001/* 002 * Copyright (C) 2011 The Guava Authors 003 * 004 * Licensed under the Apache License, Version 2.0 (the "License"); 005 * you may not use this file except in compliance with the License. 006 * You may obtain a copy of the License at 007 * 008 * http://www.apache.org/licenses/LICENSE-2.0 009 * 010 * Unless required by applicable law or agreed to in writing, software 011 * distributed under the License is distributed on an "AS IS" BASIS, 012 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 013 * See the License for the specific language governing permissions and 014 * limitations under the License. 015 */ 016 017package com.google.common.math; 018 019import static com.google.common.base.Preconditions.checkArgument; 020import static com.google.common.base.Preconditions.checkNotNull; 021import static com.google.common.math.MathPreconditions.checkNonNegative; 022import static com.google.common.math.MathPreconditions.checkPositive; 023import static com.google.common.math.MathPreconditions.checkRoundingUnnecessary; 024import static java.math.RoundingMode.CEILING; 025import static java.math.RoundingMode.FLOOR; 026import static java.math.RoundingMode.HALF_EVEN; 027 028import com.google.common.annotations.GwtCompatible; 029import com.google.common.annotations.GwtIncompatible; 030import com.google.common.annotations.VisibleForTesting; 031 032import java.math.BigDecimal; 033import java.math.BigInteger; 034import java.math.RoundingMode; 035import java.util.ArrayList; 036import java.util.List; 037 038/** 039 * A class for arithmetic on values of type {@code BigInteger}. 040 * 041 * <p>The implementations of many methods in this class are based on material from Henry S. Warren, 042 * Jr.'s <i>Hacker's Delight</i>, (Addison Wesley, 2002). 043 * 044 * <p>Similar functionality for {@code int} and for {@code long} can be found in 045 * {@link IntMath} and {@link LongMath} respectively. 046 * 047 * @author Louis Wasserman 048 * @since 11.0 049 */ 050@GwtCompatible(emulated = true) 051public final class BigIntegerMath { 052 /** 053 * Returns {@code true} if {@code x} represents a power of two. 054 */ 055 public static boolean isPowerOfTwo(BigInteger x) { 056 checkNotNull(x); 057 return x.signum() > 0 && x.getLowestSetBit() == x.bitLength() - 1; 058 } 059 060 /** 061 * Returns the base-2 logarithm of {@code x}, rounded according to the specified rounding mode. 062 * 063 * @throws IllegalArgumentException if {@code x <= 0} 064 * @throws ArithmeticException if {@code mode} is {@link RoundingMode#UNNECESSARY} and {@code x} 065 * is not a power of two 066 */ 067 @SuppressWarnings("fallthrough") 068 public static int log2(BigInteger x, RoundingMode mode) { 069 checkPositive("x", checkNotNull(x)); 070 int logFloor = x.bitLength() - 1; 071 switch (mode) { 072 case UNNECESSARY: 073 checkRoundingUnnecessary(isPowerOfTwo(x)); // fall through 074 case DOWN: 075 case FLOOR: 076 return logFloor; 077 078 case UP: 079 case CEILING: 080 return isPowerOfTwo(x) ? logFloor : logFloor + 1; 081 082 case HALF_DOWN: 083 case HALF_UP: 084 case HALF_EVEN: 085 if (logFloor < SQRT2_PRECOMPUTE_THRESHOLD) { 086 BigInteger halfPower = SQRT2_PRECOMPUTED_BITS.shiftRight( 087 SQRT2_PRECOMPUTE_THRESHOLD - logFloor); 088 if (x.compareTo(halfPower) <= 0) { 089 return logFloor; 090 } else { 091 return logFloor + 1; 092 } 093 } 094 /* 095 * Since sqrt(2) is irrational, log2(x) - logFloor cannot be exactly 0.5 096 * 097 * To determine which side of logFloor.5 the logarithm is, we compare x^2 to 2^(2 * 098 * logFloor + 1). 099 */ 100 BigInteger x2 = x.pow(2); 101 int logX2Floor = x2.bitLength() - 1; 102 return (logX2Floor < 2 * logFloor + 1) ? logFloor : logFloor + 1; 103 104 default: 105 throw new AssertionError(); 106 } 107 } 108 109 /* 110 * The maximum number of bits in a square root for which we'll precompute an explicit half power 111 * of two. This can be any value, but higher values incur more class load time and linearly 112 * increasing memory consumption. 