| 1 | <html> |
|---|
| 2 | <head> |
|---|
| 3 | <title> |
|---|
| 4 | A Tour of NTL: Summary of Changes </title> |
|---|
| 5 | </head> |
|---|
| 6 | |
|---|
| 7 | <body bgcolor="#fff9e6"> |
|---|
| 8 | <center> |
|---|
| 9 | <a href="tour-roadmap.html"><img src="arrow1.gif" alt="[Previous]" align=bottom></a> |
|---|
| 10 | <a href="tour.html"><img src="arrow2.gif" alt="[Up]" align=bottom></a> |
|---|
| 11 | <a href="tour-ack.html"> <img src="arrow3.gif" alt="[Next]" align=bottom></a> |
|---|
| 12 | </center> |
|---|
| 13 | |
|---|
| 14 | <h1> |
|---|
| 15 | <p align=center> |
|---|
| 16 | A Tour of NTL: Summary of Changes |
|---|
| 17 | </p> |
|---|
| 18 | </h1> |
|---|
| 19 | |
|---|
| 20 | <p> <hr> <p> |
|---|
| 21 | <h3> |
|---|
| 22 | 2009.05.05: Changes between NTL 5.5 and 5.5.1 |
|---|
| 23 | </h3> |
|---|
| 24 | |
|---|
| 25 | <ul> |
|---|
| 26 | <li> If using GMP (via either <tt>NTL_GMP_LIP</tt> |
|---|
| 27 | or <tt>NTL_GMP_HACK</tt>), then the new version (4.3.0) of |
|---|
| 28 | GMP implements the <tt>XGCD</tt> functionality differently, |
|---|
| 29 | so that the coefficients do not always agree with those returned by |
|---|
| 30 | the classical extended Euclidean algorithm. |
|---|
| 31 | This version of NTL corrects the coefficients, so that the |
|---|
| 32 | "classical" coefficients are always produced, regardless |
|---|
| 33 | of GMP's implementation. |
|---|
| 34 | This version of NTL also works |
|---|
| 35 | around a bug in GMP 4.3.0's <tt>XGCD</tt> code |
|---|
| 36 | (although that bug should be fixed in GMP 4.3.1). |
|---|
| 37 | |
|---|
| 38 | <li> |
|---|
| 39 | The <tt>configure</tt> script has been slightly modified: |
|---|
| 40 | there is a new configuration variable <tt>DEF_PREFIX</tt>, |
|---|
| 41 | whose value can be used to set <tt>PREFIX</tt>, <tt>GMP_PREFIX</tt>, |
|---|
| 42 | and <tt>GF2X_PREFIX</tt> in one stroke. |
|---|
| 43 | Also, the (somewhat esoteric) <tt>configure</tt> variables |
|---|
| 44 | <tt>GMP_LIBDIR</tt>, <tt>GMP_INCDIR</tt>, |
|---|
| 45 | <tt>GF2X_LIBDIR</tt>, and <tt>GF2X_INCDIR</tt> |
|---|
| 46 | have slightly different meanings now. |
|---|
| 47 | |
|---|
| 48 | </ul> |
|---|
| 49 | </h3> |
|---|
| 50 | |
|---|
| 51 | <p> <hr> <p> |
|---|
| 52 | <h3> |
|---|
| 53 | 2009.04.08: Changes between NTL 5.4.2 and 5.5 |
|---|
| 54 | </h3> |
|---|
| 55 | |
|---|
| 56 | <ul> |
|---|
| 57 | <li> |
|---|
| 58 | Added the ability to generate a <i>shared</i> library |
|---|
| 59 | (with help from Tim Abbott). <a href="tour-unix.html#shared">Details.</a> |
|---|
| 60 | |
|---|
| 61 | <li> |
|---|
| 62 | Fixed some standardization issues |
|---|
| 63 | (with help from Tim Abbot): |
|---|
| 64 | default location of installed documentation files now conforms |
|---|
| 65 | to standards; use of <tt>EOF</tt> now conforms to standards. |
|---|
| 66 | |
|---|
| 67 | <li> |
|---|
| 68 | Added a callback mechanism to NTL's error reporting function. |
|---|
| 69 | See <tt>ErrorCallback</tt> in <a href="tools.txt">tools.txt</a>. |
|---|
| 70 | |
|---|
| 71 | <li> |
|---|
| 72 | Added support for the <tt>gf2x</tt> library for speeding up |
|---|
| 73 | arithmetic in <tt>GF2X</tt> (with help from Emmanuel Thomé). |
|---|
| 74 | <a href="tour-gf2x.html">Details.</a> |
|---|
| 75 | |
|---|
| 76 | <li> |
|---|
| 77 | In conjuction with the above, I also changed the |
|---|
| 78 | <tt>GF2X</tt> so that it works better with very large polynomials: |
|---|
| 79 | large blocks of memory are released, recursive HalfGCD algorithms |
|---|
| 80 | are used for large polynomials. |
|---|
| 81 | |
|---|
| 82 | |
|---|
| 83 | <li> |
|---|
| 84 | Fixed a bug in <tt>void TraceMod(zz_p& x, const zz_pX& a, const zz_pXModulus& F)</tt> (reported by Luca De Feo). |
|---|
| 85 | |
|---|
| 86 | <li> |
|---|
| 87 | Fixed a performance issue in various versions of <tt>SetCoeff</tt> |
|---|
| 88 | (reported by Luca De Feo). |
|---|
| 89 | |
|---|
| 90 | <li> |
|---|
| 91 | Fixed the declaration of <tt>mat_zz_p transpose(const mat_zz_p& a)</tt> |
|---|
| 92 | (reported by Benoit Lacelle). |
|---|
| 93 | </ul> |
|---|
| 94 | |
|---|
| 95 | |
|---|
| 96 | <p> <hr> <p> |
|---|
| 97 | <h3> |
|---|
| 98 | 2008.03.05: Changes between NTL 5.4.1 and 5.4.2 |
|---|
| 99 | </h3> |
|---|
| 100 | |
|---|
| 101 | <ul> |
|---|
| 102 | <li> |
|---|
| 103 | Fixed a bug in the <tt>sub(ZZ_pEX, ZZ_pE, ZZ_pEX)</tt> |
|---|
| 104 | and <tt>sub(zz_pEX, zz_pE, zz_pEX)</tt> routines (reported by Charanjit Jutla). |
|---|
| 105 | Under certain circumstances, these could outout wrong answers. |
|---|
| 106 | |
|---|
| 107 | </ul> |
|---|
| 108 | |
|---|
| 109 | <p> <hr> <p> |
|---|
| 110 | <h3> |
|---|
| 111 | 2007.05.09: Changes between NTL 5.4 and 5.4.1 |
|---|
| 112 | </h3> |
|---|
| 113 | |
|---|
| 114 | <ul> |
|---|
| 115 | <li> |
|---|
| 116 | Fixed rounding bug in <tt>expm1</tt> (reported by Paul Zimmermann). |
|---|
| 117 | |
|---|
| 118 | <li> |
|---|
| 119 | Fixed memory leak in several LLL routines (reported by Friedrich Bahr). |
|---|
| 120 | |
|---|
| 121 | <li> |
|---|
| 122 | Fixed infinite loop in several LLL routines |
|---|
| 123 | (this only occurred on machines, like x86, with double rounding). |
|---|
| 124 | |
|---|
| 125 | <li> |
|---|
| 126 | Improved <tt>GF2X</tt> timing tests (suggested by Paul Zimmermann). |
|---|
| 127 | |
|---|
| 128 | </ul> |
|---|
| 129 | |
|---|
| 130 | <p> <hr> <p> |
|---|
| 131 | <h3> |
|---|
| 132 | 2005.03.24: Changes between NTL 5.3.2 and 5.4 |
|---|
| 133 | </h3> |
|---|
| 134 | |
|---|
| 135 | <ul> |
|---|
| 136 | <li> |
|---|
| 137 | By default, NTL now compiles in ISO mode (using namespaces, etc.). |
|---|
| 138 | You can always revert to traditional mode by unsetting |
|---|
| 139 | the flag <tt>NTL_STD_CXX</tt> |
|---|
| 140 | (either pass <tt>NTL_STD_CXX=off</tt> to the configure script, |
|---|
| 141 | or manually edit the <tt>config.h</tt> file). |
|---|
| 142 | <p> |
|---|
| 143 | |
|---|
| 144 | <li> |
|---|
| 145 | Some bug fixes: |
|---|
| 146 | |
|---|
| 147 | <ul> |
|---|
| 148 | <li> |
|---|
| 149 | The <tt>sqrt</tt> and <tt>log1p</tt> routines |
|---|
| 150 | for the <tt>RR</tt> class would produce incorrectly rounded |
|---|
| 151 | results in certain circumstances (although this only affected the relative |
|---|
| 152 | error of the result very marginally). |
|---|
| 153 | <li> |
|---|
| 154 | The <tt>SqrRootPrec</tt> routine for the <tt>RR</tt> class |
|---|
| 155 | could not be called, because it was defined incorrectly. |
|---|
| 156 | </ul> |
|---|
| 157 | |
|---|
| 158 | <p> |
|---|
| 159 | |
|---|
| 160 | Thanks to Paul Zimmermann for finding (and fixing) these bugs! |
|---|
| 161 | Paul has also validated NTL's <tt>RR</tt> class by cross-checking it with the |
|---|
| 162 | <a href="http://www.mpfr.org">MPFR</a> library. |
|---|
| 163 | |
|---|
| 164 | <p> |
|---|
| 165 | <li> |
|---|
| 166 | Some performance enhancements: |
|---|
| 167 | |
|---|
| 168 | <ul> |
|---|
| 169 | <li> |
|---|
| 170 | Added a new <tt>MulModPrecon</tt> inline function for |
|---|
| 171 | computing <tt>(a * b) % n</tt> for single precision numbers, |
|---|
| 172 | when <tt>b</tt> and <tt>n</tt> are fixed for several computations. |
|---|
| 173 | On some platforms this can be twice as fast or more than the |
|---|
