1 /* 2 * Copyright (c) 1999, 2020, Oracle and/or its affiliates. All rights reserved. 3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. 4 * 5 * This code is free software; you can redistribute it and/or modify it 6 * under the terms of the GNU General Public License version 2 only, as 7 * published by the Free Software Foundation. Oracle designates this 8 * particular file as subject to the "Classpath" exception as provided 9 * by Oracle in the LICENSE file that accompanied this code. 10 * 11 * This code is distributed in the hope that it will be useful, but WITHOUT 12 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or 13 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 14 * version 2 for more details (a copy is included in the LICENSE file that 15 * accompanied this code). 16 * 17 * You should have received a copy of the GNU General Public License version 18 * 2 along with this work; if not, write to the Free Software Foundation, 19 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. 20 * 21 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA 22 * or visit www.oracle.com if you need additional information or have any 23 * questions. 24 */ 25 26 package java.lang; 27 28 import java.util.Random; 29 import jdk.internal.math.DoubleConsts; 30 import jdk.internal.HotSpotIntrinsicCandidate; 31 32 /** 33 * The class {@code StrictMath} contains methods for performing basic 34 * numeric operations such as the elementary exponential, logarithm, 35 * square root, and trigonometric functions. 36 * 37 * <p>To help ensure portability of Java programs, the definitions of 38 * some of the numeric functions in this package require that they 39 * produce the same results as certain published algorithms. These 40 * algorithms are available from the well-known network library 41 * {@code netlib} as the package "Freely Distributable Math 42 * Library," <a 43 * href="https://www.netlib.org/fdlibm/">{@code fdlibm}</a>. These 44 * algorithms, which are written in the C programming language, are 45 * then to be understood as executed with all floating-point 46 * operations following the rules of Java floating-point arithmetic. 47 * 48 * <p>The Java math library is defined with respect to 49 * {@code fdlibm} version 5.3. Where {@code fdlibm} provides 50 * more than one definition for a function (such as 51 * {@code acos}), use the "IEEE 754 core function" version 52 * (residing in a file whose name begins with the letter 53 * {@code e}). The methods which require {@code fdlibm} 54 * semantics are {@code sin}, {@code cos}, {@code tan}, 55 * {@code asin}, {@code acos}, {@code atan}, 56 * {@code exp}, {@code log}, {@code log10}, 57 * {@code cbrt}, {@code atan2}, {@code pow}, 58 * {@code sinh}, {@code cosh}, {@code tanh}, 59 * {@code hypot}, {@code expm1}, and {@code log1p}. 60 * 61 * <p> 62 * The platform uses signed two's complement integer arithmetic with 63 * int and long primitive types. The developer should choose 64 * the primitive type to ensure that arithmetic operations consistently 65 * produce correct results, which in some cases means the operations 66 * will not overflow the range of values of the computation. 67 * The best practice is to choose the primitive type and algorithm to avoid 68 * overflow. In cases where the size is {@code int} or {@code long} and 69 * overflow errors need to be detected, the methods {@code addExact}, 70 * {@code subtractExact}, {@code multiplyExact}, {@code toIntExact}, 71 * {@code incrementExact}, {@code decrementExact} and {@code negateExact} 72 * throw an {@code ArithmeticException} when the results overflow. 73 * For the arithmetic operations divide and absolute value, overflow 74 * occurs only with a specific minimum or maximum value and 75 * should be checked against the minimum or maximum as appropriate. 76 * 77 * @author unascribed 78 * @author Joseph D. Darcy 79 * @since 1.3 80 */ 81 82 public final class StrictMath { 83 84 /** 85 * Don't let anyone instantiate this class. 86 */ 87 private StrictMath() {} 88 89 /** 90 * The {@code double} value that is closer than any other to 91 * <i>e</i>, the base of the natural logarithms. 92 */ 93 public static final double E = 2.7182818284590452354; 94 95 /** 96 * The {@code double} value that is closer than any other to 97 * <i>pi</i>, the ratio of the circumference of a circle to its 98 * diameter. 99 */ 100 public static final double PI = 3.14159265358979323846; 101 102 /** 103 * Constant by which to multiply an angular value in degrees to obtain an 104 * angular value in radians. 105 */ 106 private static final double DEGREES_TO_RADIANS = 0.017453292519943295; 107 108 /** 109 * Constant by which to multiply an angular value in radians to obtain an 110 * angular value in degrees. 111 */ 112 113 private static final double RADIANS_TO_DEGREES = 57.29577951308232; 114 115 /** 116 * Returns the trigonometric sine of an angle. Special cases: 117 * <ul><li>If the argument is NaN or an infinity, then the 118 * result is NaN. 119 * <li>If the argument is zero, then the result is a zero with the 120 * same sign as the argument.</ul> 121 * 122 * @param a an angle, in radians. 123 * @return the sine of the argument. 124 */ 125 public static native double sin(double a); 126 127 /** 128 * Returns the trigonometric cosine of an angle. Special cases: 129 * <ul><li>If the argument is NaN or an infinity, then the 130 * result is NaN.</ul> 131 * 132 * @param a an angle, in radians. 133 * @return the cosine of the argument. 134 */ 135 public static native double cos(double a); 136 137 /** 138 * Returns the trigonometric tangent of an angle. Special cases: 139 * <ul><li>If the argument is NaN or an infinity, then the result 140 * is NaN. 141 * <li>If the argument is zero, then the result is a zero with the 142 * same sign as the argument.</ul> 143 * 144 * @param a an angle, in radians. 145 * @return the tangent of the argument. 146 */ 147 public static native double tan(double a); 148 149 /** 150 * Returns the arc sine of a value; the returned angle is in the 151 * range -<i>pi</i>/2 through <i>pi</i>/2. Special cases: 152 * <ul><li>If the argument is NaN or its absolute value is greater 153 * than 1, then the result is NaN. 154 * <li>If the argument is zero, then the result is a zero with the 155 * same sign as the argument.</ul> 156 * 157 * @param a the value whose arc sine is to be returned. 158 * @return the arc sine of the argument. 159 */ 160 public static native double asin(double a); 161 162 /** 163 * Returns the arc cosine of a value; the returned angle is in the 164 * range 0.0 through <i>pi</i>. Special case: 165 * <ul><li>If the argument is NaN or its absolute value is greater 166 * than 1, then the result is NaN.</ul> 167 * 168 * @param a the value whose arc cosine is to be returned. 169 * @return the arc cosine of the argument. 170 */ 171 public static native double acos(double a); 172 173 /** 174 * Returns the arc tangent of a value; the returned angle is in the 175 * range -<i>pi</i>/2 through <i>pi</i>/2. Special cases: 176 * <ul><li>If the argument is NaN, then the result is NaN. 177 * <li>If the argument is zero, then the result is a zero with the 178 * same sign as the argument.</ul> 179 * 180 * @param a the value whose arc tangent is to be returned. 181 * @return the arc tangent of the argument. 182 */ 183 public static native double atan(double a); 184 185 /** 186 * Converts an angle measured in degrees to an approximately 187 * equivalent angle measured in radians. The conversion from 188 * degrees to radians is generally inexact. 189 * 190 * @param angdeg an angle, in degrees 191 * @return the measurement of the angle {@code angdeg} 192 * in radians. 193 */ 194 public static strictfp double toRadians(double angdeg) { 195 // Do not delegate to Math.toRadians(angdeg) because 196 // this method has the strictfp modifier. 197 return angdeg * DEGREES_TO_RADIANS; 198 } 199 200 /** 201 * Converts an angle measured in radians to an approximately 202 * equivalent angle measured in degrees. The conversion from 203 * radians to degrees is generally inexact; users should 204 * <i>not</i> expect {@code cos(toRadians(90.0))} to exactly 205 * equal {@code 0.0}. 206 * 207 * @param angrad an angle, in radians 208 * @return the measurement of the angle {@code angrad} 209 * in degrees. 210 */ 211 public static strictfp double toDegrees(double angrad) { 212 // Do not delegate to Math.toDegrees(angrad) because 213 // this method has the strictfp modifier. 214 return angrad * RADIANS_TO_DEGREES; 215 } 216 217 /** 218 * Returns Euler's number <i>e</i> raised to the power of a 219 * {@code double} value. Special cases: 220 * <ul><li>If the argument is NaN, the result is NaN. 221 * <li>If the argument is positive infinity, then the result is 222 * positive infinity. 223 * <li>If the argument is negative infinity, then the result is 224 * positive zero.</ul> 225 * 226 * @param a the exponent to raise <i>e</i> to. 227 * @return the value <i>e</i><sup>{@code a}</sup>, 228 * where <i>e</i> is the base of the natural logarithms. 229 */ 230 public static double exp(double a) { 231 return FdLibm.Exp.compute(a); 232 } 233 234 /** 235 * Returns the natural logarithm (base <i>e</i>) of a {@code double} 236 * value. Special cases: 237 * <ul><li>If the argument is NaN or less than zero, then the result 238 * is NaN. 239 * <li>If the argument is positive infinity, then the result is 240 * positive infinity. 241 * <li>If the argument is positive zero or negative zero, then the 242 * result is negative infinity.</ul> 243 * 244 * @param a a value 245 * @return the value ln {@code a}, the natural logarithm of 246 * {@code a}. 