1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
|
//== llvm/Support/APFloat.h - Arbitrary Precision Floating Point -*- C++ -*-==//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file declares a class to represent arbitrary precision floating
// point values and provide a variety of arithmetic operations on them.
//
//===----------------------------------------------------------------------===//
/* A self-contained host- and target-independent arbitrary-precision
floating-point software implementation. It uses bignum integer
arithmetic as provided by static functions in the APInt class.
The library will work with bignum integers whose parts are any
unsigned type at least 16 bits wide, but 64 bits is recommended.
Written for clarity rather than speed, in particular with a view
to use in the front-end of a cross compiler so that target
arithmetic can be correctly performed on the host. Performance
should nonetheless be reasonable, particularly for its intended
use. It may be useful as a base implementation for a run-time
library during development of a faster target-specific one.
All 5 rounding modes in the IEEE-754R draft are handled correctly
for all implemented operations. Currently implemented operations
are add, subtract, multiply, divide, fused-multiply-add,
conversion-to-float, conversion-to-integer and
conversion-from-integer. New rounding modes (e.g. away from zero)
can be added with three or four lines of code.
Four formats are built-in: IEEE single precision, double
precision, quadruple precision, and x87 80-bit extended double
(when operating with full extended precision). Adding a new
format that obeys IEEE semantics only requires adding two lines of
code: a declaration and definition of the format.
All operations return the status of that operation as an exception
bit-mask, so multiple operations can be done consecutively with
their results or-ed together. The returned status can be useful
for compiler diagnostics; e.g., inexact, underflow and overflow
can be easily diagnosed on constant folding, and compiler
optimizers can determine what exceptions would be raised by
folding operations and optimize, or perhaps not optimize,
accordingly.
At present, underflow tininess is detected after rounding; it
should be straight forward to add support for the before-rounding
case too.
The library reads hexadecimal floating point numbers as per C99,
and correctly rounds if necessary according to the specified
rounding mode. Syntax is required to have been validated by the
caller. It also converts floating point numbers to hexadecimal
text as per the C99 %a and %A conversions. The output precision
(or alternatively the natural minimal precision) can be specified;
if the requested precision is less than the natural precision the
output is correctly rounded for the specified rounding mode.
It also reads decimal floating point numbers and correctly rounds
according to the specified rounding mode.
Conversion to decimal text is not currently implemented.
Non-zero finite numbers are represented internally as a sign bit,
a 16-bit signed exponent, and the significand as an array of
integer parts. After normalization of a number of precision P the
exponent is within the range of the format, and if the number is
not denormal the P-th bit of the significand is set as an explicit
integer bit. For denormals the most significant bit is shifted
right so that the exponent is maintained at the format's minimum,
so that the smallest denormal has just the least significant bit
of the significand set. The sign of zeroes and infinities is
significant; the exponent and significand of such numbers is not
stored, but has a known implicit (deterministic) value: 0 for the
significands, 0 for zero exponent, all 1 bits for infinity
exponent. For NaNs the sign and significand are deterministic,
although not really meaningful, and preserved in non-conversion
operations. The exponent is implicitly all 1 bits.
TODO
====
Some features that may or may not be worth adding:
Binary to decimal conversion (hard).
Optional ability to detect underflow tininess before rounding.
New formats: x87 in single and double precision mode (IEEE apart
from extended exponent range) (hard).
New operations: sqrt, IEEE remainder, C90 fmod, nextafter,
nexttoward.
*/
#ifndef LLVM_FLOAT_H
#define LLVM_FLOAT_H
// APInt contains static functions implementing bignum arithmetic.
#include "llvm/ADT/APInt.h"
namespace llvm {
/* Exponents are stored as signed numbers. */
typedef signed short exponent_t;
struct fltSemantics;
class APSInt;
class StringRef;
/* When bits of a floating point number are truncated, this enum is
used to indicate what fraction of the LSB those bits represented.