| 174 | old <tt>MulMod2</tt> routine. |
|---|
| 175 | This indirectly affects a lot of computations that are done via |
|---|
| 176 | homomorphic imaging (polynomial multiplication |
|---|
| 177 | over <tt>zz_p</tt>, <tt>ZZ_p</tt>, and <tt>ZZ</tt>, |
|---|
| 178 | matrix computations over <tt>zz_p</tt> and <tt>ZZ</tt>). |
|---|
| 179 | |
|---|
| 180 | <li> |
|---|
| 181 | Rewrote the small prime FFT to take advantage of the new |
|---|
| 182 | <tt>MulModPrecon</tt>, and to be more cache friendly. |
|---|
| 183 | |
|---|
| 184 | <li> |
|---|
| 185 | Improved the performance of the <tt>GF2X</tt> multiplication routine. |
|---|
| 186 | On some platforms, it can be twice as fast as the old one. |
|---|
| 187 | Thanks (again) to Paul Zimmermann for suggesting some of these |
|---|
| 188 | improvements and supplying some of the code. |
|---|
| 189 | |
|---|
| 190 | </ul> |
|---|
| 191 | |
|---|
| 192 | <p> |
|---|
| 193 | <li> |
|---|
| 194 | Miscellany: |
|---|
| 195 | |
|---|
| 196 | <ul> |
|---|
| 197 | <li> |
|---|
| 198 | Rewrote several of the installation scripts in Perl (the old shell |
|---|
| 199 | scripts were getting too messy to maintain). |
|---|
| 200 | However, the syntax for all of the command-line interfaces |
|---|
| 201 | remains identical. |
|---|
| 202 | |
|---|
| 203 | </ul> |
|---|
| 204 | |
|---|
| 205 | |
|---|
| 206 | |
|---|
| 207 | |
|---|
| 208 | |
|---|
| 209 | </ul> |
|---|
| 210 | |
|---|
| 211 | <p> <hr> <p> |
|---|
| 212 | <h3> |
|---|
| 213 | 2004.05.21: Changes between NTL 5.3.1 and 5.3.2 |
|---|
| 214 | </h3> |
|---|
| 215 | |
|---|
| 216 | <ul> |
|---|
| 217 | <li> |
|---|
| 218 | Some bug fixes. |
|---|
| 219 | |
|---|
| 220 | <p> |
|---|
| 221 | <li> |
|---|
| 222 | Re-wrote <tt>SqrRootMod</tt> to make it run faster. |
|---|
| 223 | |
|---|
| 224 | </ul> |
|---|
| 225 | |
|---|
| 226 | |
|---|
| 227 | |
|---|
| 228 | <p> <hr> <p> |
|---|
| 229 | <h3> |
|---|
| 230 | 2002.12.17: Changes between NTL 5.3 and 5.3.1 |
|---|
| 231 | </h3> |
|---|
| 232 | |
|---|
| 233 | <ul> |
|---|
| 234 | <li> |
|---|
| 235 | Fixed a bug affecting the <tt>BuildIrred</tt> routines |
|---|
| 236 | for <tt>ZZ_pEX</tt> and <tt>zz_pEX</tt>. |
|---|
| 237 | </ul> |
|---|
| 238 | |
|---|
| 239 | <p> <hr> <p> |
|---|
| 240 | <h3> |
|---|
| 241 | 2002.07.05: Changes between NTL 5.2 and 5.3 |
|---|
| 242 | </h3> |
|---|
| 243 | |
|---|
| 244 | <ul> |
|---|
| 245 | <li> |
|---|
| 246 | Minimized and isolated constructs that do not adhere to <tt>C</tt>/<tt>C++</tt> |
|---|
| 247 | standards, |
|---|
| 248 | and added flags <tt>NTL_CLEAN_INT</tt> and <tt>NTL_CLEAN_PTR</tt> |
|---|
| 249 | which force stricter compliance with these standards |
|---|
| 250 | <a href="tour-impl.html">[more details]</a>. |
|---|
| 251 | |
|---|
| 252 | <p> |
|---|
| 253 | <li> |
|---|
| 254 | Added functions <tt>IsWhiteSpace</tt>, <tt>CharToIntVal</tt>, |
|---|
| 255 | and <tt>IntValToChar</tt> to the <tt>tools</tt> module |
|---|
| 256 | <a href="tools.txt">[more details]</a>. |
|---|
| 257 | |
|---|
| 258 | <p> |
|---|
| 259 | <li> |
|---|
| 260 | Added methods <tt>allocated</tt>, <tt>position1</tt> to generic vector classes |
|---|
| 261 | <a href="vector.txt">[more details]</a>. |
|---|
| 262 | |
|---|
| 263 | <p> |
|---|
| 264 | <li> |
|---|
| 265 | Added method <tt>allocated</tt> to the class <tt>vec_GF2</tt> |
|---|
| 266 | <a href="vec_GF2.txt">[more details]</a>. |
|---|
| 267 | |
|---|
| 268 | <p> |
|---|
| 269 | <li> |
|---|
| 270 | Added conversion routines from unsigned int/long to int, long, float, and double |
|---|
| 271 | <a href="conversions.txt">[more details]</a>. |
|---|
| 272 | |
|---|
| 273 | <p> |
|---|
| 274 | <li> |
|---|
| 275 | Added routines <tt>AddPrec</tt>, <tt>SubPrec</tt>, etc., to the <tt>RR</tt> |
|---|
| 276 | module, and declared the practice of directly assigning to the variable |
|---|
| 277 | <tt>RR::prec</tt> obsolete |
|---|
| 278 | <a href="RR.txt">[more details]</a>. |
|---|
| 279 | |
|---|
| 280 | <p> |
|---|
| 281 | <li> |
|---|
| 282 | Fixed a number of minor bugs. |
|---|
| 283 | |
|---|
| 284 | </ul> |
|---|
| 285 | |
|---|
| 286 | |
|---|
| 287 | |
|---|
| 288 | <p> <hr> <p> |
|---|
| 289 | <h3> |
|---|
| 290 | 2001.07.19: Changes between NTL 5.1a and 5.2 |
|---|
| 291 | </h3> |
|---|
| 292 | |
|---|
| 293 | <p> |
|---|
| 294 | |
|---|
| 295 | <ul> |
|---|
| 296 | <li> |
|---|
| 297 | Implemented Mark van Hoeij's new algorithm for factorining polynomials |
|---|
| 298 | with rational coefficients. |
|---|
| 299 | This new algorithm is much more efficient than the previous algorithm |
|---|
| 300 | used by NTL, and is the default (one can switch back to the old algorithm |
|---|
| 301 | with a run-time switch). |
|---|
| 302 | <p> |
|---|
| 303 | <a href="ZZXFactoring.txt">[documentation]</a> |
|---|
| 304 | <p> |
|---|
| 305 | <a href="tour-time.html">[performance measurements]</a> |
|---|
| 306 | <p> |
|---|
| 307 | |
|---|
| 308 | <li> |
|---|
| 309 | Added routines <tt>LLL_plus</tt> that are just like the all-integer |
|---|
| 310 | <tt>LLL</tt> routines, except that they return the exact values of the |
|---|
| 311 | squared lengths of the Gramm-Schmidt basis vectors. |
|---|
| 312 | This is useful in implementing van Hoeij's algorithm. |
|---|
| 313 | <a href="LLL.txt">[more details]</a>. |
|---|
| 314 | <p> |
|---|
| 315 | |
|---|
| 316 | <li> |
|---|
| 317 | Made a small change to <tt>quad_float.c</tt> to make it compile |
|---|
| 318 | under gcc version 3.0 |
|---|
| 319 | without errors. |
|---|
| 320 | This is the <i>one</i> place in NTL where I resort to just a little |
|---|
| 321 | assmebly code (but only on x86/Linux platforms), and wouldn't you know it, |
|---|
| 322 | this is the <i>one</i> place where gcc 3.0 had problems. |
|---|
| 323 | <p> |
|---|
| 324 | |
|---|
| 325 | <li> |
|---|
| 326 | Made a small change to the procedure for generating a distribution, |
|---|
| 327 | so that now all files in the "tar" file comprising the distribution |
|---|
| 328 | come without any annoyingly excessive access control restrictions. |
|---|
| 329 | <p> |
|---|
| 330 | |
|---|
| 331 | <li> |
|---|
| 332 | Changed the version numbering scheme so that it is now closer to |
|---|
| 333 | "standard practice". |
|---|
| 334 | This is version "5.2". |
|---|
| 335 | Any small bug fixes to this version will be named "5.2.1", "5.2.2", etc. |
|---|
| 336 | Also, macros are now defined so that the numerical components |
|---|
| 337 | of the version number are available to the programmer. |
|---|
| 338 | <a href="version.txt">[more details]</a>. |
|---|
| 339 | |
|---|
| 340 | |
|---|
| 341 | </ul> |
|---|
| 342 | |
|---|
| 343 | |
|---|
| 344 | <p> <hr> <p> |
|---|
| 345 | <h3> |
|---|
| 346 | 2001.06.08: Changes between NTL 5.0c and 5.1a |
|---|
| 347 | </h3> |
|---|
| 348 | |
|---|
| 349 | <p> |
|---|
| 350 | Some minor fixes and additions. |
|---|
| 351 | <p> |
|---|
| 352 | Completely backward compatible. |
|---|
| 353 | <p> |
|---|
| 354 | |
|---|
| 355 | <ul> |
|---|
| 356 | |
|---|
| 357 | <li> |
|---|
| 358 | Added a routine <tt>LatticeSolve()</tt> for finding integer |
|---|
| 359 | solutions to linear systems of integer equations. |
|---|
| 360 | <a href="LLL.txt">[more details]</a> |
|---|
| 361 | |
|---|
| 362 | <p> |
|---|
| 363 | <li> |