247 */ 248 public static native double log(double a); 249 250 /** 251 * Returns the base 10 logarithm of a {@code double} value. 252 * Special cases: 253 * 254 * <ul><li>If the argument is NaN or less than zero, then the result 255 * is NaN. 256 * <li>If the argument is positive infinity, then the result is 257 * positive infinity. 258 * <li>If the argument is positive zero or negative zero, then the 259 * result is negative infinity. 260 * <li> If the argument is equal to 10<sup><i>n</i></sup> for 261 * integer <i>n</i>, then the result is <i>n</i>. 262 * </ul> 263 * 264 * @param a a value 265 * @return the base 10 logarithm of {@code a}. 266 * @since 1.5 267 */ 268 public static native double log10(double a); 269 270 /** 271 * Returns the correctly rounded positive square root of a 272 * {@code double} value. 273 * Special cases: 274 * <ul><li>If the argument is NaN or less than zero, then the result 275 * is NaN. 276 * <li>If the argument is positive infinity, then the result is positive 277 * infinity. 278 * <li>If the argument is positive zero or negative zero, then the 279 * result is the same as the argument.</ul> 280 * Otherwise, the result is the {@code double} value closest to 281 * the true mathematical square root of the argument value. 282 * 283 * @param a a value. 284 * @return the positive square root of {@code a}. 285 */ 286 @HotSpotIntrinsicCandidate 287 public static native double sqrt(double a); 288 289 /** 290 * Returns the cube root of a {@code double} value. For 291 * positive finite {@code x}, {@code cbrt(-x) == 292 * -cbrt(x)}; that is, the cube root of a negative value is 293 * the negative of the cube root of that value's magnitude. 294 * Special cases: 295 * 296 * <ul> 297 * 298 * <li>If the argument is NaN, then the result is NaN. 299 * 300 * <li>If the argument is infinite, then the result is an infinity 301 * with the same sign as the argument. 302 * 303 * <li>If the argument is zero, then the result is a zero with the 304 * same sign as the argument. 305 * 306 * </ul> 307 * 308 * @param a a value. 309 * @return the cube root of {@code a}. 310 * @since 1.5 311 */ 312 public static double cbrt(double a) { 313 return FdLibm.Cbrt.compute(a); 314 } 315 316 /** 317 * Computes the remainder operation on two arguments as prescribed 318 * by the IEEE 754 standard. 319 * The remainder value is mathematically equal to 320 * <code>f1 - f2</code> × <i>n</i>, 321 * where <i>n</i> is the mathematical integer closest to the exact 322 * mathematical value of the quotient {@code f1/f2}, and if two 323 * mathematical integers are equally close to {@code f1/f2}, 324 * then <i>n</i> is the integer that is even. If the remainder is 325 * zero, its sign is the same as the sign of the first argument. 326 * Special cases: 327 * <ul><li>If either argument is NaN, or the first argument is infinite, 328 * or the second argument is positive zero or negative zero, then the 329 * result is NaN. 330 * <li>If the first argument is finite and the second argument is 331 * infinite, then the result is the same as the first argument.</ul> 332 * 333 * @param f1 the dividend. 334 * @param f2 the divisor. 335 * @return the remainder when {@code f1} is divided by 336 * {@code f2}. 337 */ 338 public static native double IEEEremainder(double f1, double f2); 339 340 /** 341 * Returns the smallest (closest to negative infinity) 342 * {@code double} value that is greater than or equal to the 343 * argument and is equal to a mathematical integer. Special cases: 344 * <ul><li>If the argument value is already equal to a 345 * mathematical integer, then the result is the same as the 346 * argument. <li>If the argument is NaN or an infinity or 347 * positive zero or negative zero, then the result is the same as 348 * the argument. <li>If the argument value is less than zero but 349 * greater than -1.0, then the result is negative zero.</ul> Note 350 * that the value of {@code StrictMath.ceil(x)} is exactly the 351 * value of {@code -StrictMath.floor(-x)}. 352 * 353 * @param a a value. 354 * @return the smallest (closest to negative infinity) 355 * floating-point value that is greater than or equal to 356 * the argument and is equal to a mathematical integer. 357 */ 358 public static double ceil(double a) { 359 return floorOrCeil(a, -0.0, 1.0, 1.0); 360 } 361 362 /** 363 * Returns the largest (closest to positive infinity) 364 * {@code double} value that is less than or equal to the 365 * argument and is equal to a mathematical integer. Special cases: 366 * <ul><li>If the argument value is already equal to a 367 * mathematical integer, then the result is the same as the 368 * argument. <li>If the argument is NaN or an infinity or 369 * positive zero or negative zero, then the result is the same as 370 * the argument.</ul> 371 * 372 * @param a a value. 373 * @return the largest (closest to positive infinity) 374 * floating-point value that less than or equal to the argument 375 * and is equal to a mathematical integer. 376 */ 377 public static double floor(double a) { 378 return floorOrCeil(a, -1.0, 0.0, -1.0); 379 } 380 381 /** 382 * Internal method to share logic between floor and ceil. 383 * 384 * @param a the value to be floored or ceiled 385 * @param negativeBoundary result for values in (-1, 0) 386 * @param positiveBoundary result for values in (0, 1) 387 * @param increment value to add when the argument is non-integral 388 */ 389 private static double floorOrCeil(double a, 390 double negativeBoundary, 391 double positiveBoundary, 392 double sign) { 393 int exponent = Math.getExponent(a); 394 395 if (exponent < 0) { 396 /* 397 * Absolute value of argument is less than 1. 398 * floorOrceil(-0.0) => -0.0 399 * floorOrceil(+0.0) => +0.0 400 */ 401 return ((a == 0.0) ? a : 402 ( (a < 0.0) ? negativeBoundary : positiveBoundary) ); 403 } else if (exponent >= 52) { 404 /* 405 * Infinity, NaN, or a value so large it must be integral. 406 */ 407 return a; 408 } 409 // Else the argument is either an integral value already XOR it 410 // has to be rounded to one. 411 assert exponent >= 0 && exponent <= 51; 412 413 long doppel = Double.doubleToRawLongBits(a); 414 long mask = DoubleConsts.SIGNIF_BIT_MASK >> exponent; 415 416 if ( (mask & doppel) == 0L ) 417 return a; // integral value 418 else { 419 double result = Double.longBitsToDouble(doppel & (~mask)); 420 if (sign*a > 0.0) 421 result = result + sign; 422 return result; 423 } 424 } 425 426 /** 427 * Returns the {@code double} value that is closest in value 428 * to the argument and is equal to a mathematical integer. If two 429 * {@code double} values that are mathematical integers are 430 * equally close to the value of the argument, the result is the 431 * integer value that is even. Special cases: 432 * <ul><li>If the argument value is already equal to a mathematical 433 * integer, then the result is the same as the argument. 434 * <li>If the argument is NaN or an infinity or positive zero or negative 435 * zero, then the result is the same as the argument.</ul> 436 * 437 * @param a a value. 438 * @return the closest floating-point value to {@code a} that is 439 * equal to a mathematical integer. 440 * @author Joseph D. Darcy 441 */ 442 public static double rint(double a) { 443 /* 444 * If the absolute value of a is not less than 2^52, it 445 * is either a finite integer (the double format does not have 446 * enough significand bits for a number that large to have any 447 * fractional portion), an infinity, or a NaN. In any of 448 * these cases, rint of the argument is the argument. 449 * 450 * Otherwise, the sum (twoToThe52 + a ) will properly round 451 * away any fractional portion of a since ulp(twoToThe52) == 452 * 1.0; subtracting out twoToThe52 from this sum will then be 453 * exact and leave the rounded integer portion of a. 454 * 455 * This method does *not* need to be declared strictfp to get 456 * fully reproducible results. Whether or not a method is 457 * declared strictfp can only make a difference in the 458 * returned result if some operation would overflow or 459 * underflow with strictfp semantics. The operation 460 * (twoToThe52 + a ) cannot overflow since large values of a 461 * are screened out; the add cannot underflow since twoToThe52 462 * is too large. The subtraction ((twoToThe52 + a ) - 463 * twoToThe52) will be exact as discussed above and thus 464 * cannot overflow or meaningfully underflow. Finally, the 465 * last multiply in the return statement is by plus or minus 466 * 1.0, which is exact too. 467 */ 468 double twoToThe52 = (double)(1L << 52); // 2^52 469 double sign = Math.copySign(1.0, a); // preserve sign info 470 a = Math.abs(a); 471 472 if (a < twoToThe52) { // E_min <= ilogb(a) <= 51 473 a = ((twoToThe52 + a ) - twoToThe52); 474 } 475 476 return sign * a; // restore original sign 477 } 478 479 /** 480 * Returns the angle <i>theta</i> from the conversion of rectangular 481 * coordinates ({@code x}, {@code y}) to polar 482 * coordinates (r, <i>theta</i>). 483 * This method computes the phase <i>theta</i> by computing an arc tangent 484 * of {@code y/x} in the range of -<i>pi</i> to <i>pi</i>. Special 485 * cases: 486 * <ul><li>If either argument is NaN, then the result is NaN. 487 * <li>If the first argument is positive zero and the second argument 488 * is positive, or the first argument is positive and finite and the 489 * second argument is positive infinity, then the result is positive 490 * zero. 491 * <li>If the first argument is negative zero and the second argument 492 * is positive, or the first argument is negative and finite and the 493 * second argument is positive infinity, then the result is negative zero. 494 * <li>If the first argument is positive zero and the second argument 495 * is negative, or the first argument is positive and finite and the 496 * second argument is negative infinity, then the result is the 497 * {@code double} value closest to <i>pi</i>. 