uintL)'.
|
| 662 |
|
| 663 |
Implementation changes
|
| 664 |
----------------------
|
| 665 |
|
| 666 |
* <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
|
| 667 |
(Reported by Tito Flagella.)
|
| 668 |
|
| 669 |
* Fixed linking error when compiling without -O.
|
| 670 |
(Reported by Tito Flagella.)
|
| 671 |
|
| 672 |
* Fixed syntax error in <cl_input.h>.
|
| 673 |
(Reported by Tito Flagella.)
|
| 674 |
|
| 675 |
* Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
|
| 676 |
(Reported by Marcus Daniels.)
|
| 677 |
|
| 678 |
* More portability fixes.
|
| 679 |
|
| 680 |
* Configuration now supports the --srcdir option.
|
| 681 |
(Reported by Marcus Daniels.)
|
| 682 |
|
| 683 |
|
| 684 |
1996-01-03
|
| 685 |
==========
|
| 686 |
|
| 687 |
* Compiles well on Sparc, using gcc.
|
| 688 |
|
| 689 |
* Mail to beta testers.
|
| 690 |
|
| 691 |
|
| 692 |
1995-12-06
|
| 693 |
==========
|
| 694 |
|
| 695 |
* Compiles well on Linux, using gcc.
|
| 696 |
|
Parent Directory
| 
Revision Log