|---|
| 364 | Modified the stragey used by the <tt>LLL()</tt> and <tt>image()</tt> |
|---|
| 365 | routines in the <a href="LLL.txt">LLL package</a> to deal |
|---|
| 366 | with linear dependencies. |
|---|
| 367 | The new strategy guarantees better worst-case bounds on the |
|---|
| 368 | sizes of intermediate values. |
|---|
| 369 | I'm not sure if it will have any serious practical impact, though. |
|---|
| 370 | |
|---|
| 371 | <p> |
|---|
| 372 | <li> |
|---|
| 373 | Added some "partial ISO modes" so that one can use |
|---|
| 374 | some of the features of Standard <tt>C++</tt>, |
|---|
| 375 | even if ones compiler does not yet support all of the features. |
|---|
| 376 | <a href="tour-stdcxx.html">[more details]</a> |
|---|
| 377 | |
|---|
| 378 | <p> |
|---|
| 379 | <li> |
|---|
| 380 | Bug fix: routine <tt>determnant()</tt> in <tt>mat_GF2.h</tt> |
|---|
| 381 | was not visible to the linker because of a typo in <tt>mat_GF2.c</tt>. |
|---|
| 382 | |
|---|
| 383 | <p> |
|---|
| 384 | <li> |
|---|
| 385 | Made a "smarter" script for selecting the <tt>GetTime()</tt> |
|---|
| 386 | function. |
|---|
| 387 | This fixes an installation problem on Cygwin/Windows 95 platforms. |
|---|
| 388 | I hope it doesn't create more problems than it solves, though. |
|---|
| 389 | |
|---|
| 390 | <p> |
|---|
| 391 | <li> |
|---|
| 392 | Added some extra documentation for installation under |
|---|
| 393 | Windows/MS Visual <tt>C++</tt>. |
|---|
| 394 | <a href="tour-win.html">[more details]</a> |
|---|
| 395 | |
|---|
| 396 | <p> |
|---|
| 397 | <li> |
|---|
| 398 | Changed some names like <tt>c_lip.c</tt> to <tt>c_lip_impl.h</tt>. |
|---|
| 399 | This should avoid some potential installation problems. |
|---|
| 400 | |
|---|
| 401 | <p> |
|---|
| 402 | <li> |
|---|
| 403 | Throw away first 256-bytes of arc4 streams to improve quality of |
|---|
| 404 | the pseudo-random number generator. |
|---|
| 405 | This may change the precise behavior of some programs. |
|---|
| 406 | |
|---|
| 407 | <p> |
|---|
| 408 | <li> |
|---|
| 409 | Other minor, internal modifications. |
|---|
| 410 | |
|---|
| 411 | </ul> |
|---|
| 412 | |
|---|
| 413 | |
|---|
| 414 | |
|---|
| 415 | |
|---|
| 416 | <p> <hr> <p> |
|---|
| 417 | <h3> |
|---|
| 418 | 2001.02.19: Changes between NTL 5.0b and 5.0c |
|---|
| 419 | </h3> |
|---|
| 420 | |
|---|
| 421 | <p> |
|---|
| 422 | Fixed a naming problem in the Windows distribution. |
|---|
| 423 | The Unix distribution is unaffected. |
|---|
| 424 | |
|---|
| 425 | |
|---|
| 426 | <p> <hr> <p> |
|---|
| 427 | <h3> |
|---|
| 428 | 2001.02.19: Changes between NTL 5.0a and 5.0b |
|---|
| 429 | </h3> |
|---|
| 430 | |
|---|
| 431 | <p> |
|---|
| 432 | Fixed a typo in <tt>vec_ulong.c</tt> that causes a compile error |
|---|
| 433 | on some platforms. |
|---|
| 434 | |
|---|
| 435 | |
|---|
| 436 | <p> <hr> <p> |
|---|
| 437 | <h3> |
|---|
| 438 | 2001.02.19: Changes between NTL 4.3a and 5.0a |
|---|
| 439 | </h3> |
|---|
| 440 | |
|---|
| 441 | <p> |
|---|
| 442 | <ul> |
|---|
| 443 | <li> |
|---|
| 444 | I've now re-structured NTL so that one can use |
|---|
| 445 | either 'traditional' LIP or GMP as the <i>primary</i> long integer package. |
|---|
| 446 | Doing this introduced some (minor) backward incompatabilies in |
|---|
| 447 | the programming interface, so there is also a 'third way' -- you |
|---|
| 448 | can use GMP as a <i>supplemental</i> long integer package (as in NTL 4.3), |
|---|
| 449 | getting |
|---|
| 450 | many (but not all) of the performance benefits of GMP, while |
|---|
| 451 | maintaining <i>complete</i> backward compatability with the traditional |
|---|
| 452 | long integer package. |
|---|
| 453 | This 'third way' is not highly recommended -- it is only intended |
|---|
| 454 | as a backward compatabilty hack. |
|---|
| 455 | |
|---|
| 456 | <p> |
|---|
| 457 | Even if you do not use GMP, you should |
|---|
| 458 | <a href="tour-gmp.html">read about using NTL with GMP</a> so |
|---|
| 459 | that you can write code that works with either the traditional or GMP |
|---|
| 460 | long integer packages. |
|---|
| 461 | <p> |
|---|
| 462 | <li> |
|---|
| 463 | Added a <tt>ZZ</tt> to unsigned long conversion routine. |
|---|
| 464 | <a href="conversions.txt">[more details]</a> |
|---|
| 465 | <li> |
|---|
| 466 | Added new vector classes <tt>vec_ulong</tt> (vectors |
|---|
| 467 | of unsigned longs) and <tt>vec_vec_ulong</tt>. |
|---|
| 468 | <a href="tour-modules.html">[more details]</a> |
|---|
| 469 | <li> |
|---|
| 470 | Some minor bug fixes: under some unusual circumstances, a memory |
|---|
| 471 | allocation error could be erroneously raised; I also added a patch |
|---|
| 472 | that works around a bug in v3.0.1 of GMP. |
|---|
| 473 | <li> |
|---|
| 474 | Some internal cleansing, minimizing the use of non-standard constructs. |
|---|
| 475 | </ul> |
|---|
| 476 | |
|---|
| 477 | |
|---|
| 478 | <p> <hr> <p> |
|---|
| 479 | <h3> |
|---|
| 480 | Changes between NTL 4.2a and 4.3a |
|---|
| 481 | </h3> |
|---|
| 482 | |
|---|
| 483 | This is backward compatible with previous versions. |
|---|
| 484 | |
|---|
| 485 | <p> |
|---|
| 486 | <ul> |
|---|
| 487 | <li> |
|---|
| 488 | Improved the performance of <tt>ZZ_pX</tt> arithmetic when using |
|---|
| 489 | GMP. |
|---|
| 490 | The GMP version is also more space efficient |
|---|
| 491 | (the pre-computed tables are much smaller). |
|---|
| 492 | These improvements are most marked for very large <tt>p</tt> (several |
|---|
| 493 | thousand bits). |
|---|
| 494 | |
|---|
| 495 | <p> |
|---|
| 496 | The only thing unsatisfactory about this state of affairs is that |
|---|
| 497 | <i>vis a vis</i> the GMP version, the pure |
|---|
| 498 | LIP code is asymptotically slower by <i>more</i> than a constant factor, |
|---|
| 499 | and is is also less space efficient. |
|---|
| 500 | Perhaps I'll get around to rectifying this imbalance someday. |
|---|
| 501 | To do this, I need a sub-quadratic division with remainder routine for LIP. |
|---|
| 502 | At any rate, the differences only become seriously noticible when |
|---|
| 503 | <tt>p</tt> has more than a few thousand bits. |
|---|
| 504 | |
|---|
| 505 | <p> |
|---|
| 506 | |
|---|
| 507 | <li> |
|---|
| 508 | Some other small adjustments here and there. |
|---|
| 509 | |
|---|
| 510 | </ul> |
|---|
| 511 | |
|---|
| 512 | <p> <hr> <p> |
|---|
| 513 | <h3> |
|---|
| 514 | Changes between NTL 4.1a and 4.2a |
|---|
| 515 | </h3> |
|---|
| 516 | |
|---|
| 517 | This is backward compatible with previous versions. |
|---|
| 518 | |
|---|
| 519 | <p> |
|---|
| 520 | <ul> |
|---|
| 521 | <li> |
|---|
| 522 | Hacked the big integer code so that NTL uses GMP |
|---|
| 523 | (the GNU Multi-Precision library). |
|---|
| 524 | This is done in such a way as to get most of the benefits of GMP |
|---|
| 525 | with a reasonable amount of effort, and while maintaining complete backward |
|---|
| 526 | compatability and minimizing the risk of introducing bugs. |
|---|
| 527 | Some arithmetic operations |
|---|
| 528 | on some platforms may execute two to three times |
|---|
| 529 | faster if using GMP. <a href="tour-gmp.html">[more details]</a> |
|---|
| 530 | <li> |
|---|
| 531 | Simplified the installation procedure on Unix systems by |
|---|