498 * <li>If the first argument is negative zero and the second argument 499 * is negative, or the first argument is negative and finite and the 500 * second argument is negative infinity, then the result is the 501 * {@code double} value closest to -<i>pi</i>. 502 * <li>If the first argument is positive and the second argument is 503 * positive zero or negative zero, or the first argument is positive 504 * infinity and the second argument is finite, then the result is the 505 * {@code double} value closest to <i>pi</i>/2. 506 * <li>If the first argument is negative and the second argument is 507 * positive zero or negative zero, or the first argument is negative 508 * infinity and the second argument is finite, then the result is the 509 * {@code double} value closest to -<i>pi</i>/2. 510 * <li>If both arguments are positive infinity, then the result is the 511 * {@code double} value closest to <i>pi</i>/4. 512 * <li>If the first argument is positive infinity and the second argument 513 * is negative infinity, then the result is the {@code double} 514 * value closest to 3*<i>pi</i>/4. 515 * <li>If the first argument is negative infinity and the second argument 516 * is positive infinity, then the result is the {@code double} value 517 * closest to -<i>pi</i>/4. 518 * <li>If both arguments are negative infinity, then the result is the 519 * {@code double} value closest to -3*<i>pi</i>/4.</ul> 520 * 521 * @param y the ordinate coordinate 522 * @param x the abscissa coordinate 523 * @return the <i>theta</i> component of the point 524 * (<i>r</i>, <i>theta</i>) 525 * in polar coordinates that corresponds to the point 526 * (<i>x</i>, <i>y</i>) in Cartesian coordinates. 527 */ 528 public static native double atan2(double y, double x); 529 530 /** 531 * Returns the value of the first argument raised to the power of the 532 * second argument. Special cases: 533 * 534 * <ul><li>If the second argument is positive or negative zero, then the 535 * result is 1.0. 536 * <li>If the second argument is 1.0, then the result is the same as the 537 * first argument. 538 * <li>If the second argument is NaN, then the result is NaN. 539 * <li>If the first argument is NaN and the second argument is nonzero, 540 * then the result is NaN. 541 * 542 * <li>If 543 * <ul> 544 * <li>the absolute value of the first argument is greater than 1 545 * and the second argument is positive infinity, or 546 * <li>the absolute value of the first argument is less than 1 and 547 * the second argument is negative infinity, 548 * </ul> 549 * then the result is positive infinity. 550 * 551 * <li>If 552 * <ul> 553 * <li>the absolute value of the first argument is greater than 1 and 554 * the second argument is negative infinity, or 555 * <li>the absolute value of the 556 * first argument is less than 1 and the second argument is positive 557 * infinity, 558 * </ul> 559 * then the result is positive zero. 560 * 561 * <li>If the absolute value of the first argument equals 1 and the 562 * second argument is infinite, then the result is NaN. 563 * 564 * <li>If 565 * <ul> 566 * <li>the first argument is positive zero and the second argument 567 * is greater than zero, or 568 * <li>the first argument is positive infinity and the second 569 * argument is less than zero, 570 * </ul> 571 * then the result is positive zero. 572 * 573 * <li>If 574 * <ul> 575 * <li>the first argument is positive zero and the second argument 576 * is less than zero, or 577 * <li>the first argument is positive infinity and the second 578 * argument is greater than zero, 579 * </ul> 580 * then the result is positive infinity. 581 * 582 * <li>If 583 * <ul> 584 * <li>the first argument is negative zero and the second argument 585 * is greater than zero but not a finite odd integer, or 586 * <li>the first argument is negative infinity and the second 587 * argument is less than zero but not a finite odd integer, 588 * </ul> 589 * then the result is positive zero. 590 * 591 * <li>If 592 * <ul> 593 * <li>the first argument is negative zero and the second argument 594 * is a positive finite odd integer, or 595 * <li>the first argument is negative infinity and the second 596 * argument is a negative finite odd integer, 597 * </ul> 598 * then the result is negative zero. 599 * 600 * <li>If 601 * <ul> 602 * <li>the first argument is negative zero and the second argument 603 * is less than zero but not a finite odd integer, or 604 * <li>the first argument is negative infinity and the second 605 * argument is greater than zero but not a finite odd integer, 606 * </ul> 607 * then the result is positive infinity. 608 * 609 * <li>If 610 * <ul> 611 * <li>the first argument is negative zero and the second argument 612 * is a negative finite odd integer, or 613 * <li>the first argument is negative infinity and the second 614 * argument is a positive finite odd integer, 615 * </ul> 616 * then the result is negative infinity. 617 * 618 * <li>If the first argument is finite and less than zero 619 * <ul> 620 * <li> if the second argument is a finite even integer, the 621 * result is equal to the result of raising the absolute value of 622 * the first argument to the power of the second argument 623 * 624 * <li>if the second argument is a finite odd integer, the result 625 * is equal to the negative of the result of raising the absolute 626 * value of the first argument to the power of the second 627 * argument 628 * 629 * <li>if the second argument is finite and not an integer, then 630 * the result is NaN. 631 * </ul> 632 * 633 * <li>If both arguments are integers, then the result is exactly equal 634 * to the mathematical result of raising the first argument to the power 635 * of the second argument if that result can in fact be represented 636 * exactly as a {@code double} value.</ul> 637 * 638 * <p>(In the foregoing descriptions, a floating-point value is 639 * considered to be an integer if and only if it is finite and a 640 * fixed point of the method {@link #ceil ceil} or, 641 * equivalently, a fixed point of the method {@link #floor 642 * floor}. A value is a fixed point of a one-argument 643 * method if and only if the result of applying the method to the 644 * value is equal to the value.) 645 * 646 * @param a base. 647 * @param b the exponent. 648 * @return the value {@code a}<sup>{@code b}</sup>. 649 */ 650 public static double pow(double a, double b) { 651 return FdLibm.Pow.compute(a, b); 652 } 653 654 /** 655 * Returns the closest {@code int} to the argument, with ties 656 * rounding to positive infinity. 657 * 658 * <p>Special cases: 659 * <ul><li>If the argument is NaN, the result is 0. 660 * <li>If the argument is negative infinity or any value less than or 661 * equal to the value of {@code Integer.MIN_VALUE}, the result is 662 * equal to the value of {@code Integer.MIN_VALUE}. 663 * <li>If the argument is positive infinity or any value greater than or 664 * equal to the value of {@code Integer.MAX_VALUE}, the result is 665 * equal to the value of {@code Integer.MAX_VALUE}.</ul> 666 * 667 * @param a a floating-point value to be rounded to an integer. 668 * @return the value of the argument rounded to the nearest 669 * {@code int} value. 670 * @see java.lang.Integer#MAX_VALUE 671 * @see java.lang.Integer#MIN_VALUE 672 */ 673 public static int round(float a) { 674 return Math.round(a); 675 } 676 677 /** 678 * Returns the closest {@code long} to the argument, with ties 679 * rounding to positive infinity. 680 * 681 * <p>Special cases: 682 * <ul><li>If the argument is NaN, the result is 0. 683 * <li>If the argument is negative infinity or any value less than or 684 * equal to the value of {@code Long.MIN_VALUE}, the result is 685 * equal to the value of {@code Long.MIN_VALUE}. 686 * <li>If the argument is positive infinity or any value greater than or 687 * equal to the value of {@code Long.MAX_VALUE}, the result is 688 * equal to the value of {@code Long.MAX_VALUE}.</ul> 689 * 690 * @param a a floating-point value to be rounded to a 691 * {@code long}. 692 * @return the value of the argument rounded to the nearest 693 * {@code long} value. 694 * @see java.lang.Long#MAX_VALUE 695 * @see java.lang.Long#MIN_VALUE 696 */ 697 public static long round(double a) { 698 return Math.round(a); 699 } 700 701 private static final class RandomNumberGeneratorHolder { 702 static final Random randomNumberGenerator = new Random(); 703 } 704 705 /** 706 * Returns a {@code double} value with a positive sign, greater 707 * than or equal to {@code 0.0} and less than {@code 1.0}. 708 * Returned values are chosen pseudorandomly with (approximately) 709 * uniform distribution from that range. 710 * 711 * <p>When this method is first called, it creates a single new 712 * pseudorandom-number generator, exactly as if by the expression 713 * 714 * <blockquote>{@code new java.util.Random()}</blockquote> 715 * 716 * This new pseudorandom-number generator is used thereafter for 717 * all calls to this method and is used nowhere else. 718 * 719 * <p>This method is properly synchronized to allow correct use by 720 * more than one thread. However, if many threads need to generate 721 * pseudorandom numbers at a great rate, it may reduce contention 722 * for each thread to have its own pseudorandom-number generator. 723 * 724 * @return a pseudorandom {@code double} greater than or equal 725 * to {@code 0.0} and less than {@code 1.0}. 726 * @see Random#nextDouble() 727 */ 728 public static double random() { 729 return RandomNumberGeneratorHolder.randomNumberGenerator.nextDouble(); 730 } 731 732 /** 733 * Returns the sum of its arguments, 734 * throwing an exception if the result overflows an {@code int}. 