113 */ 114 @VisibleForTesting static final int SQRT2_PRECOMPUTE_THRESHOLD = 256; 115 116 @VisibleForTesting static final BigInteger SQRT2_PRECOMPUTED_BITS = 117 new BigInteger("16a09e667f3bcc908b2fb1366ea957d3e3adec17512775099da2f590b0667322a", 16); 118 119 /** 120 * Returns the base-10 logarithm of {@code x}, rounded according to the specified rounding mode. 121 * 122 * @throws IllegalArgumentException if {@code x <= 0} 123 * @throws ArithmeticException if {@code mode} is {@link RoundingMode#UNNECESSARY} and {@code x} 124 * is not a power of ten 125 */ 126 @GwtIncompatible("TODO") 127 @SuppressWarnings("fallthrough") 128 public static int log10(BigInteger x, RoundingMode mode) { 129 checkPositive("x", x); 130 if (fitsInLong(x)) { 131 return LongMath.log10(x.longValue(), mode); 132 } 133 134 int approxLog10 = (int) (log2(x, FLOOR) * LN_2 / LN_10); 135 BigInteger approxPow = BigInteger.TEN.pow(approxLog10); 136 int approxCmp = approxPow.compareTo(x); 137 138 /* 139 * We adjust approxLog10 and approxPow until they're equal to floor(log10(x)) and 140 * 10^floor(log10(x)). 141 */ 142 143 if (approxCmp > 0) { 144 /* 145 * The code is written so that even completely incorrect approximations will still yield the 146 * correct answer eventually, but in practice this branch should almost never be entered, 147 * and even then the loop should not run more than once. 148 */ 149 do { 150 approxLog10--; 151 approxPow = approxPow.divide(BigInteger.TEN); 152 approxCmp = approxPow.compareTo(x); 153 } while (approxCmp > 0); 154 } else { 155 BigInteger nextPow = BigInteger.TEN.multiply(approxPow); 156 int nextCmp = nextPow.compareTo(x); 157 while (nextCmp <= 0) { 158 approxLog10++; 159 approxPow = nextPow; 160 approxCmp = nextCmp; 161 nextPow = BigInteger.TEN.multiply(approxPow); 162 nextCmp = nextPow.compareTo(x); 163 } 164 } 165 166 int floorLog = approxLog10; 167 BigInteger floorPow = approxPow; 168 int floorCmp = approxCmp; 169 170 switch (mode) { 171 case UNNECESSARY: 172 checkRoundingUnnecessary(floorCmp == 0); 173 // fall through 174 case FLOOR: 175 case DOWN: 176 return floorLog; 177 178 case CEILING: 179 case UP: 180 return floorPow.equals(x) ? floorLog : floorLog + 1; 181 182 case HALF_DOWN: 183 case HALF_UP: 184 case HALF_EVEN: 185 // Since sqrt(10) is irrational, log10(x) - floorLog can never be exactly 0.5 186 BigInteger x2 = x.pow(2); 187 BigInteger halfPowerSquared = floorPow.pow(2).multiply(BigInteger.TEN); 188 return (x2.compareTo(halfPowerSquared) <= 0) ? floorLog : floorLog + 1; 189 default: 190 throw new AssertionError(); 191 } 192 } 193 194 private static final double LN_10 = Math.log(10); 195 private static final double LN_2 = Math.log(2); 196 197 /** 198 * Returns the square root of {@code x}, rounded with the specified rounding mode. 199 * 200 * @throws IllegalArgumentException if {@code x < 0} 201 * @throws ArithmeticException if {@code mode} is {@link RoundingMode#UNNECESSARY} and 202 * {@code sqrt(x)} is not an integer 203 */ 204 @GwtIncompatible("TODO") 205 @SuppressWarnings("fallthrough") 206 public static BigInteger sqrt(BigInteger x, RoundingMode mode) { 207 checkNonNegative("x", x); 208 if (fitsInLong(x)) { 209 return BigInteger.valueOf(LongMath.sqrt(x.longValue(), mode)); 210 } 211 BigInteger sqrtFloor = sqrtFloor(x); 212 switch (mode) { 213 case UNNECESSARY: 214 checkRoundingUnnecessary(sqrtFloor.pow(2).equals(x)); // fall through 215 case FLOOR: 216 case DOWN: 217 return sqrtFloor; 218 case CEILING: 219 case UP: 220 return sqrtFloor.pow(2).equals(x) ? sqrtFloor : sqrtFloor.add(BigInteger.ONE); 221 case HALF_DOWN: 222 case HALF_UP: 223 case HALF_EVEN: 224 BigInteger halfSquare = sqrtFloor.pow(2).add(sqrtFloor); 225 /* 226 * We wish to test whether or not x <= (sqrtFloor + 0.5)^2 = halfSquare + 0.25. Since both 227 * x and halfSquare are integers, this is equivalent to testing whether or not x <= 228 * halfSquare. 