| 532 | providing a simple configuration script so that setting |
|---|
| 533 | various configuration variables can be done without |
|---|
| 534 | editing the <tt>makefile</tt> and <tt>config.h</tt> file. |
|---|
| 535 | <a href="tour-unix.html">[more details]</a> |
|---|
| 536 | <li> |
|---|
| 537 | Added function <tt>GenGermainPrime</tt> |
|---|
| 538 | to efficiently generate random Germain primes, i.e., primes <i>p</i> |
|---|
| 539 | such that <i>2p+1</i> is also prime. <a href="ZZ.txt">[more details]</a> |
|---|
| 540 | <li> |
|---|
| 541 | Added a function <tt>random</tt> to generate random <tt>quad_floats</tt>. |
|---|
| 542 | <a href="quad_float.txt">[more details]</a> |
|---|
| 543 | <li> |
|---|
| 544 | Added an <tt>ifdef</tt> in <tt>tools.h</tt> that allows |
|---|
| 545 | one to suppress the declaration of <tt>min</tt> and <tt>max</tt> |
|---|
| 546 | functions in NTL client programs; |
|---|
| 547 | these were causing problems when writing 'Windows applications'. |
|---|
| 548 | <li> |
|---|
| 549 | Implemented a faster algorithm for initializing the |
|---|
| 550 | <tt>ZZ_p</tt> auxilliary data structures. |
|---|
| 551 | <li> |
|---|
| 552 | Polished up a few other minor things in the code and documentation. |
|---|
| 553 | </ul> |
|---|
| 554 | |
|---|
| 555 | <p> <hr> <p> |
|---|
| 556 | |
|---|
| 557 | <p> |
|---|
| 558 | <h3> |
|---|
| 559 | Changes between NTL 4.0a and 4.1a |
|---|
| 560 | </h3> |
|---|
| 561 | <p> |
|---|
| 562 | |
|---|
| 563 | This is backward compatible with previous versions. |
|---|
| 564 | |
|---|
| 565 | <p> |
|---|
| 566 | <ul> |
|---|
| 567 | <li> |
|---|
| 568 | Made some changes that should make NTL compile smoothly |
|---|
| 569 | using any variation of the <tt>C++</tt> language between traditional and |
|---|
| 570 | ISO Standard. |
|---|
| 571 | These changes do not affect the documented NTL interface or the |
|---|
| 572 | behaviour of NTL. |
|---|
| 573 | |
|---|
| 574 | <li> |
|---|
| 575 | Added a flag <tt>NTL_STD_CXX</tt> in the <tt>config.h</tt> file. |
|---|
| 576 | Setting this flag causes all of NTL to be "wrapped" in namespace <tt>NTL</tt>, |
|---|
| 577 | and that part of the standard library used by NTL is "wrapped" |
|---|
| 578 | in namespace <tt>std</tt>. |
|---|
| 579 | This should greatly help with the <i>namespace pollution</i> problem. |
|---|
| 580 | <a href="tour-stdcxx.html">Go here</a> for more details. |
|---|
| 581 | |
|---|
| 582 | </ul> |
|---|
| 583 | |
|---|
| 584 | |
|---|
| 585 | |
|---|
| 586 | <p> <hr> <p> |
|---|
| 587 | |
|---|
| 588 | <p> |
|---|
| 589 | <h3> |
|---|
| 590 | Changes between NTL 3.9b and 4.0a |
|---|
| 591 | </h3> |
|---|
| 592 | <p> |
|---|
| 593 | |
|---|
| 594 | This is backward compatible with previous version. |
|---|
| 595 | |
|---|
| 596 | <p> |
|---|
| 597 | <ul> |
|---|
| 598 | <li> |
|---|
| 599 | Attached the GNU General Public License to NTL. |
|---|
| 600 | |
|---|
| 601 | <li> |
|---|
| 602 | Fixed two bugs: |
|---|
| 603 | <ul> |
|---|
| 604 | <li> |
|---|
| 605 | one in <tt>ReconstructRational</tt> which resulted in a crash on some inputs; |
|---|
| 606 | <li> |
|---|
| 607 | one in <tt>exp(RR)</tt> (and by implication in <tt>pow(RR,RR)</tt>), |
|---|
| 608 | which led to wrong answers on 64-bit machines when computing <tt>exp(x)</tt> |
|---|
| 609 | for <tt>x > 2^53</tt>. |
|---|
| 610 | </ul> |
|---|
| 611 | |
|---|
| 612 | <li> |
|---|
| 613 | Increased some inconvenient limiting bounds, including a restriction on the |
|---|
| 614 | FFT. |
|---|
| 615 | |
|---|
| 616 | </ul> |
|---|
| 617 | |
|---|
| 618 | |
|---|
| 619 | <p> <hr> <p> |
|---|
| 620 | |
|---|
| 621 | <p> |
|---|
| 622 | <h3> |
|---|
| 623 | Changes between NTL 3.9a and 3.9b |
|---|
| 624 | </h3> |
|---|
| 625 | <p> |
|---|
| 626 | |
|---|
| 627 | This is a minor revision of 3.9a. |
|---|
| 628 | |
|---|
| 629 | <ul> |
|---|
| 630 | <li> |
|---|
| 631 | Improved time and space efficiency of the HNF routine |
|---|
| 632 | (see <a href="HNF.txt"><tt>HNF.txt</tt></a>). |
|---|
| 633 | The old version was based on the description in Henri Cohen's book, |
|---|
| 634 | which was not really properly optimized. |
|---|
| 635 | </ul> |
|---|
| 636 | |
|---|
| 637 | |
|---|
| 638 | |
|---|
| 639 | <p> <hr> <p> |
|---|
| 640 | |
|---|
| 641 | <p> |
|---|
| 642 | <h3> |
|---|
| 643 | Changes between NTL 3.8b and 3.9a |
|---|
| 644 | </h3> |
|---|
| 645 | <p> |
|---|
| 646 | |
|---|
| 647 | This is backward compatible with previous versions. |
|---|
| 648 | |
|---|
| 649 | <ul> |
|---|
| 650 | <li> |
|---|
| 651 | Modified the installation script somewhat, adding |
|---|
| 652 | a <i>configuration wizard</i> that sets the flags in |
|---|
| 653 | <tt>config.h</tt> "automagically". |
|---|
| 654 | This works for the <a href="tour-unix.html">Unix version</a> only. |
|---|
| 655 | |
|---|
| 656 | <li> |
|---|
| 657 | Improved the <tt>xdouble</tt> input/output and ascii to <tt>xdouble</tt> |
|---|
| 658 | conversion. |
|---|
| 659 | The old version could be a bit flaky when reading/writing |
|---|
| 660 | very large numbers. |
|---|
| 661 | The new I/O routines also attain better accuracy. |
|---|
| 662 | |
|---|
| 663 | <li> |
|---|
| 664 | Improved conversion routines between <tt>xdouble</tt> |
|---|
| 665 | and <tt>ZZ</tt>/<tt>RR</tt>. |
|---|
| 666 | |
|---|
| 667 | <li> |
|---|
| 668 | Improved the <tt>RR</tt> output routine. |
|---|
| 669 | The new version should be more accurate and also |
|---|
| 670 | completely platform independent. |
|---|
| 671 | |
|---|
| 672 | <li> |
|---|
| 673 | Added the following routines to the <tt>RR</tt> package: |
|---|
| 674 | <pre> |
|---|
| 675 | {Trunc,Floor,Ceil,Round}ToZZ, round |
|---|
| 676 | RoundToPrecision, MakeRR |
|---|
| 677 | random |
|---|
| 678 | </pre> |
|---|
| 679 | See <a href="RR.txt"><tt>RR.txt</tt></a> for details. |
|---|
| 680 | |
|---|
| 681 | <li> |
|---|
| 682 | Improved the accuracy of <tt>quad_float</tt> input/output, |
|---|
| 683 | and the accuracy of conversion between <tt>quad_float</tt> and <tt>RR</tt>. |
|---|
| 684 | |
|---|
| 685 | <li> |
|---|
| 686 | Made the timing function somewhat more robust. |
|---|
| 687 | |
|---|
| 688 | <li> |
|---|
| 689 | Hacked the Unix installation script so that it works |
|---|
| 690 | more smoothly with Cygnus tools under Windows. |
|---|
| 691 | |
|---|
| 692 | <li> |
|---|
| 693 | Fixed a few other, small problems. |
|---|
| 694 | </ul> |
|---|
| 695 | |
|---|
| 696 | <p> <hr> <p> |
|---|
| 697 | |
|---|
| 698 | <p> |
|---|
| 699 | <h3> |
|---|
| 700 | Changes between NTL 3.8a and 3.8b |
|---|
| 701 | </h3> |
|---|
| 702 | <p> |
|---|
| 703 | |
|---|
| 704 | This is a minor revision of 3.8a. |
|---|
| 705 | |
|---|
| 706 | <ul> |
|---|
| 707 | <li> |
|---|
| 708 | Fixed a bug, a memory leak in routine <tt>gauss</tt> for <tt>mat_ZZ_pE</tt> |
|---|
| 709 | and <tt>mat_zz_pE</tt>. |
|---|
| 710 | <li> |
|---|
| 711 | Fixed a minor problem in <tt>config.h</tt>. |
|---|
| 712 | <li> |
|---|
| 713 | Tightened up some size checks, so that now some nice "size invariants" |
|---|
| 714 | are guaranteed, e.g., for a <tt>ZZ</tt> <tt>n</tt>, |
|---|
| 715 | <pre> |
|---|
| 716 | NumBits(NumBits(n)) <= NTL_BITS_PER_LONG-4 |
|---|
| 717 | </pre> |
|---|
| 718 | Similarly for the type <tt>GF2X</tt>. |
|---|