735 * 736 * @param x the first value 737 * @param y the second value 738 * @return the result 739 * @throws ArithmeticException if the result overflows an int 740 * @see Math#addExact(int,int) 741 * @since 1.8 742 */ 743 public static int addExact(int x, int y) { 744 return Math.addExact(x, y); 745 } 746 747 /** 748 * Returns the sum of its arguments, 749 * throwing an exception if the result overflows a {@code long}. 750 * 751 * @param x the first value 752 * @param y the second value 753 * @return the result 754 * @throws ArithmeticException if the result overflows a long 755 * @see Math#addExact(long,long) 756 * @since 1.8 757 */ 758 public static long addExact(long x, long y) { 759 return Math.addExact(x, y); 760 } 761 762 /** 763 * Returns the difference of the arguments, 764 * throwing an exception if the result overflows an {@code int}. 765 * 766 * @param x the first value 767 * @param y the second value to subtract from the first 768 * @return the result 769 * @throws ArithmeticException if the result overflows an int 770 * @see Math#subtractExact(int,int) 771 * @since 1.8 772 */ 773 public static int subtractExact(int x, int y) { 774 return Math.subtractExact(x, y); 775 } 776 777 /** 778 * Returns the difference of the arguments, 779 * throwing an exception if the result overflows a {@code long}. 780 * 781 * @param x the first value 782 * @param y the second value to subtract from the first 783 * @return the result 784 * @throws ArithmeticException if the result overflows a long 785 * @see Math#subtractExact(long,long) 786 * @since 1.8 787 */ 788 public static long subtractExact(long x, long y) { 789 return Math.subtractExact(x, y); 790 } 791 792 /** 793 * Returns the product of the arguments, 794 * throwing an exception if the result overflows an {@code int}. 795 * 796 * @param x the first value 797 * @param y the second value 798 * @return the result 799 * @throws ArithmeticException if the result overflows an int 800 * @see Math#multiplyExact(int,int) 801 * @since 1.8 802 */ 803 public static int multiplyExact(int x, int y) { 804 return Math.multiplyExact(x, y); 805 } 806 807 /** 808 * Returns the product of the arguments, throwing an exception if the result 809 * overflows a {@code long}. 810 * 811 * @param x the first value 812 * @param y the second value 813 * @return the result 814 * @throws ArithmeticException if the result overflows a long 815 * @see Math#multiplyExact(long,int) 816 * @since 9 817 */ 818 public static long multiplyExact(long x, int y) { 819 return Math.multiplyExact(x, y); 820 } 821 822 /** 823 * Returns the product of the arguments, 824 * throwing an exception if the result overflows a {@code long}. 825 * 826 * @param x the first value 827 * @param y the second value 828 * @return the result 829 * @throws ArithmeticException if the result overflows a long 830 * @see Math#multiplyExact(long,long) 831 * @since 1.8 832 */ 833 public static long multiplyExact(long x, long y) { 834 return Math.multiplyExact(x, y); 835 } 836 837 /** 838 * Returns the argument incremented by one, 839 * throwing an exception if the result overflows an {@code int}. 840 * The overflow only occurs for {@linkplain Integer#MAX_VALUE the maximum value}. 841 * 842 * @param a the value to increment 843 * @return the result 844 * @throws ArithmeticException if the result overflows an int 845 * @see Math#incrementExact(int) 846 * @since 14 847 */ 848 public static int incrementExact(int a) { 849 return Math.incrementExact(a); 850 } 851 852 /** 853 * Returns the argument incremented by one, 854 * throwing an exception if the result overflows a {@code long}. 855 * The overflow only occurs for {@linkplain Long#MAX_VALUE the maximum value}. 856 * 857 * @param a the value to increment 858 * @return the result 859 * @throws ArithmeticException if the result overflows a long 860 * @see Math#incrementExact(long) 861 * @since 14 862 */ 863 public static long incrementExact(long a) { 864 return Math.incrementExact(a); 865 } 866 867 /** 868 * Returns the argument decremented by one, 869 * throwing an exception if the result overflows an {@code int}. 870 * The overflow only occurs for {@linkplain Integer#MIN_VALUE the minimum value}. 871 * 872 * @param a the value to decrement 873 * @return the result 874 * @throws ArithmeticException if the result overflows an int 875 * @see Math#decrementExact(int) 876 * @since 14 877 */ 878 public static int decrementExact(int a) { 879 return Math.decrementExact(a); 880 } 881 882 /** 883 * Returns the argument decremented by one, 884 * throwing an exception if the result overflows a {@code long}. 885 * The overflow only occurs for {@linkplain Long#MIN_VALUE the minimum value}. 886 * 887 * @param a the value to decrement 888 * @return the result 889 * @throws ArithmeticException if the result overflows a long 890 * @see Math#decrementExact(long) 891 * @since 14 892 */ 893 public static long decrementExact(long a) { 894 return Math.decrementExact(a); 895 } 896 897 /** 898 * Returns the negation of the argument, 899 * throwing an exception if the result overflows an {@code int}. 900 * The overflow only occurs for {@linkplain Integer#MIN_VALUE the minimum value}. 901 * 902 * @param a the value to negate 903 * @return the result 904 * @throws ArithmeticException if the result overflows an int 905 * @see Math#negateExact(int) 906 * @since 14 907 */ 908 public static int negateExact(int a) { 909 return Math.negateExact(a); 910 } 911 912 /** 913 * Returns the negation of the argument, 914 * throwing an exception if the result overflows a {@code long}. 915 * The overflow only occurs for {@linkplain Long#MIN_VALUE the minimum value}. 916 * 917 * @param a the value to negate 918 * @return the result 919 * @throws ArithmeticException if the result overflows a long 920 * @see Math#negateExact(long) 921 * @since 14 922 */ 923 public static long negateExact(long a) { 924 return Math.negateExact(a); 925 } 926 927 /** 928 * Returns the value of the {@code long} argument, throwing an exception 929 * if the value overflows an {@code int}. 930 * 931 * @param value the long value 932 * @return the argument as an int 933 * @throws ArithmeticException if the {@code argument} overflows an int 934 * @see Math#toIntExact(long) 935 * @since 1.8 936 */ 937 public static int toIntExact(long value) { 938 return Math.toIntExact(value); 939 } 940 941 /** 942 * Returns the exact mathematical product of the arguments. 943 * 944 * @param x the first value 945 * @param y the second value 946 * @return the result 947 * @see Math#multiplyFull(int,int) 948 * @since 9 949 */ 950 public static long multiplyFull(int x, int y) { 951 return Math.multiplyFull(x, y); 952 } 953 954 /** 955 * Returns as a {@code long} the most significant 64 bits of the 128-bit 956 * product of two 64-bit factors. 957 * 958 * @param x the first value 959 * @param y the second value 960 * @return the result 961 * @see Math#multiplyHigh(long,long) 962 * @since 9 963 */ 964 public static long multiplyHigh(long x, long y) { 965 return Math.multiplyHigh(x, y); 966 } 967 968 /** 969 * Returns the largest (closest to positive infinity) 970 * {@code int} value that is less than or equal to the algebraic quotient. 971 * There is one special case, if the dividend is the 972 * {@linkplain Integer#MIN_VALUE Integer.MIN_VALUE} and the divisor is {@code -1}, 973 * then integer overflow occurs and 974 * the result is equal to the {@code Integer.MIN_VALUE}. 975 * <p> 976 * See {@link Math#floorDiv(int, int) Math.floorDiv} for examples and 977 * a comparison to the integer division {@code /} operator. 978 * 979 * @param x the dividend 980 * @param y the divisor 981 * @return the largest (closest to positive infinity) 982 * {@code int} value that is less than or equal to the algebraic quotient. 983 * @throws ArithmeticException if the divisor {@code y} is zero 984 * @see Math#floorDiv(int, int) 985 * @see Math#floor(double) 986 * @since 1.8 987 */ 988 public static int floorDiv(int x, int y) { 989 return Math.floorDiv(x, y); 990 } 991 992 /** 993 * Returns the largest (closest to positive infinity) 994 * {@code long} value that is less than or equal to the algebraic quotient. 995 * There is one special case, if the dividend is the 996 * {@linkplain Long#MIN_VALUE Long.MIN_VALUE} and the divisor is {@code -1}, 997 * then integer overflow occurs and 998 * the result is equal to {@code Long.MIN_VALUE}. 999 * <p> 1000 * See {@link Math#floorDiv(int, int) Math.floorDiv} for examples and 1001 * a comparison to the integer division {@code /} operator. 1002 * 1003 * @param x the dividend 1004 * @param y the divisor 1005 * @return the largest (closest to positive infinity) 1006 * {@code int} value that is less than or equal to the algebraic quotient. 1007 * @throws ArithmeticException if the divisor {@code y} is zero 1008 * @see Math#floorDiv(long, int) 1009 * @see Math#floor(double) 1010 * @since 9 1011 */ 1012 public static long floorDiv(long x, int y) { 1013 return Math.floorDiv(x, y); 1014 } 1015 1016 /** 1017 * Returns the largest (closest to positive infinity) 1018 * {@code long} value that is less than or equal to the algebraic quotient. 1019 * There is one special case, if the dividend is the 1020 * {@linkplain Long#MIN_VALUE Long.MIN_VALUE} and the divisor is {@code -1}, 1021 * then integer overflow occurs and 1022 * the result is equal to the {@code Long.MIN_VALUE}. 1023 * <p> 1024 * See {@link Math#floorDiv(int, int) Math.floorDiv} for examples and 1025 * a comparison to the integer division {@code /} operator. 