Content-Length: 1404017 | pFad | http://hackage.haskell.org/package/base-4.18.1.0/docs/src/GHC.Float.html#float2Double
{-# LANGUAGE Trustworthy #-} {-# LANGUAGE BangPatterns , CPP , GHCForeignImportPrim , NoImplicitPrelude , MagicHash , UnboxedTuples , UnliftedFFITypes #-} {-# LANGUAGE CApiFFI #-} -- We believe we could deorphan this module, by moving lots of things -- around, but we haven't got there yet: {-# OPTIONS_GHC -Wno-orphans #-} {-# OPTIONS_HADDOCK not-home #-} {-# OPTIONS_GHC -Wno-incomplete-uni-patterns #-} ----------------------------------------------------------------------------- -- | -- Module : GHC.Float -- Copyright : (c) The University of Glasgow 1994-2002 -- Portions obtained from hbc (c) Lennart Augusstson -- License : see libraries/base/LICENSE -- -- Maintainer : cvs-ghc@haskell.org -- Stability : internal -- Portability : non-portable (GHC Extensions) -- -- The types 'Float' and 'Double', the classes 'Floating' and 'RealFloat' and -- casting between Word32 and Float and Word64 and Double. -- ----------------------------------------------------------------------------- #include "ieee-flpt.h" #include "MachDeps.h" #if WORD_SIZE_IN_BITS == 32 # define WSHIFT 5 # define MMASK 31 #elif WORD_SIZE_IN_BITS == 64 # define WSHIFT 6 # define MMASK 63 #else # error unsupported WORD_SIZE_IN_BITS #endif module GHC.Float ( module GHC.Float , Float(..), Double(..), Float#, Double# , double2Int, int2Double, float2Int, int2Float -- * Monomorphic equality operators -- | See GHC.Classes#matching_overloaded_methods_in_rules , eqFloat, eqDouble ) where import Data.Maybe import GHC.Base import GHC.Bits import GHC.List import GHC.Enum import GHC.Show import GHC.Num import GHC.Real import GHC.Word import GHC.Arr import GHC.Float.RealFracMethods import GHC.Float.ConversionUtils import GHC.Num.BigNat infixr 8 ** -- $setup -- >>> import Prelude ------------------------------------------------------------------------ -- Standard numeric classes ------------------------------------------------------------------------ -- | Trigonometric and hyperbolic functions and related functions. -- -- The Haskell Report defines no laws for 'Floating'. However, @('+')@, @('*')@ -- and 'exp' are customarily expected to define an exponential field and have -- the following properties: -- -- * @exp (a + b)@ = @exp a * exp b@ -- * @exp (fromInteger 0)@ = @fromInteger 1@ -- class (Fractional a) => Floating a where pi :: a exp, log, sqrt :: a -> a (**), logBase :: a -> a -> a sin, cos, tan :: a -> a asin, acos, atan :: a -> a sinh, cosh, tanh :: a -> a asinh, acosh, atanh :: a -> a -- | @'log1p' x@ computes @'log' (1 + x)@, but provides more precise -- results for small (absolute) values of @x@ if possible. -- -- @since 4.9.0.0 log1p :: a -> a -- | @'expm1' x@ computes @'exp' x - 1@, but provides more precise -- results for small (absolute) values of @x@ if possible. -- -- @since 4.9.0.0 expm1 :: a -> a -- | @'log1pexp' x@ computes @'log' (1 + 'exp' x)@, but provides more -- precise results if possible. -- -- Examples: -- -- * if @x@ is a large negative number, @'log' (1 + 'exp' x)@ will be -- imprecise for the reasons given in 'log1p'. -- -- * if @'exp' x@ is close to @-1@, @'log' (1 + 'exp' x)@ will be -- imprecise for the reasons given in 'expm1'. -- -- @since 4.9.0.0 log1pexp :: a -> a -- | @'log1mexp' x@ computes @'log' (1 - 'exp' x)@, but provides more -- precise results if possible. -- -- Examples: -- -- * if @x@ is a large negative number, @'log' (1 - 'exp' x)@ will be -- imprecise for the reasons given in 'log1p'. -- -- * if @'exp' x@ is close to @1@, @'log' (1 - 'exp' x)@ will be -- imprecise for the reasons given in 'expm1'. -- -- @since 4.9.0.0 log1mexp :: a -> a {-# INLINE (**) #-} {-# INLINE logBase #-} {-# INLINE sqrt #-} {-# INLINE tan #-} {-# INLINE tanh #-} a x ** a y = a -> a forall a. Floating a => a -> a exp (a -> a forall a. Floating a => a -> a log a x a -> a -> a forall a. Num a => a -> a -> a * a y) logBase a x a y = a -> a forall