229 */ 230 return (halfSquare.compareTo(x) >= 0) ? sqrtFloor : sqrtFloor.add(BigInteger.ONE); 231 default: 232 throw new AssertionError(); 233 } 234 } 235 236 @GwtIncompatible("TODO") 237 private static BigInteger sqrtFloor(BigInteger x) { 238 /* 239 * Adapted from Hacker's Delight, Figure 11-1. 240 * 241 * Using DoubleUtils.bigToDouble, getting a double approximation of x is extremely fast, and 242 * then we can get a double approximation of the square root. Then, we iteratively improve this 243 * guess with an application of Newton's method, which sets guess := (guess + (x / guess)) / 2. 244 * This iteration has the following two properties: 245 * 246 * a) every iteration (except potentially the first) has guess >= floor(sqrt(x)). This is 247 * because guess' is the arithmetic mean of guess and x / guess, sqrt(x) is the geometric mean, 248 * and the arithmetic mean is always higher than the geometric mean. 249 * 250 * b) this iteration converges to floor(sqrt(x)). In fact, the number of correct digits doubles 251 * with each iteration, so this algorithm takes O(log(digits)) iterations. 252 * 253 * We start out with a double-precision approximation, which may be higher or lower than the 254 * true value. Therefore, we perform at least one Newton iteration to get a guess that's 255 * definitely >= floor(sqrt(x)), and then continue the iteration until we reach a fixed point. 256 */ 257 BigInteger sqrt0; 258 int log2 = log2(x, FLOOR); 259 if(log2 < Double.MAX_EXPONENT) { 260 sqrt0 = sqrtApproxWithDoubles(x); 261 } else { 262 int shift = (log2 - DoubleUtils.SIGNIFICAND_BITS) & ~1; // even! 263 /* 264 * We have that x / 2^shift < 2^54. Our initial approximation to sqrtFloor(x) will be 265 * 2^(shift/2) * sqrtApproxWithDoubles(x / 2^shift). 266 */ 267 sqrt0 = sqrtApproxWithDoubles(x.shiftRight(shift)).shiftLeft(shift >> 1); 268 } 269 BigInteger sqrt1 = sqrt0.add(x.divide(sqrt0)).shiftRight(1); 270 if (sqrt0.equals(sqrt1)) { 271 return sqrt0; 272 } 273 do { 274 sqrt0 = sqrt1; 275 sqrt1 = sqrt0.add(x.divide(sqrt0)).shiftRight(1); 276 } while (sqrt1.compareTo(sqrt0) < 0); 277 return sqrt0; 278 } 279 280 @GwtIncompatible("TODO") 281 private static BigInteger sqrtApproxWithDoubles(BigInteger x) { 282 return DoubleMath.roundToBigInteger(Math.sqrt(DoubleUtils.bigToDouble(x)), HALF_EVEN); 283 } 284 285 /** 286 * Returns the result of dividing {@code p} by {@code q}, rounding using the specified 287 * {@code RoundingMode}. 288 * 289 * @throws ArithmeticException if {@code q == 0}, or if {@code mode == UNNECESSARY} and {@code a} 290 * is not an integer multiple of {@code b} 291 */ 292 @GwtIncompatible("TODO") 293 public static BigInteger divide(BigInteger p, BigInteger q, RoundingMode mode){ 294 BigDecimal pDec = new BigDecimal(p); 295 BigDecimal qDec = new BigDecimal(q); 296 return pDec.divide(qDec, 0, mode).toBigIntegerExact(); 297 } 298 299 /** 300 * Returns {@code n!}, that is, the product of the first {@code n} positive 301 * integers, or {@code 1} if {@code n == 0}. 302 * 303 * <p><b>Warning</b>: the result takes <i>O(n log n)</i> space, so use cautiously. 304 * 305 * <p>This uses an efficient binary recursive algorithm to compute the factorial 306 * with balanced multiplies. It also removes all the 2s from the intermediate 307 * products (shifting them back in at the end). 308 * 309 * @throws IllegalArgumentException if {@code n < 0} 310 */ 311 public static BigInteger factorial(int n) { 312 checkNonNegative("n", n); 313 314 // If the factorial is small enough, just use LongMath to do it. 315 if (n < LongMath.factorials.length) { 316 return BigInteger.valueOf(LongMath.factorials[n]); 317 } 318 319 // Pre-allocate space for our list of intermediate BigIntegers. 320 int approxSize = IntMath.divide(n * IntMath.log2(n, CEILING), Long.SIZE, CEILING); 321 ArrayList<BigInteger> bignums = new ArrayList<BigInteger>(approxSize); 322 323 // Start from the pre-computed maximum long factorial. 324 int startingNumber = LongMath.factorials.length; 325 long product = LongMath.factorials[startingNumber - 1]; 326 // Strip off 2s from this value. 