| 719 | Of course, on most platforms, one will run out of memory before |
|---|
| 720 | these bounds are exceeded, but they are nevertheless convenient. |
|---|
| 721 | </ul> |
|---|
| 722 | |
|---|
| 723 | |
|---|
| 724 | <p> <hr> <p> |
|---|
| 725 | |
|---|
| 726 | <p> |
|---|
| 727 | <h3> |
|---|
| 728 | Changes between NTL 3.7a and 3.8a |
|---|
| 729 | </h3> |
|---|
| 730 | <p> |
|---|
| 731 | |
|---|
| 732 | This is backward compatible with previous versions. |
|---|
| 733 | |
|---|
| 734 | |
|---|
| 735 | <ul> |
|---|
| 736 | <li> |
|---|
| 737 | Added conversion routines from <tt>unsigned</tt> <tt>int</tt> |
|---|
| 738 | and <tt>unsigned</tt> <tt>long</tt> to |
|---|
| 739 | <tt>ZZ</tt>, <tt>RR</tt>, <tt>xdouble</tt>, and <tt>quad_float</tt>. |
|---|
| 740 | |
|---|
| 741 | <li> |
|---|
| 742 | Added routines <tt>GF2XFromBytes</tt> and <tt>BytesFromGF2X</tt> |
|---|
| 743 | for conversion between byte vectors and polynomials over <tt>GF(2)</tt>, |
|---|
| 744 | along with routines <tt>NumBits</tt> and <tt>NumBytes</tt> |
|---|
| 745 | for such polynomials. |
|---|
| 746 | See <a href="GF2X.txt"><tt>GF2X.txt</tt></a> for details. |
|---|
| 747 | |
|---|
| 748 | <li> |
|---|
| 749 | Added a hack in the <tt>ZZX</tt> factorizer |
|---|
| 750 | to exploit polynomials of the form <tt>g(x^k)</tt>. |
|---|
| 751 | This can be disabled by setting the variable <tt>ZZXFac_PowerHack</tt> |
|---|
| 752 | to zero. |
|---|
| 753 | See <a href="ZZXFactoring.txt"><tt>ZZXFactoring.txt</tt></a> |
|---|
| 754 | for details. |
|---|
| 755 | |
|---|
| 756 | <li> |
|---|
| 757 | Improved the hensel system solver <tt>solve1</tt>. |
|---|
| 758 | See <a href="mat_ZZ.txt"><tt>mat_ZZ.txt</tt></a> for details. |
|---|
| 759 | |
|---|
| 760 | <li> |
|---|
| 761 | Changed documentation for <tt>RationalReconstruction</tt> |
|---|
| 762 | to reflect the Wang, Guy, Davenport bounds. |
|---|
| 763 | See <a href="ZZ.txt"><tt>ZZ.txt</tt></a> for details. |
|---|
| 764 | |
|---|
| 765 | <li> |
|---|
| 766 | Improved the routine <tt>GenPrime</tt> a bit. |
|---|
| 767 | |
|---|
| 768 | <li> |
|---|
| 769 | Some other small tweaks here and there. |
|---|
| 770 | No real bug fixes. |
|---|
| 771 | |
|---|
| 772 | <li> |
|---|
| 773 | Polished the documentation a bit, adding more examples. |
|---|
| 774 | |
|---|
| 775 | </ul> |
|---|
| 776 | |
|---|
| 777 | <p> <hr> <p> |
|---|
| 778 | |
|---|
| 779 | <p> |
|---|
| 780 | <h3> |
|---|
| 781 | Changes between NTL 3.6b and 3.7a |
|---|
| 782 | </h3> |
|---|
| 783 | <p> |
|---|
| 784 | |
|---|
| 785 | This is backward compatible with previous versions. |
|---|
| 786 | |
|---|
| 787 | <ul> |
|---|
| 788 | <li> |
|---|
| 789 | Added a "rational reconstruction" routine. |
|---|
| 790 | See the routine <tt>ReconstructRational</tt> in <a href="ZZ.txt">ZZ.txt</a>. |
|---|
| 791 | <li> |
|---|
| 792 | Added another routine for solving linear systems over <tt>ZZ</tt> |
|---|
| 793 | that is based on Hensel lifting, rather than Chinese Remaindering. |
|---|
| 794 | It can be significantly faster in some cases. |
|---|
| 795 | See the routine <tt>solve1</tt> in <a href="mat_ZZ.txt">mat_ZZ.txt</a>). |
|---|
| 796 | <li> |
|---|
| 797 | Some performace tuning, especially CRT and polynomial interpolation code. |
|---|
| 798 | <li> |
|---|
| 799 | Various documentation corrections. |
|---|
| 800 | <li> |
|---|
| 801 | Added more "overflow checks" here and there to ensure programs crash gracefully |
|---|
| 802 | when certain things get too big. |
|---|
| 803 | <li> |
|---|
| 804 | Fixed a "benign" bug (i.e., it would never get triggered on any of today's |
|---|
| 805 | machines). |
|---|
| 806 | <li> |
|---|
| 807 | Removed references to <tt><malloc.h></tt>, which were unnecessary, |
|---|
| 808 | non-standard, and caused problems on some platforms. |
|---|
| 809 | </ul> |
|---|
| 810 | |
|---|
| 811 | <p> |
|---|
| 812 | <hr> |
|---|
| 813 | |
|---|
| 814 | <p> |
|---|
| 815 | <h3> |
|---|
| 816 | Changes between NTL 3.6a and 3.6b |
|---|
| 817 | </h3> |
|---|
| 818 | <p> |
|---|
| 819 | |
|---|
| 820 | Bug fixes. |
|---|
| 821 | |
|---|
| 822 | <p> |
|---|
| 823 | <hr> |
|---|
| 824 | |
|---|
| 825 | <p> |
|---|
| 826 | <h3> |
|---|
| 827 | Changes between NTL 3.5a and 3.6a |
|---|
| 828 | </h3> |
|---|
| 829 | <p> |
|---|
| 830 | |
|---|
| 831 | This version is backward compatible with 3.5a. |
|---|
| 832 | |
|---|
| 833 | <p> |
|---|
| 834 | |
|---|
| 835 | <ul> |
|---|
| 836 | |
|---|
| 837 | <li> |
|---|
| 838 | A few small bug fixes and performance enhancements. |
|---|
| 839 | |
|---|
| 840 | <li> |
|---|
| 841 | Changes to the <tt>ZZX</tt> factoring routines that in some |
|---|
| 842 | cases yield dramatic performance improvements |
|---|
| 843 | (<a href="tour-time.html">more details</a>). |
|---|
| 844 | |
|---|
| 845 | </ul> |
|---|
| 846 | |
|---|
| 847 | <p> |
|---|
| 848 | <hr> |
|---|
| 849 | |
|---|
| 850 | |
|---|
| 851 | <p> |
|---|
| 852 | <h3> |
|---|
| 853 | Changes between NTL 3.1b and 3.5a |
|---|
| 854 | </h3> |
|---|
| 855 | <p> |
|---|
| 856 | |
|---|
| 857 | <b>Please note.</b> This version is <b>NOT</b> completely backward compatible. |
|---|
| 858 | |
|---|
| 859 | <p> |
|---|
| 860 | |
|---|
| 861 | Summary of changes: |
|---|
| 862 | |
|---|
| 863 | <ul> |
|---|
| 864 | |
|---|
| 865 | <li> |
|---|
| 866 | Improved performance of the "all integer" LLL routine. |
|---|
| 867 | |
|---|
| 868 | <li> |
|---|
| 869 | Put in a better pseudo-random number generator, |
|---|
| 870 | and added ZZ/byte array conversions. |
|---|
| 871 | |
|---|
| 872 | <li> |
|---|
| 873 | Improved performance of primality test, and added a |
|---|
| 874 | more convenient routine <tt>GenPrime</tt>. |
|---|
| 875 | |
|---|
| 876 | <li> |
|---|
| 877 | Overloaded NTL's vector placement "new" operator in a different |
|---|
| 878 | way to avoid conflicts with standard <tt>C++</tt> library. |
|---|
| 879 | |
|---|
| 880 | <li> |
|---|
| 881 | Renamed many macros. |
|---|
| 882 | |
|---|
| 883 | <li> |
|---|
| 884 | Renamed header files. |
|---|
| 885 | |
|---|
| 886 | <li> |
|---|
| 887 | Made some changes to the packaging |
|---|
| 888 | the installation procedure. |
|---|
| 889 | |
|---|
| 890 | </ul> |
|---|
| 891 | |
|---|
| 892 | <p> |
|---|
| 893 | <b>Renamed Macros.</b> |
|---|
| 894 | I renamed many macros defined in NTL header files. |
|---|
| 895 | |
|---|
| 896 | <p> |
|---|
| 897 | The reason is that I want to minimize namespace pollution. |
|---|
| 898 | Someday, NTL will be wrapped in a namespace, and when that happens |
|---|
| 899 | the only remaining namespace pollution problems will be caused by macros. |
|---|
| 900 | Eliminating all macros from NTL is not feasible. |
|---|
| 901 | Instead, all NTL defined macros now begin with the prefix "NTL_", |
|---|
| 902 | which reduces the namespace pollution to an ecceptable level. |
|---|
| 903 | You will probably not be affected by this, unless you |
|---|
| 904 | do some low level hacking using a macro like <tt>ZZ_NBITS</tt> |
|---|
| 905 | (now called <tt>NTL_NBITS</tt>), or unless you create your |
|---|
| 906 | own NTL vectors using a macro like <tt>ntl_vector_decl</tt> |
|---|
| 907 | (now called <tt>NTL_vector_decl</tt>). |
|---|
| 908 | |
|---|
| 909 | <p> |
|---|