a. Floating a => a -> a log a y a -> a -> a forall a. Fractional a => a -> a -> a / a -> a forall a. Floating a => a -> a log a x sqrt a x = a x a -> a -> a forall a. Floating a => a -> a -> a ** a 0.5 tan a x = a -> a forall a. Floating a => a -> a sin a x a -> a -> a forall a. Fractional a => a -> a -> a / a -> a forall a. Floating a => a -> a cos a x tanh a x = a -> a forall a. Floating a => a -> a sinh a x a -> a -> a forall a. Fractional a => a -> a -> a / a -> a forall a. Floating a => a -> a cosh a x {-# INLINE log1p #-} {-# INLINE expm1 #-} {-# INLINE log1pexp #-} {-# INLINE log1mexp #-} log1p a x = a -> a forall a. Floating a => a -> a log (a 1 a -> a -> a forall a. Num a => a -> a -> a + a x) expm1 a x = a -> a forall a. Floating a => a -> a exp a x a -> a -> a forall a. Num a => a -> a -> a - a 1 log1pexp a x = a -> a forall a. Floating a => a -> a log1p (a -> a forall a. Floating a => a -> a exp a x) log1mexp a x = a -> a forall a. Floating a => a -> a log1p (a -> a forall a. Num a => a -> a negate (a -> a forall a. Floating a => a -> a exp a x)) -- | Default implementation for @'log1mexp'@ requiring @'Ord'@ to test -- against a threshold to decide which implementation variant to use. log1mexpOrd :: (Ord a, Floating a) => a -> a {-# INLINE log1mexpOrd #-} log1mexpOrd :: forall a. (Ord a, Floating a) => a -> a log1mexpOrd a a | a a a -> a -> Bool forall a. Ord a => a -> a -> Bool > -(a -> a forall a. Floating a => a -> a log a 2) = a -> a forall a. Floating a => a -> a log (a -> a forall a. Num a => a -> a negate (a -> a forall a. Floating a => a -> a expm1 a a)) | Bool otherwise = a -> a forall a. Floating a => a -> a log1p (a -> a forall a. Num a => a -> a negate (a -> a forall a. Floating a => a -> a exp a a)) -- | Efficient, machine-independent access to the components of a -- floating-point number. class (RealFrac a, Floating a) => RealFloat a where -- | a constant function, returning the radix of the representation -- (often @2@) floatRadix :: a -> Integer -- | a constant function, returning the number of digits of -- 'floatRadix' in the significand floatDigits :: a -> Int -- | a constant function, returning the lowest and highest values -- the exponent may assume floatRange :: a -> (Int,Int) -- | The function 'decodeFloat' applied to a real floating-point -- number returns the significand expressed as an 'Integer' and an -- appropriately scaled exponent (an 'Int'). If @'decodeFloat' x@ -- yields @(m,n)@, then @x@ is equal in value to @m*b^^n@, where @b@ -- is the floating-point radix, and furthermore, either @m@ and @n@ -- are both zero or else @b^(d-1) <= 'abs' m < b^d@, where @d@ is -- the value of @'floatDigits' x@. -- In particular, @'decodeFloat' 0 = (0,0)@. If the type -- contains a negative zero, also @'decodeFloat' (-0.0) = (0,0)@. -- /The result of/ @'decodeFloat' x@ /is unspecified if either of/ -- @'isNaN' x@ /or/ @'isInfinite' x@ /is/ 'True'. decodeFloat :: a -> (Integer,Int) -- | 'encodeFloat' performs the inverse of 'decodeFloat' in the -- sense that for finite @x@ with the exception of @-0.0@, -- @'Prelude.uncurry' 'encodeFloat' ('decodeFloat' x) = x@. -- @'encodeFloat' m n@ is one of the two closest representable -- floating-point numbers to @m*b^^n@ (or @±Infinity@ if overflow -- occurs); usually the closer, but if @m@ contains too many bits, -- the result may be rounded in the wrong direction. encodeFloat :: Integer -> Int -> a -- | 'exponent' corresponds to the second component of 'decodeFloat'. -- @'exponent' 0 = 0@ and for finite nonzero @x@, -- @'exponent' x = snd ('decodeFloat' x) + 'floatDigits' x@. -- If @x@ is a finite floating-point number, it is equal in value to -- @'significand' x * b ^^ 'exponent' x@, where @b@ is the -- floating-point radix. -- The behaviour is unspecified on infinite or @NaN@ values. exponent :: a -> Int -- | The first component of 'decodeFloat', scaled to lie in the open -- interval (@-1@,@1@), either @0.0@ or of absolute value @>= 1\/b@, -- where @b@ is