327 int shift = Long.numberOfTrailingZeros(product); 328 product >>= shift; 329 330 // Use floor(log2(num)) + 1 to prevent overflow of multiplication. 331 int productBits = LongMath.log2(product, FLOOR) + 1; 332 int bits = LongMath.log2(startingNumber, FLOOR) + 1; 333 // Check for the next power of two boundary, to save us a CLZ operation. 334 int nextPowerOfTwo = 1 << (bits - 1); 335 336 // Iteratively multiply the longs as big as they can go. 337 for (long num = startingNumber; num <= n; num++) { 338 // Check to see if the floor(log2(num)) + 1 has changed. 339 if ((num & nextPowerOfTwo) != 0) { 340 nextPowerOfTwo <<= 1; 341 bits++; 342 } 343 // Get rid of the 2s in num. 344 int tz = Long.numberOfTrailingZeros(num); 345 long normalizedNum = num >> tz; 346 shift += tz; 347 // Adjust floor(log2(num)) + 1. 348 int normalizedBits = bits - tz; 349 // If it won't fit in a long, then we store off the intermediate product. 350 if (normalizedBits + productBits >= Long.SIZE) { 351 bignums.add(BigInteger.valueOf(product)); 352 product = 1; 353 productBits = 0; 354 } 355 product *= normalizedNum; 356 productBits = LongMath.log2(product, FLOOR) + 1; 357 } 358 // Check for leftovers. 359 if (product > 1) { 360 bignums.add(BigInteger.valueOf(product)); 361 } 362 // Efficiently multiply all the intermediate products together. 363 return listProduct(bignums).shiftLeft(shift); 364 } 365 366 static BigInteger listProduct(List<BigInteger> nums) { 367 return listProduct(nums, 0, nums.size()); 368 } 369 370 static BigInteger listProduct(List<BigInteger> nums, int start, int end) { 371 switch (end - start) { 372 case 0: 373 return BigInteger.ONE; 374 case 1: 375 return nums.get(start); 376 case 2: 377 return nums.get(start).multiply(nums.get(start + 1)); 378 case 3: 379 return nums.get(start).multiply(nums.get(start + 1)).multiply(nums.get(start + 2)); 380 default: 381 // Otherwise, split the list in half and recursively do this. 382 int m = (end + start) >>> 1; 383 return listProduct(nums, start, m).multiply(listProduct(nums, m, end)); 384 } 385 } 386 387 /** 388 * Returns {@code n} choose {@code k}, also known as the binomial coefficient of {@code n} and 389 * {@code k}, that is, {@code n! / (k! (n - k)!)}. 390 * 391 * <p><b>Warning</b>: the result can take as much as <i>O(k log n)</i> space. 392 * 393 * @throws IllegalArgumentException if {@code n < 0}, {@code k < 0}, or {@code k > n} 394 */ 395 public static BigInteger binomial(int n, int k) { 396 checkNonNegative("n", n); 397 checkNonNegative("k", k); 398 checkArgument(k <= n, "k (%s) > n (%s)", k, n); 399 if (k > (n >> 1)) { 400 k = n - k; 401 } 402 if (k < LongMath.biggestBinomials.length && n <= LongMath.biggestBinomials[k]) { 403 return BigInteger.valueOf(LongMath.binomial(n, k)); 404 } 405 406 BigInteger accum = BigInteger.ONE; 407 408 long numeratorAccum = n; 409 long denominatorAccum = 1; 410 411 int bits = LongMath.log2(n, RoundingMode.CEILING); 412 413 int numeratorBits = bits; 414 415 for (int i = 1; i < k; i++) { 416 int p = n - i; 417 int q = i + 1; 418 419 // log2(p) >= bits - 1, because p >= n/2 420 421 if (numeratorBits + bits >= Long.SIZE - 1) { 422 // The numerator is as big as it can get without risking overflow. 423 // Multiply numeratorAccum / denominatorAccum into accum. 424 accum = accum 425 .multiply(BigInteger.valueOf(numeratorAccum)) 426 .divide(BigInteger.valueOf(denominatorAccum)); 427 numeratorAccum = p; 428 denominatorAccum = q; 429 numeratorBits = bits; 430 } else { 431 // We can definitely multiply into the long accumulators without overflowing them. 432 numeratorAccum *= p; 433 denominatorAccum *= q; 434 numeratorBits += bits; 435 } 436 } 437 return accum 438 .multiply(BigInteger.valueOf(numeratorAccum)) 439 .divide(BigInteger.valueOf(denominatorAccum)); 440 } 441 442 // Returns true if BigInteger.valueOf(x.longValue()).equals(x). 443 @GwtIncompatible("TODO") 444 static boolean fitsInLong(BigInteger x) { 445 return x.bitLength() <= Long.SIZE - 1; 446 } 447 448 private BigIntegerMath() {} 449}