It essentially combines the roles of guard and sticky bits. */
enum lostFraction { // Example of truncated bits:
lfExactlyZero, // 000000
lfLessThanHalf, // 0xxxxx x's not all zero
lfExactlyHalf, // 100000
lfMoreThanHalf // 1xxxxx x's not all zero
};
class APFloat {
public:
/* We support the following floating point semantics. */
static const fltSemantics IEEEhalf;
static const fltSemantics IEEEsingle;
static const fltSemantics IEEEdouble;
static const fltSemantics IEEEquad;
static const fltSemantics PPCDoubleDouble;
static const fltSemantics x87DoubleExtended;
/* And this pseudo, used to construct APFloats that cannot
conflict with anything real. */
static const fltSemantics Bogus;
static unsigned int semanticsPrecision(const fltSemantics &);
/* Floating point numbers have a four-state comparison relation. */
enum cmpResult {
cmpLessThan,
cmpEqual,
cmpGreaterThan,
cmpUnordered
};
/* IEEE-754R gives five rounding modes. */
enum roundingMode {
rmNearestTiesToEven,
rmTowardPositive,
rmTowardNegative,
rmTowardZero,
rmNearestTiesToAway
};
// Operation status. opUnderflow or opOverflow are always returned
// or-ed with opInexact.
enum opStatus {
opOK = 0x00,
opInvalidOp = 0x01,
opDivByZero = 0x02,
opOverflow = 0x04,
opUnderflow = 0x08,
opInexact = 0x10
};
// Category of internally-represented number.
enum fltCategory {
fcInfinity,
fcNaN,
fcNormal,
fcZero
};
enum uninitializedTag {
uninitialized
};
// Constructors.
APFloat(const fltSemantics &); // Default construct to 0.0
APFloat(const fltSemantics &, StringRef);
APFloat(const fltSemantics &, integerPart);
APFloat(const fltSemantics &, fltCategory, bool negative);
APFloat(const fltSemantics &, uninitializedTag);
explicit APFloat(double d);
explicit APFloat(float f);
explicit APFloat(const APInt &, bool isIEEE = false);
APFloat(const APFloat &);
~APFloat();
// Convenience "constructors"
static APFloat getZero(const fltSemantics &Sem, bool Negative = false) {
return APFloat(Sem, fcZero, Negative);
}
static APFloat getInf(const fltSemantics &Sem, bool Negative = false) {
return APFloat(Sem, fcInfinity, Negative);
}
/// getNaN - Factory for QNaN values.
///
/// \param Negative - True iff the NaN generated should be negative.
/// \param type - The unspecified fill bits for creating the NaN, 0 by
/// default. The value is truncated as necessary.
static APFloat getNaN(const fltSemantics &Sem, bool Negative = false,
unsigned type = 0) {
if (type) {
APInt fill(64, type);
return getQNaN(Sem, Negative, &fill);
} else {
return getQNaN(Sem, Negative, 0);
}
}
/// getQNan - Factory for QNaN values.
static APFloat getQNaN(const fltSemantics &Sem,
bool Negative = false,
const APInt *payload = 0) {
return makeNaN(Sem, false, Negative, payload);
}
/// getSNan - Factory for SNaN values.
static APFloat getSNaN(const fltSemantics &Sem,
bool Negative = false,
const APInt *payload = 0) {
return makeNaN(Sem, true, Negative, payload);
}
/// getLargest - Returns the largest finite number in the given
/// semantics.
///
/// \param Negative - True iff the number should be negative
static APFloat getLargest(const fltSemantics &Sem, bool Negative = false);
/// getSmallest - Returns the smallest (by magnitude) finite number
/// in the given semantics. Might be denormalized, which implies a
/// relative loss of precision.
///
/// \param Negative - True iff the number should be negative
static APFloat getSmallest(const fltSemantics &Sem, bool Negative = false);
/// getSmallestNormalized - Returns the smallest (by magnitude)
/// normalized finite number in the given semantics.
///
/// \param Negative - True iff the number should be negative
static APFloat getSmallestNormalized(const fltSemantics &Sem,
bool Negative = false);
/// getAllOnesValue - Returns a float which is bitcasted from
/// an all one value int.