| 910 | For a complete list of affected names, see <a href="names.txt">names.txt</a>. |
|---|
| 911 | |
|---|
| 912 | <p> |
|---|
| 913 | Adapting to this name change should be painless, as there is a |
|---|
| 914 | program to translate source files from the old naming convention to the new. |
|---|
| 915 | The file "newnames.c", |
|---|
| 916 | can be compiled as either a <tt>C</tt> or <tt>C++</tt> |
|---|
| 917 | program. |
|---|
| 918 | The program is a "filter" that copies its input to its output, |
|---|
| 919 | replacing all the old macro names by the new macro names. |
|---|
| 920 | <p> |
|---|
| 921 | In the WinNTL distribibution, "newnames.c" is called |
|---|
| 922 | "newnames.cpp" and is located in the directory |
|---|
| 923 | "newnames". |
|---|
| 924 | |
|---|
| 925 | |
|---|
| 926 | <p> |
|---|
| 927 | <b>Renamed header files.</b> |
|---|
| 928 | The names of header files themeselves pollute another (extra-linguitsic) namespace. |
|---|
| 929 | To alleviate this problem, the header files have been renamed. |
|---|
| 930 | Instead of |
|---|
| 931 | <pre> |
|---|
| 932 | #include "foo.h" |
|---|
| 933 | </pre> |
|---|
| 934 | one now should write |
|---|
| 935 | <pre> |
|---|
| 936 | #include <NTL/foo.h> |
|---|
| 937 | </pre> |
|---|
| 938 | The only exceptions are the old header files "ntl_vector.h", |
|---|
| 939 | "ntl_matrix.h", and "ntl_pair.h", which are now called |
|---|
| 940 | <tt><NTL/vector.h></tt>, <tt><NTL/matrix.h></tt>, and |
|---|
| 941 | <tt><NTL/pair.h></tt>. |
|---|
| 942 | |
|---|
| 943 | <p> |
|---|
| 944 | <b>Installation procedure.</b> |
|---|
| 945 | Now all |
|---|
| 946 | NTL flags like NTL_LONG_LONG, NTL_AVOID_FLOAT, etc., can now be set |
|---|
| 947 | by editing the special file "include/NTL/config.h". |
|---|
| 948 | See details in that file. |
|---|
| 949 | The reason for this change is that this allows all of these settings |
|---|
| 950 | to be made when NTL is configured and built. |
|---|
| 951 | Clients of NTL will then automatically use consistent settings. |
|---|
| 952 | One should not set these flags on the compiler command line as previously. |
|---|
| 953 | |
|---|
| 954 | |
|---|
| 955 | <p> |
|---|
| 956 | Pentium/Linux people should no longer have to worry |
|---|
| 957 | about the NTL_X86_FIX flag. NTL now psychically deduces |
|---|
| 958 | the "right thing to do", although if its psychic abilities fail, |
|---|
| 959 | you can override it with flags in "include/NTL/config.h". |
|---|
| 960 | |
|---|
| 961 | <p> |
|---|
| 962 | The "packaging" in the Unix distribution is slightly |
|---|
| 963 | different, but hopefully nicer. |
|---|
| 964 | Among other things, the tar file now unpacks into a sub-directory of the current directory. |
|---|
| 965 | See <a href="tour-unix.html">the unix installation section</a> |
|---|
| 966 | for more details. |
|---|
| 967 | The Windows zip file now also |
|---|
| 968 | unpacks into sub-directory. |
|---|
| 969 | |
|---|
| 970 | |
|---|
| 971 | <p> |
|---|
| 972 | <b>My apologies.</b> |
|---|
| 973 | Although these changes are minor, they will cause some NTL |
|---|
| 974 | users some inconvenience. |
|---|
| 975 | I apologize for this. |
|---|
| 976 | I really, really hope there are no more changes like this |
|---|
| 977 | (see my <a href="tour-roadmap.html">roadmap</a> of NTL's future). |
|---|
| 978 | |
|---|
| 979 | <p> |
|---|
| 980 | <hr> |
|---|
| 981 | |
|---|
| 982 | |
|---|
| 983 | <p> |
|---|
| 984 | <h3> |
|---|
| 985 | Changes between NTL 3.1a and 3.1b |
|---|
| 986 | </h3> |
|---|
| 987 | <p> |
|---|
| 988 | |
|---|
| 989 | Defined functions <tt>div(GF2X,GF2X,GF2)</tt> and <tt>div(GF2X,GF2X,long)</tt>, |
|---|
| 990 | which had not been defined in earlier versions. |
|---|
| 991 | Affected file: <tt>GF2X.c</tt>. |
|---|
| 992 | Most programs never use this, and most linkers do not complain |
|---|
| 993 | if these are missing (but some do). |
|---|
| 994 | |
|---|
| 995 | <p> |
|---|
| 996 | <hr> |
|---|
| 997 | |
|---|
| 998 | <p> |
|---|
| 999 | <h3> |
|---|
| 1000 | Changes between NTL 3.0f and 3.1a |
|---|
| 1001 | </h3> |
|---|
| 1002 | <p> |
|---|
| 1003 | |
|---|
| 1004 | This version is backward compatible with previous versions. |
|---|
| 1005 | |
|---|
| 1006 | <p> |
|---|
| 1007 | |
|---|
| 1008 | <ul> |
|---|
| 1009 | <li> |
|---|
| 1010 | Added floating point LLL routines based on Givens rotations, |
|---|
| 1011 | instead of classical Gramm-Schmidt orthogonalization. |
|---|
| 1012 | This is a more stable, but somewhat slower, method. |
|---|
| 1013 | See <a href="LLL.txt">LLL.txt</a> for details. |
|---|
| 1014 | |
|---|
| 1015 | <li> |
|---|
| 1016 | Added support for irreducible trinomials and pentanomials |
|---|
| 1017 | over GF(2). The <tt>GF2XModulus</tt> routines, |
|---|
| 1018 | and by extension, the <tt>GF2E</tt> routines, |
|---|
| 1019 | now exploit moduli of this special form. |
|---|
| 1020 | The new routine <tt>BuildSparseIrred</tt> in <tt>GF2XFactoring</tt> |
|---|
| 1021 | builds irreducibles of this form. |
|---|
| 1022 | |
|---|
| 1023 | <li> |
|---|
| 1024 | Also implemented a faster modular inversion routine |
|---|
| 1025 | for <tt>GF2X</tt>, and improved the performance of <tt>ZZ_pX</tt> |
|---|
| 1026 | multiplication for small degree polynomials. |
|---|
| 1027 | </ul> |
|---|
| 1028 | |
|---|
| 1029 | <p> |
|---|
| 1030 | <hr> |
|---|
| 1031 | |
|---|
| 1032 | <p> |
|---|
| 1033 | <h3> |
|---|
| 1034 | Changes between NTL 3.0e and 3.0f |
|---|
| 1035 | </h3> |
|---|
| 1036 | <p> |
|---|
| 1037 | |
|---|
| 1038 | <ul> |
|---|
| 1039 | <li> |
|---|
| 1040 | Fixed a bug (another one) affecting routines |
|---|
| 1041 | <pre> |
|---|
| 1042 | RandomBits, RandomBits_ZZ |
|---|
| 1043 | </pre> |
|---|
| 1044 | in module <tt>ZZ</tt>. |
|---|
| 1045 | Affected source file: <tt>lip.c</tt>. |
|---|
| 1046 | |
|---|
| 1047 | <li> |
|---|
| 1048 | Bug fix and performance tweak in <tt>ZZX</tt> factorizer. |
|---|
| 1049 | Affected source file: <tt>ZZXFactoring.c</tt>. |
|---|
| 1050 | |
|---|
| 1051 | </ul> |
|---|
| 1052 | |
|---|
| 1053 | <p> |
|---|
| 1054 | <hr> |
|---|
| 1055 | |
|---|
| 1056 | <p> |
|---|
| 1057 | <h3> |
|---|
| 1058 | Changes between NTL 3.0 and 3.0e |
|---|
| 1059 | </h3> |
|---|
| 1060 | <p> |
|---|
| 1061 | |
|---|
| 1062 | <ul> |
|---|
| 1063 | <li> |
|---|
| 1064 | Fixed a bug affecting routines |
|---|
| 1065 | <pre> |
|---|
| 1066 | RandomBits, RandomBits_ZZ, RandomBits_long |
|---|
| 1067 | </pre> |
|---|
| 1068 | in module <tt>ZZ</tt>. |
|---|
| 1069 | The only source files that are affected and require re-compilation are |
|---|
| 1070 | <pre> |
|---|
| 1071 | ZZ.c, lip.c |
|---|
| 1072 | </pre> |
|---|
| 1073 | |
|---|
| 1074 | <li> |
|---|
| 1075 | Note about names: |
|---|
| 1076 | 3.0a-c were "pre-releases", which makes the "first release" 3.0d, |
|---|
| 1077 | and hence this bug fix 3.0e. |
|---|
| 1078 | |
|---|
| 1079 | </ul> |
|---|
| 1080 | |
|---|
| 1081 | <p> |
|---|
| 1082 | <hr> |
|---|
| 1083 | |
|---|
| 1084 | |
|---|
| 1085 | <p> |
|---|
| 1086 | |
|---|
| 1087 | <h3> |
|---|
| 1088 | Changes between NTL 2.0 and 3.0 |
|---|
| 1089 | </h3> |
|---|
| 1090 | <p> |
|---|
| 1091 | |
|---|
| 1092 | |
|---|
| 1093 | <ul> |
|---|
| 1094 | |
|---|
| 1095 | <li> |
|---|
| 1096 | Added functionality: |
|---|
| 1097 | <p> |