1026 * 1027 * @param x the dividend 1028 * @param y the divisor 1029 * @return the largest (closest to positive infinity) 1030 * {@code long} value that is less than or equal to the algebraic quotient. 1031 * @throws ArithmeticException if the divisor {@code y} is zero 1032 * @see Math#floorDiv(long, long) 1033 * @see Math#floor(double) 1034 * @since 1.8 1035 */ 1036 public static long floorDiv(long x, long y) { 1037 return Math.floorDiv(x, y); 1038 } 1039 1040 /** 1041 * Returns the floor modulus of the {@code int} arguments. 1042 * <p> 1043 * The floor modulus is {@code x - (floorDiv(x, y) * y)}, 1044 * has the same sign as the divisor {@code y}, and 1045 * is in the range of {@code -abs(y) < r < +abs(y)}. 1046 * <p> 1047 * The relationship between {@code floorDiv} and {@code floorMod} is such that: 1048 * <ul> 1049 * <li>{@code floorDiv(x, y) * y + floorMod(x, y) == x} 1050 * </ul> 1051 * <p> 1052 * See {@link Math#floorMod(int, int) Math.floorMod} for examples and 1053 * a comparison to the {@code %} operator. 1054 * 1055 * @param x the dividend 1056 * @param y the divisor 1057 * @return the floor modulus {@code x - (floorDiv(x, y) * y)} 1058 * @throws ArithmeticException if the divisor {@code y} is zero 1059 * @see Math#floorMod(int, int) 1060 * @see StrictMath#floorDiv(int, int) 1061 * @since 1.8 1062 */ 1063 public static int floorMod(int x, int y) { 1064 return Math.floorMod(x , y); 1065 } 1066 1067 /** 1068 * Returns the floor modulus of the {@code long} and {@code int} arguments. 1069 * <p> 1070 * The floor modulus is {@code x - (floorDiv(x, y) * y)}, 1071 * has the same sign as the divisor {@code y}, and 1072 * is in the range of {@code -abs(y) < r < +abs(y)}. 1073 * 1074 * <p> 1075 * The relationship between {@code floorDiv} and {@code floorMod} is such that: 1076 * <ul> 1077 * <li>{@code floorDiv(x, y) * y + floorMod(x, y) == x} 1078 * </ul> 1079 * <p> 1080 * See {@link Math#floorMod(int, int) Math.floorMod} for examples and 1081 * a comparison to the {@code %} operator. 1082 * 1083 * @param x the dividend 1084 * @param y the divisor 1085 * @return the floor modulus {@code x - (floorDiv(x, y) * y)} 1086 * @throws ArithmeticException if the divisor {@code y} is zero 1087 * @see Math#floorMod(long, int) 1088 * @see StrictMath#floorDiv(long, int) 1089 * @since 9 1090 */ 1091 public static int floorMod(long x, int y) { 1092 return Math.floorMod(x , y); 1093 } 1094 1095 /** 1096 * Returns the floor modulus of the {@code long} arguments. 1097 * <p> 1098 * The floor modulus is {@code x - (floorDiv(x, y) * y)}, 1099 * has the same sign as the divisor {@code y}, and 1100 * is in the range of {@code -abs(y) < r < +abs(y)}. 1101 * <p> 1102 * The relationship between {@code floorDiv} and {@code floorMod} is such that: 1103 * <ul> 1104 * <li>{@code floorDiv(x, y) * y + floorMod(x, y) == x} 1105 * </ul> 1106 * <p> 1107 * See {@link Math#floorMod(int, int) Math.floorMod} for examples and 1108 * a comparison to the {@code %} operator. 1109 * 1110 * @param x the dividend 1111 * @param y the divisor 1112 * @return the floor modulus {@code x - (floorDiv(x, y) * y)} 1113 * @throws ArithmeticException if the divisor {@code y} is zero 1114 * @see Math#floorMod(long, long) 1115 * @see StrictMath#floorDiv(long, long) 1116 * @since 1.8 1117 */ 1118 public static long floorMod(long x, long y) { 1119 return Math.floorMod(x, y); 1120 } 1121 1122 /** 1123 * Returns the absolute value of an {@code int} value. 1124 * If the argument is not negative, the argument is returned. 1125 * If the argument is negative, the negation of the argument is returned. 1126 * 1127 * <p>Note that if the argument is equal to the value of 1128 * {@link Integer#MIN_VALUE}, the most negative representable 1129 * {@code int} value, the result is that same value, which is 1130 * negative. 1131 * In contrast, the {@link StrictMath#absExact(int)} method throws an 1132 * {@code ArithmeticException} for this value. 1133 * 1134 * @param a the argument whose absolute value is to be determined. 1135 * @return the absolute value of the argument. 1136 * @see Math#absExact(int) 1137 */ 1138 public static int abs(int a) { 1139 return Math.abs(a); 1140 } 1141 1142 /** 1143 * Returns the mathematical absolute value of an {@code int} value 1144 * if it is exactly representable as an {@code int}, throwing 1145 * {@code ArithmeticException} if the result overflows the 1146 * positive {@code int} range. 1147 * 1148 * <p>Since the range of two's complement integers is asymmetric 1149 * with one additional negative value (JLS {@jls 4.2.1}), the 1150 * mathematical absolute value of {@link Integer#MIN_VALUE} 1151 * overflows the positive {@code int} range, so an exception is 1152 * thrown for that argument. 1153 * 1154 * @param a the argument whose absolute value is to be determined 1155 * @return the absolute value of the argument, unless overflow occurs 1156 * @throws ArithmeticException if the argument is {@link Integer#MIN_VALUE} 1157 * @see Math#abs(int) 1158 * @see Math#absExact(int) 1159 * @since 15 1160 */ 1161 public static int absExact(int a) { 1162 return Math.absExact(a); 1163 } 1164 1165 /** 1166 * Returns the absolute value of a {@code long} value. 1167 * If the argument is not negative, the argument is returned. 1168 * If the argument is negative, the negation of the argument is returned. 1169 * 1170 * <p>Note that if the argument is equal to the value of 1171 * {@link Long#MIN_VALUE}, the most negative representable 1172 * {@code long} value, the result is that same value, which 1173 * is negative. 1174 * In contrast, the {@link StrictMath#absExact(long)} method throws an 1175 * {@code ArithmeticException} for this value. 1176 * 1177 * @param a the argument whose absolute value is to be determined. 1178 * @return the absolute value of the argument. 1179 * @see Math#absExact(long) 1180 */ 1181 public static long abs(long a) { 1182 return Math.abs(a); 1183 } 1184 1185 /** 1186 * Returns the mathematical absolute value of an {@code long} value 1187 * if it is exactly representable as an {@code long}, throwing 1188 * {@code ArithmeticException} if the result overflows the 1189 * positive {@code long} range. 1190 * 1191 * <p>Since the range of two's complement integers is asymmetric 1192 * with one additional negative value (JLS {@jls 4.2.1}), the 1193 * mathematical absolute value of {@link Long#MIN_VALUE} overflows 1194 * the positive {@code long} range, so an exception is thrown for 1195 * that argument. 1196 * 1197 * @param a the argument whose absolute value is to be determined 1198 * @return the absolute value of the argument, unless overflow occurs 1199 * @throws ArithmeticException if the argument is {@link Long#MIN_VALUE} 1200 * @see Math#abs(long) 1201 * @see Math#absExact(long) 1202 * @since 15 1203 */ 1204 public static long absExact(long a) { 1205 return Math.absExact(a); 1206 } 1207 1208 /** 1209 * Returns the absolute value of a {@code float} value. 1210 * If the argument is not negative, the argument is returned. 1211 * If the argument is negative, the negation of the argument is returned. 1212 * Special cases: 1213 * <ul><li>If the argument is positive zero or negative zero, the 1214 * result is positive zero. 1215 * <li>If the argument is infinite, the result is positive infinity. 1216 * <li>If the argument is NaN, the result is NaN.</ul> 1217 * 1218 * @apiNote As implied by the above, one valid implementation of 1219 * this method is given by the expression below which computes a 1220 * {@code float} with the same exponent and significand as the 1221 * argument but with a guaranteed zero sign bit indicating a 1222 * positive value: <br> 1223 * {@code Float.intBitsToFloat(0x7fffffff & Float.floatToRawIntBits(a))} 1224 * 1225 * @param a the argument whose absolute value is to be determined 1226 * @return the absolute value of the argument. 1227 */ 1228 public static float abs(float a) { 1229 return Math.abs(a); 1230 } 1231 1232 /** 1233 * Returns the absolute value of a {@code double} value. 1234 * If the argument is not negative, the argument is returned. 1235 * If the argument is negative, the negation of the argument is returned. 1236 * Special cases: 1237 * <ul><li>If the argument is positive zero or negative zero, the result 1238 * is positive zero. 1239 * <li>If the argument is infinite, the result is positive infinity. 1240 * <li>If the argument is NaN, the result is NaN.</ul> 1241 * 1242 * @apiNote As implied by the above, one valid implementation of 1243 * this method is given by the expression below which computes a 1244 * {@code double} with the same exponent and significand as the 1245 * argument but with a guaranteed zero sign bit indicating a 1246 * positive value: <br> 1247 * {@code Double.longBitsToDouble((Double.doubleToRawLongBits(a)<<1)>>>1)} 1248 * 1249 * @param a the argument whose absolute value is to be determined 1250 * @return the absolute value of the argument. 1251 */ 1252 public static double abs(double a) { 1253 return Math.abs(a); 1254 } 1255 1256 /** 1257 * Returns the greater of two {@code int} values. That is, the 1258 * result is the argument closer to the value of 1259 * {@link Integer#MAX_VALUE}. If the arguments have the same value, 1260 * the result is that same value. 1261 * 1262 * @param a an argument. 1263 * @param b another argument. 1264 * @return the larger of {@code a} and {@code b}. 1265 */ 1266 @HotSpotIntrinsicCandidate 1267 public static int max(int a, int b) { 1268 return Math.max(a, b); 1269 } 1270 1271 /** 1272 * Returns the greater of two {@code long} values. That is, the 1273 * result is the argument closer to the value of 1274 * {@link Long#MAX_VALUE}. If the arguments have the same value, 1275 * the result is that same value. 