// Copyright 2018 Ulf Adams // // The contents of this file may be used under the terms of the Apache License, // Version 2.0. // // (See accompanying file LICENSE-Apache or copy at // http://www.apache.org/licenses/LICENSE-2.0) // // Alternatively, the contents of this file may be used under the terms of // the Boost Software License, Version 1.0. // (See accompanying file LICENSE-Boost or copy at // https://www.boost.org/LICENSE_1_0.txt) // // Unless required by applicable law or agreed to in writing, this software // is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY // KIND, either express or implied. // Runtime compiler options: // -DRYU_DEBUG Generate verbose debugging output to stdout. // // -DRYU_ONLY_64_BIT_OPS Avoid using uint128_t or 64-bit intrinsics. Slower, // depending on your compiler. // // -DRYU_OPTIMIZE_SIZE Use smaller lookup tables. Instead of storing every // required power of 5, only store every 26th entry, and compute // intermediate values with a multiplication. This reduces the lookup table // size by about 10x (only one case, and only double) at the cost of some // performance. Currently requires MSVC intrinsics. #ifdef RYU_DEBUG #endif // Include either the small or the full lookup tables depending on the mode. #if defined(RYU_OPTIMIZE_SIZE) #else #endif #define DOUBLE_MANTISSA_BITS 52 #define DOUBLE_EXPONENT_BITS 11 #define DOUBLE_BIAS 1023 static inline uint32_t decimalLength17(const uint64_t v) { // This is slightly faster than a loop. // The average output length is 16.38 digits, so we check high-to-low. // Function precondition: v is not an 18, 19, or 20-digit number. // (17 digits are sufficient for round-tripping.) assert(v < 100000000000000000L); if (v >= 10000000000000000L) { return 17; } if (v >= 1000000000000000L) { return 16; } if (v >= 100000000000000L) { return 15; } if (v >= 10000000000000L) { return 14; } if (v >= 1000000000000L) { return 13; } if (v >= 100000000000L) { return 12; } if (v >= 10000000000L) { return 11; } if (v >= 1000000000L) { return 10; } if (v >= 100000000L) { return 9; } if (v >= 10000000L) { return 8; } if (v >= 1000000L) { return 7; } if (v >= 100000L) { return 6; } if (v >= 10000L) { return 5; } if (v >= 1000L) { return 4; } if (v >= 100L) { return 3; } if (v >= 10L) { return 2; } return 1; } // A floating decimal representing m * 10^e. typedef struct floating_decimal_64 { uint64_t mantissa; // Decimal exponent's range is -324 to 308 // inclusive, and can fit in a short if needed. int32_t exponent; bool sign; } floating_decimal_64; static inline floating_decimal_64 d2d(const uint64_t ieeeMantissa, const uint32_t ieeeExponent, const bool ieeeSign) { int32_t e2; uint64_t m2; if (ieeeExponent == 0) { // We subtract 2 so that the bounds computation has 2 additional bits. e2 = 1 - DOUBLE_BIAS - DOUBLE_MANTISSA_BITS - 2; m2 = ieeeMantissa; } else { e2 = (int32_t) ieeeExponent - DOUBLE_BIAS - DOUBLE_MANTISSA_BITS - 2; m2 = (1ull << DOUBLE_MANTISSA_BITS) | ieeeMantissa; } const bool even = (m2 & 1) == 0; const bool acceptBounds = even; #ifdef RYU_DEBUG printf("-> %" PRIu64 " * 2^%d\n", m2, e2 + 2); #endif // Step 2: Determine the interval of valid decimal representations. const uint64_t mv = 4 * m2; // Implicit bool -> int conversion. True is 1, false is 0. const uint32_t mmShift = ieeeMantissa != 0 || ieeeExponent <= 1; // We would compute mp and mm like this: // uint64_t mp = 4 * m2 + 2; // uint64_t mm = mv - 1 - mmShift; // Step 3: Convert to a decimal power base using 128-bit arithmetic. uint64_t vr, vp, vm; int32_t e10; bool vmIsTrailingZeros = false; bool vrIsTrailingZeros = false; if (e2 >= 0) { // I tried special-casing q == 0, but there was no effect on performance. // This expression is slightly faster than max(0, log10Pow2(e2) - 1). const uint32_t q = log10Pow2(e2) - (e2 > 3); e10 = (int32_t) q; const int32_t k = DOUBLE_POW5_INV_BITCOUNT + pow5bits((int32_t) q) - 1; const int32_t i = -e2 + (int32_t) q + k; #if defined(RYU_OPTIMIZE_SIZE) uint64_t pow5[2]; double_computeInvPow5(q, pow5); vr = mulShiftAll64(m2, pow5, i, &vp, &vm, mmShift); #else vr = mulShiftAll64(m2, DOUBLE_POW5_INV_SPLIT[q], i, &vp, &vm, mmShift); #endif #ifdef RYU_DEBUG printf("%" PRIu64 " * 2^%d / 10^%u\n", mv, e2, q); printf("V+=%" PRIu64 "\nV =%" PRIu64 "\nV-=%" PRIu64 "\n", vp, vr, vm); #endif if (q <= 21) { // This should use q <= 22, but I think 21 is also safe. Smaller values // may still be safe, but it's more difficult to reason about them. // Only one of mp, mv, and mm can be a multiple of 5, if any. const uint32_t mvMod5 = ((uint32_t) mv) - 5 * ((uint32_t) div5(mv)); if (mvMod5 == 0) { vrIsTrailingZeros = multipleOfPowerOf5(mv, q); } else if (acceptBounds) { // Same as min(e2 + (~mm & 1), pow5Factor(mm)) >= q // <=> e2 + (~mm & 1) >= q && pow5Factor(mm) >= q // <=> true && pow5Factor(mm) >= q, since e2 >= q. vmIsTrailingZeros = multipleOfPowerOf5(mv - 1 - mmShift, q); } else { // Same as min(e2 + 1, pow5Factor(mp)) >= q. vp -= multipleOfPowerOf5(mv + 2, q); } } } else { // This expression is slightly faster than max(0, log10Pow5(-e2) - 1). const uint32_t q = log10Pow5(-e2) - (-e2 > 1); e10 = (int32_t) q + e2; const int32_t i = -e2 - (int32_t) q; const int32_t k = pow5bits(i) - DOUBLE_POW5_BITCOUNT; const int32_t j = (int32_t) q - k; #if defined(RYU_OPTIMIZE_SIZE) uint64_t pow5[2]; double_computePow5(i, pow5); vr = mulShiftAll64(m2, pow5, j, &vp, &vm, mmShift); #else vr = mulShiftAll64(m2, DOUBLE_POW5_SPLIT[i], j, &vp, &vm, mmShift); #endif #ifdef RYU_DEBUG printf("%" PRIu64 " * 5^%d / 10^%u\n", mv, -e2, q); printf("%u %d %d %d\n", q, i, k, j); printf("V+=%" PRIu64 "\nV =%" PRIu64 "\nV-=%" PRIu64 "\n", vp, vr, vm); #endif if (q <= 1) { // {vr,vp,vm} is trailing zeros if {mv,mp,mm} has at least q trailing 0 bits. // mv = 4 * m2, so it always has at least two trailing 0 bits. vrIsTrailingZeros = true; if (acceptBounds) { // mm = mv - 1 - mmShift, so it has 1 trailing 0 bit iff mmShift == 1. vmIsTrailingZeros = mmShift == 1; } else { // mp = mv + 2, so it always has at least one trailing 0 bit. --vp; } } else if (q < 63) { // TODO(ulfjack): Use a tighter bound here. // We want to know if the full product has at least q trailing zeros. // We need to compute min(p2(mv), p5(mv) - e2) >= q // <=> p2(mv) >= q && p5(mv) - e2 >= q // <=> p2(mv) >= q (because -e2 >= q) vrIsTrailingZeros = multipleOfPowerOf2(mv, q); #ifdef RYU_DEBUG printf("vr is trailing zeros=%s\n", vrIsTrailingZeros ? "true" : "false"); #endif } } #ifdef RYU_DEBUG printf("e10=%d\n", e10); printf("V+=%" PRIu64 "\nV =%" PRIu64 "\nV-=%" PRIu64 "\n", vp, vr, vm); printf("vm is trailing zeros=%s\n", vmIsTrailingZeros ? "true" : "false"); printf("vr is trailing zeros=%s\n", vrIsTrailingZeros ? "true" : "false"); #endif // Step 4: Find the shortest decimal representation in the interval of valid representations. int32_t removed = 0; uint8_t lastRemovedDigit = 0; uint64_t output; // On average, we remove ~2 digits. if (vmIsTrailingZeros || vrIsTrailingZeros) { // General case, which happens rarely (~0.7%). for (;;) { const uint64_t vpDiv10 = div10(vp); const uint64_t vmDiv10 = div10(vm); if (vpDiv10 <= vmDiv10) { break; } const uint32_t vmMod10 = ((uint32_t) vm) - 10 * ((uint32_t) vmDiv10); const uint64_t vrDiv10 = div10(vr); const uint32_t vrMod10 = ((uint32_t) vr) - 10 * ((uint32_t) vrDiv10); vmIsTrailingZeros &= vmMod10 == 0; vrIsTrailingZeros &= lastRemovedDigit == 0; lastRemovedDigit = (uint8_t) vrMod10; vr = vrDiv10; vp = vpDiv10; vm = vmDiv10; ++removed; } #ifdef RYU_DEBUG printf("V+=%" PRIu64 "\nV =%" PRIu64 "\nV-=%" PRIu64 "\n", vp, vr, vm); printf("d-10=%s\n", vmIsTrailingZeros ? "true" : "false"); #endif if (vmIsTrailingZeros) { for (;;) { const uint64_t vmDiv10 = div10(vm); const uint32_t vmMod10 = ((uint32_t) vm) - 10 * ((uint32_t) vmDiv10); if (vmMod10 != 0) { break; } const uint64_t vpDiv10 = div10(vp); const uint64_t vrDiv10 = div10(vr); const uint32_t vrMod10 = ((uint32_t) vr) - 10 * ((uint32_t) vrDiv10); vrIsTrailingZeros &= lastRemovedDigit == 0; lastRemovedDigit = (uint8_t) vrMod10; vr = vrDiv10; vp = vpDiv10; vm = vmDiv10; ++removed; } } #ifdef RYU_DEBUG printf("%" PRIu64 " %d\n", vr, lastRemovedDigit); printf("vr is trailing zeros=%s\n", vrIsTrailingZeros ? "true" : "false"); #endif if (vrIsTrailingZeros && lastRemovedDigit == 5 && vr % 2 == 0) { // Round even if the exact number is .....50..0. lastRemovedDigit = 4; } // We need to take vr + 1 if vr is outside bounds or we need to round up. output = vr + ((vr == vm && (!acceptBounds || !vmIsTrailingZeros)) || lastRemovedDigit >= 5); } else { // Specialized for the common case (~99.3%). Percentages below are relative to this. bool roundUp = false; const uint64_t vpDiv100 = div100(vp); const uint64_t vmDiv100 = div100(vm); if (vpDiv100 > vmDiv100) { // Optimization: remove two digits at a time (~86.2%). const uint64_t vrDiv100 = div100(vr); const uint32_t vrMod100 = ((uint32_t) vr) - 100 * ((uint32_t) vrDiv100); roundUp = vrMod100 >= 50; vr = vrDiv100; vp = vpDiv100; vm = vmDiv100; removed += 2; } // Loop iterations below (approximately), without optimization above: // 0: 0.03%, 1: 13.8%, 2: 70.6%, 3: 14.0%, 4: 1.40%, 5: 0.14%, 6+: 0.02% // Loop iterations below (approximately), with optimization above: // 0: 70.6%, 1: 27.8%, 2: 1.40%, 3: 0.14%, 4+: 0.02% for (;;) { const uint64_t vpDiv10 = div10(vp); const uint64_t vmDiv10 = div10(vm); if (vpDiv10 <= vmDiv10) { break; } const uint64_t vrDiv10 = div10(vr); const uint32_t vrMod10 = ((uint32_t) vr) - 10 * ((uint32_t) vrDiv10); roundUp = vrMod10 >= 5; vr = vrDiv10; vp = vpDiv10; vm = vmDiv10; ++removed; } #ifdef RYU_DEBUG printf("%" PRIu64 " roundUp=%s\n", vr, roundUp ? "true" : "false"); printf("vr is trailing zeros=%s\n", vrIsTrailingZeros ? "true" : "false"); #endif // We need to take vr + 1 if vr is outside bounds or we need to round up. output = vr + (vr == vm || roundUp); } const int32_t exp = e10 + removed; #ifdef RYU_DEBUG printf("V+=%" PRIu64 "\nV =%" PRIu64 "\nV-=%" PRIu64 "\n", vp, vr, vm); printf("O=%" PRIu64 "\n", output); printf("EXP=%d\n", exp); #endif floating_decimal_64 fd; fd.exponent = exp; fd.mantissa = output; fd.sign = ieeeSign; return fd; } static inline int to_chars(const floating_decimal_64 v, char* const result) { // Step 5: Print the decimal representation. int index = 0; if (v.sign) { result[index++] = '-'; } uint64_t output = v.mantissa; const uint32_t olength = decimalLength17(output); #ifdef RYU_DEBUG printf("DIGITS=%" PRIu64 "\n", v.mantissa); printf("OLEN=%u\n", olength); printf("EXP=%u\n", v.exponent + olength); #endif // Print the decimal digits. // The following code is equivalent to: // for (uint32_t i = 0; i < olength - 1; ++i) { // const uint32_t c = output % 10; output /= 10; // result[index + olength - i] = (char) ('0' + c); // } // result[index] = '0' + output % 10; uint32_t i = 0; // We prefer 32-bit operations, even on 64-bit platforms. // We have at most 17 digits, and uint32_t can store 9 digits. // If output doesn't fit into uint32_t, we cut off 8 digits, // so the rest will fit into uint32_t. if ((output >> 32) != 0) { // Expensive 64-bit division. const uint64_t q = div1e8(output); uint32_t output2 = ((uint32_t) output) - 100000000 * ((uint32_t) q); output = q; const uint32_t c = output2 % 10000; output2 /= 10000; const uint32_t d = output2 % 10000; const uint32_t c0 = (c % 100) << 1; const uint32_t c1 = (c / 100) << 1; const uint32_t d0 = (d % 100) << 1; const uint32_t d1 = (d / 100) << 1; memcpy(result + index + olength - i - 1, DIGIT_TABLE + c0, 2); memcpy(result + index + olength - i - 3, DIGIT_TABLE + c1, 2); memcpy(result + index + olength - i - 5, DIGIT_TABLE + d0, 2); memcpy(result + index + olength - i - 7, DIGIT_TABLE + d1, 2); i += 8; } uint32_t output2 = (uint32_t) output; while (output2 >= 10000) { #ifdef __clang__ // https://bugs.llvm.org/show_bug.cgi?id=38217 const uint32_t c = output2 - 10000 * (output2 / 10000); #else const uint32_t c = output2 % 10000; #endif output2 /= 10000; const uint32_t c0 = (c % 100) << 1; const uint32_t c1 = (c / 100) << 1; memcpy(result + index + olength - i - 1, DIGIT_TABLE + c0, 2); memcpy(result + index + olength - i - 3, DIGIT_TABLE + c1, 2); i += 4; } if (output2 >= 100) { const uint32_t c = (output2 % 100) << 1; output2 /= 100; memcpy(result + index + olength - i - 1, DIGIT_TABLE + c, 2); i += 2; } if (output2 >= 10) { const uint32_t c = output2 << 1; // We can't use memcpy here: the decimal dot goes between these two digits. result[index + olength - i] = DIGIT_TABLE[c + 1]; result[index] = DIGIT_TABLE[c]; } else { result[index] = (char) ('0' + output2); } // Print decimal point if needed. if (olength > 1) { result[index + 1] = '.'; index += olength + 1; } else { ++index; } // Print the exponent. result[index++] = 'e'; int32_t exp = v.exponent + (int32_t) olength - 1; if (exp < 0) { result[index++] = '-'; exp = -exp; } else result[index++] = '+'; if (exp >= 100) { const int32_t c = exp % 10; memcpy(result + index, DIGIT_TABLE + 2 * (exp / 10), 2); result[index + 2] = (char) ('0' + c); index += 3; } else { memcpy(result + index, DIGIT_TABLE + 2 * exp, 2); index += 2; } return index; } static inline bool d2d_small_int(const uint64_t ieeeMantissa, const uint32_t ieeeExponent, const bool ieeeSign, floating_decimal_64* const v) { const uint64_t m2 = (1ull << DOUBLE_MANTISSA_BITS) | ieeeMantissa; const int32_t e2 = (int32_t) ieeeExponent - DOUBLE_BIAS - DOUBLE_MANTISSA_BITS; if (e2 > 0) { // f = m2 * 2^e2 >= 2^53 is an integer. // Ignore this case for now. return false; } if (e2 < -52) { // f < 1. return false; } // Since 2^52 <= m2 < 2^53 and 0 <= -e2 <= 52: 1 <= f = m2 / 2^-e2 < 2^53. // Test if the lower -e2 bits of the significand are 0, i.e. whether the fraction is 0. const uint64_t mask = (1ull << -e2) - 1; const uint64_t fraction = m2 & mask; if (fraction != 0) { return false; } // f is an integer in the range [1, 2^53). // Note: mantissa might contain trailing (decimal) 0's. // Note: since 2^53 < 10^16, there is no need to adjust decimalLength17(). v->mantissa = m2 >> -e2; v->exponent = 0; v->sign = ieeeSign; return true; } floating_decimal_64 floating_to_fd64(double f) { // Step 1: Decode the floating-point number, and unify normalized and subnormal cases. const uint64_t bits = double_to_bits(f); #ifdef RYU_DEBUG printf("IN="); for (int32_t bit = 63; bit >= 0; --bit) { printf("%d", (int) ((bits >> bit) & 1)); } printf("\n"); #endif // Decode bits into sign, mantissa, and exponent. const bool ieeeSign = ((bits >> (DOUBLE_MANTISSA_BITS + DOUBLE_EXPONENT_BITS)) & 1) != 0; const uint64_t ieeeMantissa = bits & ((1ull << DOUBLE_MANTISSA_BITS) - 1); const uint32_t ieeeExponent = (uint32_t) ((bits >> DOUBLE_MANTISSA_BITS) & ((1u << DOUBLE_EXPONENT_BITS) - 1)); // Case distinction; exit early for the easy cases. if (ieeeExponent == ((1u << DOUBLE_EXPONENT_BITS) - 1u) || (ieeeExponent == 0 && ieeeMantissa == 0)) { __builtin_abort(); } floating_decimal_64 v; const bool isSmallInt = d2d_small_int(ieeeMantissa, ieeeExponent, ieeeSign, &v); if (isSmallInt) { // For small integers in the range [1, 2^53), v.mantissa might contain trailing (decimal) zeros. // For scientific notation we need to move these zeros into the exponent. // (This is not needed for fixed-point notation, so it might be beneficial to trim // trailing zeros in to_chars only if needed - once fixed-point notation output is implemented.) for (;;) { const uint64_t q = div10(v.mantissa); const uint32_t r = ((uint32_t) v.mantissa) - 10 * ((uint32_t) q); if (r != 0) { break; } v.mantissa = q; ++v.exponent; } } else { v = d2d(ieeeMantissa, ieeeExponent, ieeeSign); } return v; }