summaryrefslogtreecommitdiff
path: root/libraries/base/GHC/Float/RealFracMethods.hs
blob: 5f971ad1269f9623bbbf862435eea8766914fe3b (plain)
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
{-# LANGUAGE Trustworthy #-}
{-# LANGUAGE CPP, MagicHash, UnboxedTuples, NoImplicitPrelude #-}
{-# OPTIONS_HADDOCK not-home #-}

-----------------------------------------------------------------------------
-- |
-- Module      :  GHC.Float.RealFracMethods
-- Copyright   :  (c) Daniel Fischer 2010
-- License     :  see libraries/base/LICENSE
--
-- Maintainer  :  cvs-ghc@haskell.org
-- Stability   :  internal
-- Portability :  non-portable (GHC Extensions)
--
-- Methods for the RealFrac instances for 'Float' and 'Double',
-- with specialised versions for 'Int'.
--
-- Moved to their own module to not bloat GHC.Float further.
--
-----------------------------------------------------------------------------

#include "MachDeps.h"

module GHC.Float.RealFracMethods
    ( -- * Double methods
      -- ** Integer results
      properFractionDoubleInteger
    , truncateDoubleInteger
    , floorDoubleInteger
    , ceilingDoubleInteger
    , roundDoubleInteger
      -- ** Int results
    , properFractionDoubleInt
    , floorDoubleInt
    , ceilingDoubleInt
    , roundDoubleInt
      -- * Double/Int conversions, wrapped primops
    , double2Int
    , int2Double
      -- * Float methods
      -- ** Integer results
    , properFractionFloatInteger
    , truncateFloatInteger
    , floorFloatInteger
    , ceilingFloatInteger
    , roundFloatInteger
      -- ** Int results
    , properFractionFloatInt
    , floorFloatInt
    , ceilingFloatInt
    , roundFloatInt
      -- * Float/Int conversions, wrapped primops
    , float2Int
    , int2Float
    ) where

import GHC.Num.Integer

import GHC.Base
import GHC.Num ()

#if WORD_SIZE_IN_BITS < 64

#define TO64 integerToInt64#
#define FROM64 integerFromInt64#
#define MINUS64 subInt64#
#define NEGATE64 negateInt64#

#else

#define TO64 integerToInt#
#define FROM64 IS
#define MINUS64 ( -# )
#define NEGATE64 negateInt#

uncheckedIShiftRA64# :: Int# -> Int# -> Int#
uncheckedIShiftRA64# = uncheckedIShiftRA#

uncheckedIShiftL64# :: Int# -> Int# -> Int#
uncheckedIShiftL64# = uncheckedIShiftL#

#endif

default ()

------------------------------------------------------------------------------
--                              Float Methods                               --
------------------------------------------------------------------------------

-- Special Functions for Int, nice, easy and fast.
-- They should be small enough to be inlined automatically.

-- We have to test for ±0.0 to avoid returning -0.0 in the second
-- component of the pair. Unfortunately the branching costs a lot
-- of performance.
properFractionFloatInt :: Float -> (Int, Float)
properFractionFloatInt (F# x) =
    if isTrue# (x `eqFloat#` 0.0#)
        then (I# 0#, F# 0.0#)
        else case float2Int# x of
                n -> (I# n, F# (x `minusFloat#` int2Float# n))

-- truncateFloatInt = float2Int

floorFloatInt :: Float -> Int
floorFloatInt (F# x) =
    case float2Int# x of
      n | isTrue# (x `ltFloat#` int2Float# n) -> I# (n -# 1#)
        | otherwise                           -> I# n

ceilingFloatInt :: Float -> Int
ceilingFloatInt (F# x) =
    case float2Int# x of
      n | isTrue# (int2Float# n `ltFloat#` x) -> I# (n +# 1#)
        | otherwise                           -> I# n

roundFloatInt :: Float -> Int
roundFloatInt x = float2Int (c_rintFloat x)

-- Functions with Integer results

-- With the new code generator in GHC 7, the explicit bit-fiddling is
-- slower than the old code for values of small modulus, but when the
-- 'Int' range is left, the bit-fiddling quickly wins big, so we use that.
-- If the methods are called on smallish values, hopefully people go
-- through Int and not larger types.