///
/// \param BitWidth - Select float type
/// \param isIEEE - If 128 bit number, select between PPC and IEEE
static APFloat getAllOnesValue(unsigned BitWidth, bool isIEEE = false);
/// Profile - Used to insert APFloat objects, or objects that contain
/// APFloat objects, into FoldingSets.
void Profile(FoldingSetNodeID& NID) const;
/// @brief Used by the Bitcode serializer to emit APInts to Bitcode.
void Emit(Serializer& S) const;
/// @brief Used by the Bitcode deserializer to deserialize APInts.
static APFloat ReadVal(Deserializer& D);
/* Arithmetic. */
opStatus add(const APFloat &, roundingMode);
opStatus subtract(const APFloat &, roundingMode);
opStatus multiply(const APFloat &, roundingMode);
opStatus divide(const APFloat &, roundingMode);
/* IEEE remainder. */
opStatus remainder(const APFloat &);
/* C fmod, or llvm frem. */
opStatus mod(const APFloat &, roundingMode);
opStatus fusedMultiplyAdd(const APFloat &, const APFloat &, roundingMode);
/* Sign operations. */
void changeSign();
void clearSign();
void copySign(const APFloat &);
/* Conversions. */
opStatus convert(const fltSemantics &, roundingMode, bool *);
opStatus convertToInteger(integerPart *, unsigned int, bool,
roundingMode, bool *) const;
opStatus convertToInteger(APSInt&, roundingMode, bool *) const;
opStatus convertFromAPInt(const APInt &,
bool, roundingMode);
opStatus convertFromSignExtendedInteger(const integerPart *, unsigned int,
bool, roundingMode);
opStatus convertFromZeroExtendedInteger(const integerPart *, unsigned int,
bool, roundingMode);
opStatus convertFromString(StringRef, roundingMode);
APInt bitcastToAPInt() const;
double convertToDouble() const;
float convertToFloat() const;
/* The definition of equality is not straightforward for floating point,
so we won't use operator==. Use one of the following, or write
whatever it is you really mean. */
// bool operator==(const APFloat &) const; // DO NOT IMPLEMENT
/* IEEE comparison with another floating point number (NaNs
compare unordered, 0==-0). */
cmpResult compare(const APFloat &) const;
/* Bitwise comparison for equality (QNaNs compare equal, 0!=-0). */
bool bitwiseIsEqual(const APFloat &) const;
/* Write out a hexadecimal representation of the floating point
value to DST, which must be of sufficient size, in the C99 form
[-]0xh.hhhhp[+-]d. Return the number of characters written,
excluding the terminating NUL. */
unsigned int convertToHexString(char *dst, unsigned int hexDigits,
bool upperCase, roundingMode) const;
/* Simple queries. */
fltCategory getCategory() const { return category; }
const fltSemantics &getSemantics() const { return *semantics; }
bool isZero() const { return category == fcZero; }
bool isNonZero() const { return category != fcZero; }
bool isNaN() const { return category == fcNaN; }
bool isInfinity() const { return category == fcInfinity; }
bool isNegative() const { return sign; }
bool isPosZero() const { return isZero() && !isNegative(); }
bool isNegZero() const { return isZero() && isNegative(); }
APFloat& operator=(const APFloat &);
/* Return an arbitrary integer value usable for hashing. */
uint32_t getHashValue() const;
/// Converts this value into a decimal string.
///
/// \param FormatPrecision The maximum number of digits of
/// precision to output. If there are fewer digits available,
/// zero padding will not be used unless the value is
/// integral and small enough to be expressed in
/// FormatPrecision digits. 0 means to use the natural
/// precision of the number.
/// \param FormatMaxPadding The maximum number of zeros to
/// consider inserting before falling back to scientific
/// notation. 0 means to always use scientific notation.