1276 * 1277 * @param a an argument. 1278 * @param b another argument. 1279 * @return the larger of {@code a} and {@code b}. 1280 */ 1281 public static long max(long a, long b) { 1282 return Math.max(a, b); 1283 } 1284 1285 /** 1286 * Returns the greater of two {@code float} values. That is, 1287 * the result is the argument closer to positive infinity. If the 1288 * arguments have the same value, the result is that same 1289 * value. If either value is NaN, then the result is NaN. Unlike 1290 * the numerical comparison operators, this method considers 1291 * negative zero to be strictly smaller than positive zero. If one 1292 * argument is positive zero and the other negative zero, the 1293 * result is positive zero. 1294 * 1295 * @param a an argument. 1296 * @param b another argument. 1297 * @return the larger of {@code a} and {@code b}. 1298 */ 1299 @HotSpotIntrinsicCandidate 1300 public static float max(float a, float b) { 1301 return Math.max(a, b); 1302 } 1303 1304 /** 1305 * Returns the greater of two {@code double} values. That 1306 * is, the result is the argument closer to positive infinity. If 1307 * the arguments have the same value, the result is that same 1308 * value. If either value is NaN, then the result is NaN. Unlike 1309 * the numerical comparison operators, this method considers 1310 * negative zero to be strictly smaller than positive zero. If one 1311 * argument is positive zero and the other negative zero, the 1312 * result is positive zero. 1313 * 1314 * @param a an argument. 1315 * @param b another argument. 1316 * @return the larger of {@code a} and {@code b}. 1317 */ 1318 @HotSpotIntrinsicCandidate 1319 public static double max(double a, double b) { 1320 return Math.max(a, b); 1321 } 1322 1323 /** 1324 * Returns the smaller of two {@code int} values. That is, 1325 * the result the argument closer to the value of 1326 * {@link Integer#MIN_VALUE}. If the arguments have the same 1327 * value, the result is that same value. 1328 * 1329 * @param a an argument. 1330 * @param b another argument. 1331 * @return the smaller of {@code a} and {@code b}. 1332 */ 1333 @HotSpotIntrinsicCandidate 1334 public static int min(int a, int b) { 1335 return Math.min(a, b); 1336 } 1337 1338 /** 1339 * Returns the smaller of two {@code long} values. That is, 1340 * the result is the argument closer to the value of 1341 * {@link Long#MIN_VALUE}. If the arguments have the same 1342 * value, the result is that same value. 1343 * 1344 * @param a an argument. 1345 * @param b another argument. 1346 * @return the smaller of {@code a} and {@code b}. 1347 */ 1348 public static long min(long a, long b) { 1349 return Math.min(a, b); 1350 } 1351 1352 /** 1353 * Returns the smaller of two {@code float} values. That is, 1354 * the result is the value closer to negative infinity. If the 1355 * arguments have the same value, the result is that same 1356 * value. If either value is NaN, then the result is NaN. Unlike 1357 * the numerical comparison operators, this method considers 1358 * negative zero to be strictly smaller than positive zero. If 1359 * one argument is positive zero and the other is negative zero, 1360 * the result is negative zero. 1361 * 1362 * @param a an argument. 1363 * @param b another argument. 1364 * @return the smaller of {@code a} and {@code b.} 1365 */ 1366 @HotSpotIntrinsicCandidate 1367 public static float min(float a, float b) { 1368 return Math.min(a, b); 1369 } 1370 1371 /** 1372 * Returns the smaller of two {@code double} values. That 1373 * is, the result is the value closer to negative infinity. If the 1374 * arguments have the same value, the result is that same 1375 * value. If either value is NaN, then the result is NaN. Unlike 1376 * the numerical comparison operators, this method considers 1377 * negative zero to be strictly smaller than positive zero. If one 1378 * argument is positive zero and the other is negative zero, the 1379 * result is negative zero. 1380 * 1381 * @param a an argument. 1382 * @param b another argument. 1383 * @return the smaller of {@code a} and {@code b}. 1384 */ 1385 @HotSpotIntrinsicCandidate 1386 public static double min(double a, double b) { 1387 return Math.min(a, b); 1388 } 1389 1390 /** 1391 * Returns the fused multiply add of the three arguments; that is, 1392 * returns the exact product of the first two arguments summed 1393 * with the third argument and then rounded once to the nearest 1394 * {@code double}. 1395 * 1396 * The rounding is done using the {@linkplain 1397 * java.math.RoundingMode#HALF_EVEN round to nearest even 1398 * rounding mode}. 1399 * 1400 * In contrast, if {@code a * b + c} is evaluated as a regular 1401 * floating-point expression, two rounding errors are involved, 1402 * the first for the multiply operation, the second for the 1403 * addition operation. 1404 * 1405 * <p>Special cases: 1406 * <ul> 1407 * <li> If any argument is NaN, the result is NaN. 1408 * 1409 * <li> If one of the first two arguments is infinite and the 1410 * other is zero, the result is NaN. 1411 * 1412 * <li> If the exact product of the first two arguments is infinite 1413 * (in other words, at least one of the arguments is infinite and 1414 * the other is neither zero nor NaN) and the third argument is an 1415 * infinity of the opposite sign, the result is NaN. 1416 * 1417 * </ul> 1418 * 1419 * <p>Note that {@code fusedMac(a, 1.0, c)} returns the same 1420 * result as ({@code a + c}). However, 1421 * {@code fusedMac(a, b, +0.0)} does <em>not</em> always return the 1422 * same result as ({@code a * b}) since 1423 * {@code fusedMac(-0.0, +0.0, +0.0)} is {@code +0.0} while 1424 * ({@code -0.0 * +0.0}) is {@code -0.0}; {@code fusedMac(a, b, -0.0)} is 1425 * equivalent to ({@code a * b}) however. 1426 * 1427 * @apiNote This method corresponds to the fusedMultiplyAdd 1428 * operation defined in IEEE 754-2008. 1429 * 1430 * @param a a value 1431 * @param b a value 1432 * @param c a value 1433 * 1434 * @return (<i>a</i> × <i>b</i> + <i>c</i>) 1435 * computed, as if with unlimited range and precision, and rounded 1436 * once to the nearest {@code double} value 1437 * 1438 * @since 9 1439 */ 1440 public static double fma(double a, double b, double c) { 1441 return Math.fma(a, b, c); 1442 } 1443 1444 /** 1445 * Returns the fused multiply add of the three arguments; that is, 1446 * returns the exact product of the first two arguments summed 1447 * with the third argument and then rounded once to the nearest 1448 * {@code float}. 1449 * 1450 * The rounding is done using the {@linkplain 1451 * java.math.RoundingMode#HALF_EVEN round to nearest even 1452 * rounding mode}. 1453 * 1454 * In contrast, if {@code a * b + c} is evaluated as a regular 1455 * floating-point expression, two rounding errors are involved, 1456 * the first for the multiply operation, the second for the 1457 * addition operation. 1458 * 1459 * <p>Special cases: 1460 * <ul> 1461 * <li> If any argument is NaN, the result is NaN. 1462 * 1463 * <li> If one of the first two arguments is infinite and the 1464 * other is zero, the result is NaN. 1465 * 1466 * <li> If the exact product of the first two arguments is infinite 1467 * (in other words, at least one of the arguments is infinite and 1468 * the other is neither zero nor NaN) and the third argument is an 1469 * infinity of the opposite sign, the result is NaN. 1470 * 1471 * </ul> 1472 * 1473 * <p>Note that {@code fma(a, 1.0f, c)} returns the same 1474 * result as ({@code a + c}). However, 1475 * {@code fma(a, b, +0.0f)} does <em>not</em> always return the 1476 * same result as ({@code a * b}) since 1477 * {@code fma(-0.0f, +0.0f, +0.0f)} is {@code +0.0f} while 1478 * ({@code -0.0f * +0.0f}) is {@code -0.0f}; {@code fma(a, b, -0.0f)} is 1479 * equivalent to ({@code a * b}) however. 1480 * 1481 * @apiNote This method corresponds to the fusedMultiplyAdd 1482 * operation defined in IEEE 754-2008. 1483 * 1484 * @param a a value 1485 * @param b a value 1486 * @param c a value 1487 * 1488 * @return (<i>a</i> × <i>b</i> + <i>c</i>) 1489 * computed, as if with unlimited range and precision, and rounded 1490 * once to the nearest {@code float} value 1491 * 1492 * @since 9 1493 */ 1494 public static float fma(float a, float b, float c) { 1495 return Math.fma(a, b, c); 1496 } 1497 1498 /** 1499 * Returns the size of an ulp of the argument. An ulp, unit in 1500 * the last place, of a {@code double} value is the positive 1501 * distance between this floating-point value and the {@code 1502 * double} value next larger in magnitude. Note that for non-NaN 1503 * <i>x</i>, <code>ulp(-<i>x</i>) == ulp(<i>x</i>)</code>. 1504 * 1505 * <p>Special Cases: 1506 * <ul> 1507 * <li> If the argument is NaN, then the result is NaN. 1508 * <li> If the argument is positive or negative infinity, then the 1509 * result is positive infinity. 1510 * <li> If the argument is positive or negative zero, then the result is 1511 * {@code Double.MIN_VALUE}. 1512 * <li> If the argument is ±{@code Double.MAX_VALUE}, then 1513 * the result is equal to 2<sup>971</sup>. 1514 * </ul> 1515 * 1516 * @param d the floating-point value whose ulp is to be returned 1517 * @return the size of an ulp of the argument 1518 * @author Joseph D. Darcy 1519 * @since 1.5 1520 */ 1521 public static double ulp(double d) { 1522 return Math.ulp(d); 1523 } 1524 1525 /** 1526 * Returns the size of an ulp of the argument. An ulp, unit in 1527 * the last place, of a {@code float} value is the positive 1528 * distance between this floating-point value and the {@code 1529 * float} value next larger in magnitude. Note that for non-NaN 1530 * <i>x</i>, <code>ulp(-<i>x</i>) == ulp(<i>x</i>)</code>. 