-- Note: For negative exponents, we must check the validity of the shift
-- distance for the right shifts of the mantissa.

{-# INLINE properFractionFloatInteger #-}
properFractionFloatInteger :: Float -> (Integer, Float)
properFractionFloatInteger v@(F# x) =
    case decodeFloat_Int# x of
      (# m, e #)
        | isTrue# (e <# 0#) ->
          case negateInt# e of
            s | isTrue# (s ># 23#) -> (0, v)
              | isTrue# (m <#  0#) ->
                case negateInt# (negateInt# m `uncheckedIShiftRA#` s) of
                  k -> (IS k,
                            case m -# (k `uncheckedIShiftL#` s) of
                              r -> F# (integerEncodeFloat# (IS r) e))
              | otherwise           ->
                case m `uncheckedIShiftRL#` s of
                  k -> (IS k,
                            case m -# (k `uncheckedIShiftL#` s) of
                              r -> F# (integerEncodeFloat# (IS r) e))
        | otherwise -> (integerShiftL# (IS m) (int2Word# e), F# 0.0#)

{-# INLINE truncateFloatInteger #-}
truncateFloatInteger :: Float -> Integer
truncateFloatInteger x =
    case properFractionFloatInteger x of
      (n, _) -> n

-- floor is easier for negative numbers than truncate, so this gets its
-- own implementation, it's a little faster.
{-# INLINE floorFloatInteger #-}
floorFloatInteger :: Float -> Integer
floorFloatInteger (F# x) =
    case decodeFloat_Int# x of
      (# m, e #)
        | isTrue# (e <# 0#) ->
          case negateInt# e of
            s | isTrue# (s ># 23#) -> if isTrue# (m <# 0#) then (-1) else 0
              | otherwise          -> IS (m `uncheckedIShiftRA#` s)
        | otherwise -> integerShiftL# (IS m) (int2Word# e)

-- ceiling x = -floor (-x)
-- If giving this its own implementation is faster at all,
-- it's only marginally so, hence we keep it short.
{-# INLINE ceilingFloatInteger #-}
ceilingFloatInteger :: Float -> Integer
ceilingFloatInteger (F# x) =
    integerNegate (floorFloatInteger (F# (negateFloat# x)))

{-# INLINE roundFloatInteger #-}
roundFloatInteger :: Float -> Integer
roundFloatInteger x = float2Integer (c_rintFloat x)

------------------------------------------------------------------------------
--                              Double Methods                              --
------------------------------------------------------------------------------

-- Special Functions for Int, nice, easy and fast.
-- They should be small enough to be inlined automatically.

-- We have to test for ±0.0 to avoid returning -0.0 in the second
-- component of the pair. Unfortunately the branching costs a lot
-- of performance.
properFractionDoubleInt :: Double -> (Int, Double)
properFractionDoubleInt (D# x) =
    if isTrue# (x ==## 0.0##)
        then (I# 0#, D# 0.0##)
        else case double2Int# x of
                n -> (I# n, D# (x -## int2Double# n))

-- truncateDoubleInt = double2Int

floorDoubleInt :: Double -> Int
floorDoubleInt (D# x) =
    case double2Int# x of
      n | isTrue# (x <## int2Double# n) -> I# (n -# 1#)
        | otherwise                     -> I# n

ceilingDoubleInt :: Double -> Int
ceilingDoubleInt (D# x) =
    case double2Int# x of
      n | isTrue# (int2Double# n <## x) -> I# (n +# 1#)
        | otherwise                     -> I# n

roundDoubleInt :: Double -> Int
roundDoubleInt x = double2Int (c_rintDouble x)

-- Functions with Integer results

-- The new Code generator isn't quite as good for the old 'Double' code
-- as for the 'Float' code, so for 'Double' the bit-fiddling also wins
-- when the values have small modulus.

-- When the exponent is negative, all mantissae have less than 64 bits
-- and the right shifting of sized types is much faster than that of
-- 'Integer's, especially when we can

-- Note: For negative exponents, we must check the validity of the shift
-- distance for the right shifts of the mantissa.