///
/// Number Precision MaxPadding Result
/// ------ --------- ---------- ------
/// 1.01E+4 5 2 10100
/// 1.01E+4 4 2 1.01E+4
/// 1.01E+4 5 1 1.01E+4
/// 1.01E-2 5 2 0.0101
/// 1.01E-2 4 2 0.0101
/// 1.01E-2 4 1 1.01E-2
void toString(SmallVectorImpl<char> &Str,
unsigned FormatPrecision = 0,
unsigned FormatMaxPadding = 3) const;
/// getExactInverse - If this value has an exact multiplicative inverse,
/// store it in inv and return true.
bool getExactInverse(APFloat *inv) const;
private:
/* Trivial queries. */
integerPart *significandParts();
const integerPart *significandParts() const;
unsigned int partCount() const;
/* Significand operations. */
integerPart addSignificand(const APFloat &);
integerPart subtractSignificand(const APFloat &, integerPart);
lostFraction addOrSubtractSignificand(const APFloat &, bool subtract);
lostFraction multiplySignificand(const APFloat &, const APFloat *);
lostFraction divideSignificand(const APFloat &);
void incrementSignificand();
void initialize(const fltSemantics *);
void shiftSignificandLeft(unsigned int);
lostFraction shiftSignificandRight(unsigned int);
unsigned int significandLSB() const;
unsigned int significandMSB() const;
void zeroSignificand();
/* Arithmetic on special values. */
opStatus addOrSubtractSpecials(const APFloat &, bool subtract);
opStatus divideSpecials(const APFloat &);
opStatus multiplySpecials(const APFloat &);
opStatus modSpecials(const APFloat &);
/* Miscellany. */
static APFloat makeNaN(const fltSemantics &Sem, bool SNaN, bool Negative,
const APInt *fill);
void makeNaN(bool SNaN = false, bool Neg = false, const APInt *fill = 0);
opStatus normalize(roundingMode, lostFraction);
opStatus addOrSubtract(const APFloat &, roundingMode, bool subtract);
cmpResult compareAbsoluteValue(const APFloat &) const;
opStatus handleOverflow(roundingMode);
bool roundAwayFromZero(roundingMode, lostFraction, unsigned int) const;
opStatus convertToSignExtendedInteger(integerPart *, unsigned int, bool,
roundingMode, bool *) const;
opStatus convertFromUnsignedParts(const integerPart *, unsigned int,
roundingMode);
opStatus convertFromHexadecimalString(StringRef, roundingMode);
opStatus convertFromDecimalString(StringRef, roundingMode);
char *convertNormalToHexString(char *, unsigned int, bool,
roundingMode) const;
opStatus roundSignificandWithExponent(const integerPart *, unsigned int,
int, roundingMode);
APInt convertHalfAPFloatToAPInt() const;
APInt convertFloatAPFloatToAPInt() const;
APInt convertDoubleAPFloatToAPInt() const;
APInt convertQuadrupleAPFloatToAPInt() const;
APInt convertF80LongDoubleAPFloatToAPInt() const;
APInt convertPPCDoubleDoubleAPFloatToAPInt() const;
void initFromAPInt(const APInt& api, bool isIEEE = false);
void initFromHalfAPInt(const APInt& api);
void initFromFloatAPInt(const APInt& api);
void initFromDoubleAPInt(const APInt& api);
void initFromQuadrupleAPInt(const APInt &api);
void initFromF80LongDoubleAPInt(const APInt& api);
void initFromPPCDoubleDoubleAPInt(const APInt& api);
void assign(const APFloat &);
void copySignificand(const APFloat &);
void freeSignificand();
/* What kind of semantics does this value obey? */
const fltSemantics *semantics;
/* Significand - the fraction with an explicit integer bit. Must be
at least one bit wider than the target precision. */
union Significand
{
integerPart part;
integerPart *parts;
} significand;
/* The exponent - a signed number. */
exponent_t exponent;
/* What kind of floating point number this is. */
/* Only 2 bits are required, but VisualStudio incorrectly sign extends
it. Using the extra bit keeps it from failing under VisualStudio */
fltCategory category: 3;
/* The sign bit of this number. */
unsigned int sign: 1;
/* For PPCDoubleDouble, we have a second exponent and sign (the second
significand is appended to the first one, although it would be wrong to
regard these as a single number for arithmetic purposes). These fields
are not meaningful for any other type. */
exponent_t exponent2 : 11;
unsigned int sign2: 1;
};
} /* namespace llvm */
#endif /* LLVM_FLOAT_H */
|