1531 * 1532 * <p>Special Cases: 1533 * <ul> 1534 * <li> If the argument is NaN, then the result is NaN. 1535 * <li> If the argument is positive or negative infinity, then the 1536 * result is positive infinity. 1537 * <li> If the argument is positive or negative zero, then the result is 1538 * {@code Float.MIN_VALUE}. 1539 * <li> If the argument is ±{@code Float.MAX_VALUE}, then 1540 * the result is equal to 2<sup>104</sup>. 1541 * </ul> 1542 * 1543 * @param f the floating-point value whose ulp is to be returned 1544 * @return the size of an ulp of the argument 1545 * @author Joseph D. Darcy 1546 * @since 1.5 1547 */ 1548 public static float ulp(float f) { 1549 return Math.ulp(f); 1550 } 1551 1552 /** 1553 * Returns the signum function of the argument; zero if the argument 1554 * is zero, 1.0 if the argument is greater than zero, -1.0 if the 1555 * argument is less than zero. 1556 * 1557 * <p>Special Cases: 1558 * <ul> 1559 * <li> If the argument is NaN, then the result is NaN. 1560 * <li> If the argument is positive zero or negative zero, then the 1561 * result is the same as the argument. 1562 * </ul> 1563 * 1564 * @param d the floating-point value whose signum is to be returned 1565 * @return the signum function of the argument 1566 * @author Joseph D. Darcy 1567 * @since 1.5 1568 */ 1569 public static double signum(double d) { 1570 return Math.signum(d); 1571 } 1572 1573 /** 1574 * Returns the signum function of the argument; zero if the argument 1575 * is zero, 1.0f if the argument is greater than zero, -1.0f if the 1576 * argument is less than zero. 1577 * 1578 * <p>Special Cases: 1579 * <ul> 1580 * <li> If the argument is NaN, then the result is NaN. 1581 * <li> If the argument is positive zero or negative zero, then the 1582 * result is the same as the argument. 1583 * </ul> 1584 * 1585 * @param f the floating-point value whose signum is to be returned 1586 * @return the signum function of the argument 1587 * @author Joseph D. Darcy 1588 * @since 1.5 1589 */ 1590 public static float signum(float f) { 1591 return Math.signum(f); 1592 } 1593 1594 /** 1595 * Returns the hyperbolic sine of a {@code double} value. 1596 * The hyperbolic sine of <i>x</i> is defined to be 1597 * (<i>e<sup>x</sup> - e<sup>-x</sup></i>)/2 1598 * where <i>e</i> is {@linkplain Math#E Euler's number}. 1599 * 1600 * <p>Special cases: 1601 * <ul> 1602 * 1603 * <li>If the argument is NaN, then the result is NaN. 1604 * 1605 * <li>If the argument is infinite, then the result is an infinity 1606 * with the same sign as the argument. 1607 * 1608 * <li>If the argument is zero, then the result is a zero with the 1609 * same sign as the argument. 1610 * 1611 * </ul> 1612 * 1613 * @param x The number whose hyperbolic sine is to be returned. 1614 * @return The hyperbolic sine of {@code x}. 1615 * @since 1.5 1616 */ 1617 public static native double sinh(double x); 1618 1619 /** 1620 * Returns the hyperbolic cosine of a {@code double} value. 1621 * The hyperbolic cosine of <i>x</i> is defined to be 1622 * (<i>e<sup>x</sup> + e<sup>-x</sup></i>)/2 1623 * where <i>e</i> is {@linkplain Math#E Euler's number}. 1624 * 1625 * <p>Special cases: 1626 * <ul> 1627 * 1628 * <li>If the argument is NaN, then the result is NaN. 1629 * 1630 * <li>If the argument is infinite, then the result is positive 1631 * infinity. 1632 * 1633 * <li>If the argument is zero, then the result is {@code 1.0}. 1634 * 1635 * </ul> 1636 * 1637 * @param x The number whose hyperbolic cosine is to be returned. 1638 * @return The hyperbolic cosine of {@code x}. 1639 * @since 1.5 1640 */ 1641 public static native double cosh(double x); 1642 1643 /** 1644 * Returns the hyperbolic tangent of a {@code double} value. 1645 * The hyperbolic tangent of <i>x</i> is defined to be 1646 * (<i>e<sup>x</sup> - e<sup>-x</sup></i>)/(<i>e<sup>x</sup> + e<sup>-x</sup></i>), 1647 * in other words, {@linkplain Math#sinh 1648 * sinh(<i>x</i>)}/{@linkplain Math#cosh cosh(<i>x</i>)}. Note 1649 * that the absolute value of the exact tanh is always less than 1650 * 1. 1651 * 1652 * <p>Special cases: 1653 * <ul> 1654 * 1655 * <li>If the argument is NaN, then the result is NaN. 1656 * 1657 * <li>If the argument is zero, then the result is a zero with the 1658 * same sign as the argument. 1659 * 1660 * <li>If the argument is positive infinity, then the result is 1661 * {@code +1.0}. 1662 * 1663 * <li>If the argument is negative infinity, then the result is 1664 * {@code -1.0}. 1665 * 1666 * </ul> 1667 * 1668 * @param x The number whose hyperbolic tangent is to be returned. 1669 * @return The hyperbolic tangent of {@code x}. 1670 * @since 1.5 1671 */ 1672 public static native double tanh(double x); 1673 1674 /** 1675 * Returns sqrt(<i>x</i><sup>2</sup> +<i>y</i><sup>2</sup>) 1676 * without intermediate overflow or underflow. 1677 * 1678 * <p>Special cases: 1679 * <ul> 1680 * 1681 * <li> If either argument is infinite, then the result 1682 * is positive infinity. 1683 * 1684 * <li> If either argument is NaN and neither argument is infinite, 1685 * then the result is NaN. 1686 * 1687 * </ul> 1688 * 1689 * @param x a value 1690 * @param y a value 1691 * @return sqrt(<i>x</i><sup>2</sup> +<i>y</i><sup>2</sup>) 1692 * without intermediate overflow or underflow 1693 * @since 1.5 1694 */ 1695 public static double hypot(double x, double y) { 1696 return FdLibm.Hypot.compute(x, y); 1697 } 1698 1699 /** 1700 * Returns <i>e</i><sup>x</sup> -1. Note that for values of 1701 * <i>x</i> near 0, the exact sum of 1702 * {@code expm1(x)} + 1 is much closer to the true 1703 * result of <i>e</i><sup>x</sup> than {@code exp(x)}. 1704 * 1705 * <p>Special cases: 1706 * <ul> 1707 * <li>If the argument is NaN, the result is NaN. 1708 * 1709 * <li>If the argument is positive infinity, then the result is 1710 * positive infinity. 1711 * 1712 * <li>If the argument is negative infinity, then the result is 1713 * -1.0. 1714 * 1715 * <li>If the argument is zero, then the result is a zero with the 1716 * same sign as the argument. 1717 * 1718 * </ul> 1719 * 1720 * @param x the exponent to raise <i>e</i> to in the computation of 1721 * <i>e</i><sup>{@code x}</sup> -1. 1722 * @return the value <i>e</i><sup>{@code x}</sup> - 1. 1723 * @since 1.5 1724 */ 1725 public static native double expm1(double x); 1726 1727 /** 1728 * Returns the natural logarithm of the sum of the argument and 1. 1729 * Note that for small values {@code x}, the result of 1730 * {@code log1p(x)} is much closer to the true result of ln(1 1731 * + {@code x}) than the floating-point evaluation of 1732 * {@code log(1.0+x)}. 1733 * 1734 * <p>Special cases: 1735 * <ul> 1736 * 1737 * <li>If the argument is NaN or less than -1, then the result is 1738 * NaN. 1739 * 1740 * <li>If the argument is positive infinity, then the result is 1741 * positive infinity. 1742 * 1743 * <li>If the argument is negative one, then the result is 1744 * negative infinity. 1745 * 1746 * <li>If the argument is zero, then the result is a zero with the 1747 * same sign as the argument. 1748 * 1749 * </ul> 1750 * 1751 * @param x a value 1752 * @return the value ln({@code x} + 1), the natural 1753 * log of {@code x} + 1 1754 * @since 1.5 1755 */ 1756 public static native double log1p(double x); 1757 1758 /** 1759 * Returns the first floating-point argument with the sign of the 1760 * second floating-point argument. For this method, a NaN 1761 * {@code sign} argument is always treated as if it were 1762 * positive. 1763 * 1764 * @param magnitude the parameter providing the magnitude of the result 1765 * @param sign the parameter providing the sign of the result 1766 * @return a value with the magnitude of {@code magnitude} 1767 * and the sign of {@code sign}. 1768 * @since 1.6 1769 */ 1770 public static double copySign(double magnitude, double sign) { 1771 return Math.copySign(magnitude, (Double.isNaN(sign)?1.0d:sign)); 1772 } 1773 1774 /** 1775 * Returns the first floating-point argument with the sign of the 1776 * second floating-point argument. For this method, a NaN 1777 * {@code sign} argument is always treated as if it were 1778 * positive. 1779 * 1780 * @param magnitude the parameter providing the magnitude of the result 1781 * @param sign the parameter providing the sign of the result 1782 * @return a value with the magnitude of {@code magnitude} 1783 * and the sign of {@code sign}. 1784 * @since 1.6 1785 */ 1786 public static float copySign(float magnitude, float sign) { 1787 return Math.copySign(magnitude, (Float.isNaN(sign)?1.0f:sign)); 1788 } 1789 /** 1790 * Returns the unbiased exponent used in the representation of a 1791 * {@code float}. Special cases: 1792 * 1793 * <ul> 1794 * <li>If the argument is NaN or infinite, then the result is 1795 * {@link Float#MAX_EXPONENT} + 1. 1796 * <li>If the argument is zero or subnormal, then the result is 1797 * {@link Float#MIN_EXPONENT} -1. 1798 * </ul> 1799 * @param f a {@code float} value 1800 * @return the unbiased exponent of the argument 1801 * @since 1.6 1802 */ 1803 public static int getExponent(float f) { 1804 return Math.getExponent(f); 1805 } 1806 1807 /** 1808 * Returns the unbiased exponent used in the representation of a 1809 * {@code double}. Special cases: 1810 * 1811 * <ul> 1812 * <li>If the argument is NaN or infinite, then the result is 1813 * {@link Double#MAX_EXPONENT} + 1. 1814 * <li>If the argument is zero or subnormal, then the result is 1815 * {@link Double#MIN_EXPONENT} -1. 1816 * </ul> 1817 * @param d a {@code double} value 1818 * @return the unbiased exponent of the argument 1819 * @since 1.6 1820 */ 1821 public static int getExponent(double d) { 1822 return Math.getExponent(d); 1823 } 1824 1825 /** 1826 * Returns the floating-point number adjacent to the first 1827 * argument in the direction of the second argument. If both 1828 * arguments compare as equal the second argument is returned. 