{-# INLINE properFractionDoubleInteger #-}
properFractionDoubleInteger :: Double -> (Integer, Double)
properFractionDoubleInteger v@(D# x) =
    case integerDecodeDouble# x of
      (# m, e #)
        | isTrue# (e <# 0#) ->
          case negateInt# e of
            s | isTrue# (s ># 52#) -> (0, v)
              | m < 0                 ->
                case TO64 (integerNegate m) of
                  n ->
                    case n `uncheckedIShiftRA64#` s of
                      k ->
                        (FROM64 (NEGATE64 k),
                          case MINUS64 n (k `uncheckedIShiftL64#` s) of
                            r ->
                              D# (integerEncodeDouble# (FROM64 (NEGATE64 r)) e))
              | otherwise           ->
                case TO64 m of
                  n ->
                    case n `uncheckedIShiftRA64#` s of
                      k -> (FROM64 k,
                            case MINUS64 n (k `uncheckedIShiftL64#` s) of
                              r -> D# (integerEncodeDouble# (FROM64 r) e))
        | otherwise -> (integerShiftL# m (int2Word# e), D# 0.0##)

{-# INLINE truncateDoubleInteger #-}
truncateDoubleInteger :: Double -> Integer
truncateDoubleInteger x =
    case properFractionDoubleInteger x of
      (n, _) -> n

-- floor is easier for negative numbers than truncate, so this gets its
-- own implementation, it's a little faster.
{-# INLINE floorDoubleInteger #-}
floorDoubleInteger :: Double -> Integer
floorDoubleInteger (D# x) =
    case integerDecodeDouble# x of
      (# m, e #)
        | isTrue# (e <# 0#) ->
          case negateInt# e of
            s | isTrue# (s ># 52#) -> if m < 0 then (-1) else 0
              | otherwise          ->
                case TO64 m of
                  n -> FROM64 (n `uncheckedIShiftRA64#` s)
        | otherwise -> integerShiftL# m (int2Word# e)

{-# INLINE ceilingDoubleInteger #-}
ceilingDoubleInteger :: Double -> Integer
ceilingDoubleInteger (D# x) =
    integerNegate (floorDoubleInteger (D# (negateDouble# x)))

{-# INLINE roundDoubleInteger #-}
roundDoubleInteger :: Double -> Integer
roundDoubleInteger x = double2Integer (c_rintDouble x)

-- Wrappers around double2Int#, int2Double#, float2Int# and int2Float#,
-- we need them here, so we move them from GHC.Float and re-export them
-- explicitly from there.

double2Int :: Double -> Int
double2Int (D# x) = I# (double2Int# x)

int2Double :: Int -> Double
int2Double (I# i) = D# (int2Double# i)

float2Int :: Float -> Int
float2Int (F# x) = I# (float2Int# x)

int2Float :: Int -> Float
int2Float (I# i) = F# (int2Float# i)

-- Quicker conversions from 'Double' and 'Float' to 'Integer',
-- assuming the floating point value is integral.
--
-- Note: Since the value is integral, the exponent can't be less than
-- (-TYP_MANT_DIG), so we need not check the validity of the shift
-- distance for the right shifts here.

{-# INLINE double2Integer #-}
double2Integer :: Double -> Integer
double2Integer (D# x) =
    case integerDecodeDouble# x of
      (# m, e #)
        | isTrue# (e <# 0#) ->
          case TO64 m of
            n -> FROM64 (n `uncheckedIShiftRA64#` negateInt# e)
        | otherwise -> integerShiftL# m (int2Word# e)

{-# INLINE float2Integer #-}
float2Integer :: Float -> Integer
float2Integer (F# x) =
    case decodeFloat_Int# x of
      (# m, e #)
        | isTrue# (e <# 0#) -> IS (m `uncheckedIShiftRA#` negateInt# e)
        | otherwise         -> integerShiftL# (IS m) (int2Word# e)

-- Foreign imports, the rounding is done faster in C when the value
-- isn't integral, so we call out for rounding. For values of large
-- modulus, calling out to C is slower than staying in Haskell, but
-- presumably 'round' is mostly called for values with smaller modulus,
-- when calling out to C is a major win.
-- For all other functions, calling out to C gives at most a marginal
-- speedup for values of small modulus and is much slower than staying
-- in Haskell for values of large modulus, so those are done in Haskell.

foreign import ccall unsafe "rintDouble"
    c_rintDouble :: Double -> Double

foreign import ccall unsafe "rintFloat"
    c_rintFloat :: Float -> Float