|---|
| 1098 | |
|---|
| 1099 | <ul> |
|---|
| 1100 | |
|---|
| 1101 | <li> |
|---|
| 1102 | Added classes vec_GF2 and mat_GF2 for fast linear algebra over GF(2). |
|---|
| 1103 | |
|---|
| 1104 | <li> |
|---|
| 1105 | Added classes ZZ_pE, ZZ_pEX, zz_pE, zz_pEX, supporting polynomial |
|---|
| 1106 | arithmetic over extension rings/fields over prime fields. |
|---|
| 1107 | |
|---|
| 1108 | <li> |
|---|
| 1109 | Added John Abbott's pruning heuristic to the ZZX factoring routine. |
|---|
| 1110 | |
|---|
| 1111 | <li> |
|---|
| 1112 | Speeded up multiplication in zz_pX for small p (this also helps |
|---|
| 1113 | the ZZX factoring routine). |
|---|
| 1114 | |
|---|
| 1115 | <li> |
|---|
| 1116 | Added some some transcendental functions (e.g., exp, log, pi) to RR. |
|---|
| 1117 | |
|---|
| 1118 | <li> |
|---|
| 1119 | Added verbose mode and pruning to the XD and RR variants of LLL. |
|---|
| 1120 | |
|---|
| 1121 | </ul> |
|---|
| 1122 | <p> |
|---|
| 1123 | |
|---|
| 1124 | <li> |
|---|
| 1125 | Improved programming interface: |
|---|
| 1126 | with this version, I've taken an the opportunity to |
|---|
| 1127 | give the programming interface a "professional facelift". |
|---|
| 1128 | In previous releases, I've tried to maintain backward compatability |
|---|
| 1129 | as much as possible, but to make the badly needed improvements |
|---|
| 1130 | to the interface that I've made with this release, this was not |
|---|
| 1131 | possible. |
|---|
| 1132 | <p> |
|---|
| 1133 | NTL 3.0 is not backward compatable with NTL 2.0. |
|---|
| 1134 | <p> |
|---|
| 1135 | I apologize to NTL users for this, but it is a bit of painful |
|---|
| 1136 | medicine that should only be necessary to take just this one time |
|---|
| 1137 | (but then as a <tt>C++</tt> programmer, you must already |
|---|
| 1138 | be used to suffering ;-). |
|---|
| 1139 | Just about all of the incompatabilities are detectable by the compiler. |
|---|
| 1140 | See below for a detailed list of the changes and |
|---|
| 1141 | some tips on making the transition. |
|---|
| 1142 | <p> |
|---|
| 1143 | The new interface is much more enjoyable to work with, |
|---|
| 1144 | and I don't foresee any changes to the interace in the future. |
|---|
| 1145 | Here is a broad overview of the changes: |
|---|
| 1146 | <p> |
|---|
| 1147 | |
|---|
| 1148 | <ul> |
|---|
| 1149 | <li> |
|---|
| 1150 | Added functional/operator notation consistently throughout NTL, |
|---|
| 1151 | making it possible to write much more concise and readable code. |
|---|
| 1152 | <li> |
|---|
| 1153 | Got rid of automatic type conversions: these cause just too |
|---|
| 1154 | many problems. But I've overloaded all of the basic arithmetic |
|---|
| 1155 | operators and procedures so as to emulate a natural kind |
|---|
| 1156 | of "type promotion" logic. With these promotions, along with |
|---|
| 1157 | a full compliment of conversion functions, one hardly misses |
|---|
| 1158 | the automatic conversions. |
|---|
| 1159 | <li> |
|---|
| 1160 | Got rid of the macros |
|---|
| 1161 | <pre> |
|---|
| 1162 | vector(T), matrix(T), pair(T), |
|---|
| 1163 | </pre> |
|---|
| 1164 | which were causing too many name space problems. |
|---|
| 1165 | |
|---|
| 1166 | <li> |
|---|
| 1167 | Made assignment operators have the "correct" return type. |
|---|
| 1168 | <li> |
|---|
| 1169 | Introduced a more powerful and flexible mechanism for modulus changing. |
|---|
| 1170 | <li> |
|---|
| 1171 | Cleaned up numerous other minor problems. |
|---|
| 1172 | </ul> |
|---|
| 1173 | |
|---|
| 1174 | </ul> |
|---|
| 1175 | |
|---|
| 1176 | <p> |
|---|
| 1177 | <h4> |
|---|
| 1178 | Compatibility |
|---|
| 1179 | </h4> |
|---|
| 1180 | <p> |
|---|
| 1181 | |
|---|
| 1182 | Here is a detailed list of the changes to the programming |
|---|
| 1183 | interface. |
|---|
| 1184 | <p> |
|---|
| 1185 | |
|---|
| 1186 | |
|---|
| 1187 | <ul> |
|---|
| 1188 | |
|---|
| 1189 | <li> |
|---|
| 1190 | The names of the classes |
|---|
| 1191 | <pre> |
|---|
| 1192 | BB, BB_p, BB_pX |
|---|
| 1193 | </pre> |
|---|
| 1194 | have been changed to |
|---|
| 1195 | <pre> |
|---|
| 1196 | GF2X, GF2E, GF2EX |
|---|
| 1197 | </pre> |
|---|
| 1198 | |
|---|
| 1199 | <li> |
|---|
| 1200 | There is also a class <tt>GF2</tt> to represent GF(2). |
|---|
| 1201 | Many of the functions relating to <tt>BB, BB_p, BB_pX</tt> |
|---|
| 1202 | had argument and return-value types of type <tt>long</tt> |
|---|
| 1203 | that are now of the more appropriate type <tt>GF2</tt>. |
|---|
| 1204 | This change was needed so that the interface would be consistent |
|---|
| 1205 | with that of the new classes |
|---|
| 1206 | <pre> |
|---|
| 1207 | ZZ_pE, ZZ_pEX, zz_pE, zz_pEX. |
|---|
| 1208 | </pre> |
|---|
| 1209 | |
|---|
| 1210 | <li> |
|---|
| 1211 | The explicit conversion operator from <tt>GF2X</tt> |
|---|
| 1212 | (the new <tt>BB</tt>) to <tt>GF2EX</tt> (the new <tt>BB_pX</tt>) |
|---|
| 1213 | has different semantics: it now performs a coefficient lift, |
|---|
| 1214 | instead of creating a constant polynomial. |
|---|
| 1215 | |
|---|
| 1216 | <li> |
|---|
| 1217 | The conversion operator "<tt><<</tt>" has been retired. |
|---|
| 1218 | Now instead of |
|---|
| 1219 | <pre> |
|---|
| 1220 | x << a; |
|---|
| 1221 | </pre> |
|---|
| 1222 | one writes |
|---|
| 1223 | <pre> |
|---|
| 1224 | conv(x, a); |
|---|
| 1225 | </pre> |
|---|
| 1226 | <p> |
|---|
| 1227 | Operator "<tt><<</tt>" is now used for shift operations. |
|---|
| 1228 | <li> |
|---|
| 1229 | Every conversion routine now has a corresponding functional version |
|---|
| 1230 | which has the name <tt>to_T</tt>, where <tt>T</tt> is the result type. |
|---|
| 1231 | These new names replace old names that were less consistent. |
|---|
| 1232 | So instead of |
|---|
| 1233 | <pre> |
|---|
| 1234 | x = Long(a); |
|---|
| 1235 | </pre> |
|---|
| 1236 | one writes |
|---|
| 1237 | <pre> |
|---|
| 1238 | x = to_long(a); |
|---|
| 1239 | </pre> |
|---|
| 1240 | |
|---|
| 1241 | |
|---|
| 1242 | <li> |
|---|
| 1243 | The names of the routines |
|---|
| 1244 | <pre> |
|---|
| 1245 | ZZ_pInit, zz_pInit, zz_pFFTInit, GF2EInit |
|---|
| 1246 | </pre> |
|---|
| 1247 | have been changed to |
|---|
| 1248 | <pre> |
|---|
| 1249 | zz_p::init, zz_p::init, zz_p::FFTInit, GF2E::init |
|---|
| 1250 | </pre> |
|---|
| 1251 | |
|---|
| 1252 | <li> |
|---|
| 1253 | The names of the routines |
|---|
| 1254 | <pre> |
|---|
| 1255 | and, or, xor |
|---|
| 1256 | </pre> |
|---|
| 1257 | for class <tt>ZZ</tt> have |
|---|
| 1258 | changed to |
|---|
| 1259 | <pre> |
|---|
| 1260 | bit_and, bit_or, bit_xor, |
|---|
| 1261 | </pre> |
|---|
| 1262 | because the new <tt>C++</tt> |
|---|
| 1263 | standard defines these as reserved words. |
|---|
| 1264 | |
|---|
| 1265 | <li> |
|---|
| 1266 | The function <tt>LowBits</tt> for <tt>ZZ</tt> is now called <tt>trunc</tt>. |
|---|
| 1267 | |
|---|
| 1268 | <li> |
|---|
| 1269 | Polynomial inversion mod <tt>X^n</tt> has changed from <tt>inv</tt> |
|---|
| 1270 | to <tt>InvTrunc</tt>. |
|---|
| 1271 | |
|---|
| 1272 | <li> |
|---|
| 1273 | Modular trace, norm, minimum polynomial and characteristic |
|---|
| 1274 | polynomial have changed from |
|---|
| 1275 | <pre> |
|---|
| 1276 | trace, norm, MinPoly, IrredPoly, CharPoly |