1829 * 1830 * <p>Special cases: 1831 * <ul> 1832 * <li> If either argument is a NaN, then NaN is returned. 1833 * 1834 * <li> If both arguments are signed zeros, {@code direction} 1835 * is returned unchanged (as implied by the requirement of 1836 * returning the second argument if the arguments compare as 1837 * equal). 1838 * 1839 * <li> If {@code start} is 1840 * ±{@link Double#MIN_VALUE} and {@code direction} 1841 * has a value such that the result should have a smaller 1842 * magnitude, then a zero with the same sign as {@code start} 1843 * is returned. 1844 * 1845 * <li> If {@code start} is infinite and 1846 * {@code direction} has a value such that the result should 1847 * have a smaller magnitude, {@link Double#MAX_VALUE} with the 1848 * same sign as {@code start} is returned. 1849 * 1850 * <li> If {@code start} is equal to ± 1851 * {@link Double#MAX_VALUE} and {@code direction} has a 1852 * value such that the result should have a larger magnitude, an 1853 * infinity with same sign as {@code start} is returned. 1854 * </ul> 1855 * 1856 * @param start starting floating-point value 1857 * @param direction value indicating which of 1858 * {@code start}'s neighbors or {@code start} should 1859 * be returned 1860 * @return The floating-point number adjacent to {@code start} in the 1861 * direction of {@code direction}. 1862 * @since 1.6 1863 */ 1864 public static double nextAfter(double start, double direction) { 1865 return Math.nextAfter(start, direction); 1866 } 1867 1868 /** 1869 * Returns the floating-point number adjacent to the first 1870 * argument in the direction of the second argument. If both 1871 * arguments compare as equal a value equivalent to the second argument 1872 * is returned. 1873 * 1874 * <p>Special cases: 1875 * <ul> 1876 * <li> If either argument is a NaN, then NaN is returned. 1877 * 1878 * <li> If both arguments are signed zeros, a value equivalent 1879 * to {@code direction} is returned. 1880 * 1881 * <li> If {@code start} is 1882 * ±{@link Float#MIN_VALUE} and {@code direction} 1883 * has a value such that the result should have a smaller 1884 * magnitude, then a zero with the same sign as {@code start} 1885 * is returned. 1886 * 1887 * <li> If {@code start} is infinite and 1888 * {@code direction} has a value such that the result should 1889 * have a smaller magnitude, {@link Float#MAX_VALUE} with the 1890 * same sign as {@code start} is returned. 1891 * 1892 * <li> If {@code start} is equal to ± 1893 * {@link Float#MAX_VALUE} and {@code direction} has a 1894 * value such that the result should have a larger magnitude, an 1895 * infinity with same sign as {@code start} is returned. 1896 * </ul> 1897 * 1898 * @param start starting floating-point value 1899 * @param direction value indicating which of 1900 * {@code start}'s neighbors or {@code start} should 1901 * be returned 1902 * @return The floating-point number adjacent to {@code start} in the 1903 * direction of {@code direction}. 1904 * @since 1.6 1905 */ 1906 public static float nextAfter(float start, double direction) { 1907 return Math.nextAfter(start, direction); 1908 } 1909 1910 /** 1911 * Returns the floating-point value adjacent to {@code d} in 1912 * the direction of positive infinity. This method is 1913 * semantically equivalent to {@code nextAfter(d, 1914 * Double.POSITIVE_INFINITY)}; however, a {@code nextUp} 1915 * implementation may run faster than its equivalent 1916 * {@code nextAfter} call. 1917 * 1918 * <p>Special Cases: 1919 * <ul> 1920 * <li> If the argument is NaN, the result is NaN. 1921 * 1922 * <li> If the argument is positive infinity, the result is 1923 * positive infinity. 1924 * 1925 * <li> If the argument is zero, the result is 1926 * {@link Double#MIN_VALUE} 1927 * 1928 * </ul> 1929 * 1930 * @param d starting floating-point value 1931 * @return The adjacent floating-point value closer to positive 1932 * infinity. 1933 * @since 1.6 1934 */ 1935 public static double nextUp(double d) { 1936 return Math.nextUp(d); 1937 } 1938 1939 /** 1940 * Returns the floating-point value adjacent to {@code f} in 1941 * the direction of positive infinity. This method is 1942 * semantically equivalent to {@code nextAfter(f, 1943 * Float.POSITIVE_INFINITY)}; however, a {@code nextUp} 1944 * implementation may run faster than its equivalent 1945 * {@code nextAfter} call. 1946 * 1947 * <p>Special Cases: 1948 * <ul> 1949 * <li> If the argument is NaN, the result is NaN. 1950 * 1951 * <li> If the argument is positive infinity, the result is 1952 * positive infinity. 1953 * 1954 * <li> If the argument is zero, the result is 1955 * {@link Float#MIN_VALUE} 1956 * 1957 * </ul> 1958 * 1959 * @param f starting floating-point value 1960 * @return The adjacent floating-point value closer to positive 1961 * infinity. 1962 * @since 1.6 1963 */ 1964 public static float nextUp(float f) { 1965 return Math.nextUp(f); 1966 } 1967 1968 /** 1969 * Returns the floating-point value adjacent to {@code d} in 1970 * the direction of negative infinity. This method is 1971 * semantically equivalent to {@code nextAfter(d, 1972 * Double.NEGATIVE_INFINITY)}; however, a 1973 * {@code nextDown} implementation may run faster than its 1974 * equivalent {@code nextAfter} call. 1975 * 1976 * <p>Special Cases: 1977 * <ul> 1978 * <li> If the argument is NaN, the result is NaN. 1979 * 1980 * <li> If the argument is negative infinity, the result is 1981 * negative infinity. 1982 * 1983 * <li> If the argument is zero, the result is 1984 * {@code -Double.MIN_VALUE} 1985 * 1986 * </ul> 1987 * 1988 * @param d starting floating-point value 1989 * @return The adjacent floating-point value closer to negative 1990 * infinity. 1991 * @since 1.8 1992 */ 1993 public static double nextDown(double d) { 1994 return Math.nextDown(d); 1995 } 1996 1997 /** 1998 * Returns the floating-point value adjacent to {@code f} in 1999 * the direction of negative infinity. This method is 2000 * semantically equivalent to {@code nextAfter(f, 2001 * Float.NEGATIVE_INFINITY)}; however, a 2002 * {@code nextDown} implementation may run faster than its 2003 * equivalent {@code nextAfter} call. 2004 * 2005 * <p>Special Cases: 2006 * <ul> 2007 * <li> If the argument is NaN, the result is NaN. 2008 * 2009 * <li> If the argument is negative infinity, the result is 2010 * negative infinity. 2011 * 2012 * <li> If the argument is zero, the result is 2013 * {@code -Float.MIN_VALUE} 2014 * 2015 * </ul> 2016 * 2017 * @param f starting floating-point value 2018 * @return The adjacent floating-point value closer to negative 2019 * infinity. 2020 * @since 1.8 2021 */ 2022 public static float nextDown(float f) { 2023 return Math.nextDown(f); 2024 } 2025 2026 /** 2027 * Returns {@code d} × 2028 * 2<sup>{@code scaleFactor}</sup> rounded as if performed 2029 * by a single correctly rounded floating-point multiply to a 2030 * member of the double value set. See the Java 2031 * Language Specification for a discussion of floating-point 2032 * value sets. If the exponent of the result is between {@link 2033 * Double#MIN_EXPONENT} and {@link Double#MAX_EXPONENT}, the 2034 * answer is calculated exactly. If the exponent of the result 2035 * would be larger than {@code Double.MAX_EXPONENT}, an 2036 * infinity is returned. Note that if the result is subnormal, 2037 * precision may be lost; that is, when {@code scalb(x, n)} 2038 * is subnormal, {@code scalb(scalb(x, n), -n)} may not equal 2039 * <i>x</i>. When the result is non-NaN, the result has the same 2040 * sign as {@code d}. 2041 * 2042 * <p>Special cases: 2043 * <ul> 2044 * <li> If the first argument is NaN, NaN is returned. 2045 * <li> If the first argument is infinite, then an infinity of the 2046 * same sign is returned. 2047 * <li> If the first argument is zero, then a zero of the same 2048 * sign is returned. 2049 * </ul> 2050 * 2051 * @param d number to be scaled by a power of two. 2052 * @param scaleFactor power of 2 used to scale {@code d} 2053 * @return {@code d} × 2<sup>{@code scaleFactor}</sup> 2054 * @since 1.6 2055 */ 2056 public static double scalb(double d, int scaleFactor) { 2057 return Math.scalb(d, scaleFactor); 2058 } 2059 2060 /** 2061 * Returns {@code f} × 2062 * 2<sup>{@code scaleFactor}</sup> rounded as if performed 2063 * by a single correctly rounded floating-point multiply to a 2064 * member of the float value set. See the Java 2065 * Language Specification for a discussion of floating-point 2066 * value sets. If the exponent of the result is between {@link 2067 * Float#MIN_EXPONENT} and {@link Float#MAX_EXPONENT}, the 2068 * answer is calculated exactly. If the exponent of the result 2069 * would be larger than {@code Float.MAX_EXPONENT}, an 2070 * infinity is returned. Note that if the result is subnormal, 2071 * precision may be lost; that is, when {@code scalb(x, n)} 2072 * is subnormal, {@code scalb(scalb(x, n), -n)} may not equal 2073 * <i>x</i>. When the result is non-NaN, the result has the same 2074 * sign as {@code f}. 2075 * 2076 * <p>Special cases: 2077 * <ul> 2078 * <li> If the first argument is NaN, NaN is returned. 2079 * <li> If the first argument is infinite, then an infinity of the 2080 * same sign is returned. 2081 * <li> If the first argument is zero, then a zero of the same 2082 * sign is returned. 2083 * </ul> 2084 * 2085 * @param f number to be scaled by a power of two. 2086 * @param scaleFactor power of 2 used to scale {@code f} 2087 * @return {@code f} × 2<sup>{@code scaleFactor}</sup> 2088 * @since 1.6 2089 */ 2090 public static float scalb(float f, int scaleFactor) { 2091 return Math.scalb(f, scaleFactor); 2092 } 2093 }