|---|
| 1277 | </pre> |
|---|
| 1278 | to |
|---|
| 1279 | <pre> |
|---|
| 1280 | TraceMod, NormMod, MinPolyMod, IrredPolyMod, CharPolyMod |
|---|
| 1281 | </pre> |
|---|
| 1282 | |
|---|
| 1283 | |
|---|
| 1284 | <li> |
|---|
| 1285 | For the class <tt>ZZX</tt>, the functions |
|---|
| 1286 | <pre> |
|---|
| 1287 | DivRem, div, rem, /, %, /=, %= |
|---|
| 1288 | </pre> |
|---|
| 1289 | have new semantics when dividing by non-monic polynomials. |
|---|
| 1290 | The old semantics are provided by new routines |
|---|
| 1291 | <pre> |
|---|
| 1292 | PseudoDivRem, PseudoDiv, PseudoRem. |
|---|
| 1293 | </pre> |
|---|
| 1294 | |
|---|
| 1295 | <li> |
|---|
| 1296 | The <tt>UpdateMap</tt> routines have slightly different semantics: |
|---|
| 1297 | in versions < 3.0, the output always had length n; |
|---|
| 1298 | now high-order zeroes are stripped. |
|---|
| 1299 | |
|---|
| 1300 | <li> |
|---|
| 1301 | The classes <tt>ZZ_pBak</tt>, <tt>zz_pBak,</tt> etc., |
|---|
| 1302 | have just slightly different semantics; I can't imagine |
|---|
| 1303 | any reasonable program detecting a difference. |
|---|
| 1304 | |
|---|
| 1305 | <li> |
|---|
| 1306 | The assignment operator and copy constructor for the class <tt>RR</tt> |
|---|
| 1307 | have different semantics: they now produce exact copies, instead |
|---|
| 1308 | of rounding to current precision. |
|---|
| 1309 | |
|---|
| 1310 | <li> |
|---|
| 1311 | All of the NTL compiler flags now start with <tt>NTL_</tt> |
|---|
| 1312 | to avoid name space problems. |
|---|
| 1313 | |
|---|
| 1314 | <li> |
|---|
| 1315 | All of the files "zz_p.h", vec_zz_p.h", etc., have been eliminated. |
|---|
| 1316 | Use instead the names "lzz_p.h", "vec_lzz_p.h", etc. |
|---|
| 1317 | |
|---|
| 1318 | </ul> |
|---|
| 1319 | |
|---|
| 1320 | <p> |
|---|
| 1321 | <h4> |
|---|
| 1322 | Tips on making the transition |
|---|
| 1323 | </h4> |
|---|
| 1324 | <p> |
|---|
| 1325 | |
|---|
| 1326 | <ul> |
|---|
| 1327 | |
|---|
| 1328 | <li> |
|---|
| 1329 | Apply this <a href="sedscript.txt">sed script</a> to make |
|---|
| 1330 | most of the necessary syntactic changes. |
|---|
| 1331 | |
|---|
| 1332 | <li> |
|---|
| 1333 | Re-compile old NTL programs with the flag |
|---|
| 1334 | <pre> |
|---|
| 1335 | -DNTL_TRANSITION |
|---|
| 1336 | </pre> |
|---|
| 1337 | See <a href="flags.txt">flags.txt</a> for details on how |
|---|
| 1338 | this will help your compiler detect remaining incompatabilities. |
|---|
| 1339 | In particular, any uses of operator <tt><<</tt> |
|---|
| 1340 | in its old role as a conversion operator will cause the compiler |
|---|
| 1341 | to raise an error. |
|---|
| 1342 | You can then convert all of these to the new notation. |
|---|
| 1343 | |
|---|
| 1344 | </ul> |
|---|
| 1345 | |
|---|
| 1346 | |
|---|
| 1347 | <p> |
|---|
| 1348 | <hr> |
|---|
| 1349 | |
|---|
| 1350 | |
|---|
| 1351 | <p> |
|---|
| 1352 | <h3> |
|---|
| 1353 | Changes between NTL 1.7 and 2.0 |
|---|
| 1354 | </h3> |
|---|
| 1355 | <p> |
|---|
| 1356 | |
|---|
| 1357 | <ul> |
|---|
| 1358 | <li> |
|---|
| 1359 | Implementation of classes BB (polynomials over GF(2)) |
|---|
| 1360 | and BB_pX (polynomials over GF(2^n)). |
|---|
| 1361 | |
|---|
| 1362 | <li> |
|---|
| 1363 | A more consistent and natural interface, including arithmetic operators |
|---|
| 1364 | and a disciplined use of automatic conversion. |
|---|
| 1365 | So now one can write |
|---|
| 1366 | <pre> |
|---|
| 1367 | x = a * b + c; |
|---|
| 1368 | </pre> |
|---|
| 1369 | instead of |
|---|
| 1370 | <pre> |
|---|
| 1371 | mul(x, a, b); |
|---|
| 1372 | add(x, x, c); |
|---|
| 1373 | </pre> |
|---|
| 1374 | as one must in older versions of NTL. |
|---|
| 1375 | The operator notation leads to somewhat less efficient code, |
|---|
| 1376 | and one can always use the old notation in situations |
|---|
| 1377 | where efficiency is critical. |
|---|
| 1378 | Despite the new programming interface, |
|---|
| 1379 | care has been taken to ensure backward compitability; |
|---|
| 1380 | pre-existing programs that use NTL should still work. |
|---|
| 1381 | |
|---|
| 1382 | <li> |
|---|
| 1383 | Windows port. |
|---|
| 1384 | |
|---|
| 1385 | <li> |
|---|
| 1386 | Added compile-time flag that allows one to exploit |
|---|
| 1387 | "long long" data type if it exists (this especially helps on Pentium/Linux |
|---|
| 1388 | platforms). |
|---|
| 1389 | |
|---|
| 1390 | <li> |
|---|
| 1391 | Added compile-time flag to get better quad_float code on |
|---|
| 1392 | Pentium/Linux platforms. |
|---|
| 1393 | |
|---|
| 1394 | <li> |
|---|
| 1395 | A few bug fixes and performance tuning. |
|---|
| 1396 | </ul> |
|---|
| 1397 | |
|---|
| 1398 | <p> |
|---|
| 1399 | <hr> |
|---|
| 1400 | |
|---|
| 1401 | |
|---|
| 1402 | <p> |
|---|
| 1403 | <h3> |
|---|
| 1404 | Changes between NTL 1.5 and NTL 1.7 |
|---|
| 1405 | </h3> |
|---|
| 1406 | <p> |
|---|
| 1407 | |
|---|
| 1408 | <ul> |
|---|
| 1409 | <li> |
|---|
| 1410 | Incorporation of Keith Briggs' quadratic precision package. |
|---|
| 1411 | |
|---|
| 1412 | <li> |
|---|
| 1413 | Much faster and more robust lattice basis reduction, |
|---|
| 1414 | including Schnorr-Horner "volume heuristic" for Block Korkin |
|---|
| 1415 | Zolotarev reductions, and a new quadratic precision LLL variant |
|---|
| 1416 | that is much more robust. |
|---|
| 1417 | |
|---|
| 1418 | <li> |
|---|
| 1419 | A few bug fixes. |
|---|
| 1420 | |
|---|
| 1421 | </ul> |
|---|
| 1422 | |
|---|
| 1423 | |
|---|
| 1424 | <p> |
|---|
| 1425 | <hr> |
|---|
| 1426 | |
|---|
| 1427 | <p> |
|---|
| 1428 | <h3> |
|---|
| 1429 | Changes between NTL 1.0 and NTL 1.5 |
|---|
| 1430 | </h3> |
|---|
| 1431 | <p> |
|---|
| 1432 | |
|---|
| 1433 | |
|---|
| 1434 | <ul> |
|---|
| 1435 | <li> |
|---|
| 1436 | Implementation of Schnorr-Euchner algorithms for |
|---|
| 1437 | lattice basis reduction, including deep insertions and |
|---|
| 1438 | block Korkin Zolotarev reduction. |
|---|
| 1439 | These are significantly faster than the LLL algorithm |
|---|
| 1440 | in NTL 1.0. |
|---|
| 1441 | |
|---|
| 1442 | <li> |
|---|
| 1443 | Implementation of arbitrary-precision floating point. |
|---|
| 1444 | |
|---|
| 1445 | <li> |
|---|
| 1446 | Implementation of double precision with extended exponent range, |
|---|
| 1447 | which is useful for lattice basis reduction when the coefficients |
|---|
| 1448 | are large. |
|---|
| 1449 | |
|---|
| 1450 | <li> |
|---|
| 1451 | Faster polynomial multiplication over the integers, |
|---|
| 1452 | incorporating the Schoenhagge-Strassen method. |
|---|
| 1453 | |
|---|
| 1454 | <li> |
|---|
| 1455 | Compilation flags that increase performance on machines |
|---|
| 1456 | with poor floating-point performance. |
|---|
| 1457 | |
|---|
| 1458 | <li> |
|---|
| 1459 | Sundry performance tuning and a few bug fixes. |
|---|
| 1460 | |
|---|
| 1461 | </ul> |
|---|
| 1462 | |
|---|
| 1463 | <center> |
|---|
| 1464 | <a href="tour-roadmap.html"><img src="arrow1.gif" alt="[Previous]" align=bottom></a> |
|---|
| 1465 | <a href="tour.html"><img src="arrow2.gif" alt="[Up]" align=bottom></a> |
|---|
| 1466 | <a href="tour-ack.html"> <img src="arrow3.gif" alt="[Next]" align=bottom></a> |
|---|
| 1467 | </center> |
|---|
| 1468 | |
|---|
| 1469 | </body> |
|---|
| 1470 | </html> |
|---|