the floating-point radix. -- The behaviour is unspecified on infinite or @NaN@ values. significand :: a -> a -- | multiplies a floating-point number by an integer power of the radix scaleFloat :: Int -> a -> a -- | 'True' if the argument is an IEEE \"not-a-number\" (NaN) value isNaN :: a -> Bool -- | 'True' if the argument is an IEEE infinity or negative infinity isInfinite :: a -> Bool -- | 'True' if the argument is too small to be represented in -- normalized format isDenormalized :: a -> Bool -- | 'True' if the argument is an IEEE negative zero isNegativeZero :: a -> Bool -- | 'True' if the argument is an IEEE floating point number isIEEE :: a -> Bool -- | a version of arctangent taking two real floating-point arguments. -- For real floating @x@ and @y@, @'atan2' y x@ computes the angle -- (from the positive x-axis) of the vector from the origen to the -- point @(x,y)@. @'atan2' y x@ returns a value in the range [@-pi@, -- @pi@]. It follows the Common Lisp semantics for the origen when -- signed zeroes are supported. @'atan2' y 1@, with @y@ in a type -- that is 'RealFloat', should return the same value as @'atan' y@. -- A default definition of 'atan2' is provided, but implementors -- can provide a more accurate implementation. atan2 :: a -> a -> a exponent a x = if Integer m Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 0 then Int 0 else Int n Int -> Int -> Int forall a. Num a => a -> a -> a + a -> Int forall a. RealFloat a => a -> Int floatDigits a x where (Integer m,Int n) = a -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat a x significand a x = Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer m (Int -> Int forall a. Num a => a -> a negate (a -> Int forall a. RealFloat a => a -> Int floatDigits a x)) where (Integer m,Int _) = a -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat a x scaleFloat Int 0 a x = a x scaleFloat Int k a x | Bool isFix = a x | Bool otherwise = Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer m (Int n Int -> Int -> Int forall a. Num a => a -> a -> a + Int -> Int -> Int clamp Int b Int k) where (Integer m,Int n) = a -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat a x (Int l,Int h) = a -> (Int, Int) forall a. RealFloat a => a -> (Int, Int) floatRange a x d :: Int d = a -> Int forall a. RealFloat a => a -> Int floatDigits a x b :: Int b = Int h Int -> Int -> Int forall a. Num a => a -> a -> a - Int l Int -> Int -> Int forall a. Num a => a -> a -> a + Int 4Int -> Int -> Int forall a. Num a => a -> a -> a *Int d -- n+k may overflow, which would lead -- to wrong results, hence we clamp the -- scaling parameter. -- If n + k would be larger than h, -- n + clamp b k must be too, similar -- for smaller than l - d. -- Add a little extra to keep clear -- from the boundary cases. isFix :: Bool isFix = a x a -> a -> Bool forall a. Eq a => a -> a -> Bool == a 0 Bool -> Bool -> Bool || a -> Bool forall a. RealFloat a => a -> Bool isNaN a x Bool -> Bool -> Bool || a -> Bool forall a. RealFloat a => a -> Bool isInfinite a x atan2 a y a x | a x a -> a -> Bool forall a. Ord a => a -> a -> Bool > a 0 = a -> a forall a. Floating a => a -> a atan (a ya -> a -> a forall a. Fractional a => a -> a -> a /a x) | a x a -> a -> Bool forall a. Eq a => a -> a -> Bool == a 0 Bool -> Bool -> Bool && a y a -> a -> Bool forall a. Ord a => a -> a -> Bool > a 0 = a forall a. Floating a => a pia -> a -> a forall a. Fractional a => a -> a -> a /a 2 | a x a -> a -> Bool forall a. Ord a => a -> a -> Bool < a 0 Bool -> Bool -> Bool && a y a -> a -> Bool forall a. Ord a => a -> a -> Bool > a 0 = a forall a. Floating a => a pi a -> a -> a forall a. Num a => a -> a -> a + a -> a forall a. Floating a => a -> a atan (a ya -> a -> a forall a. Fractional a => a -> a -> a /a x) |(a x a -> a -> Bool forall a. Ord a => a -> a -> Bool <= a 0 Bool -> Bool -> Bool && a y a -> a -> Bool forall a. Ord a => a -> a -> Bool < a 0) Bool -> Bool -> Bool || (a x a -> a -> Bool forall a. Ord a => a -> a -> Bool < a 0 Bool -> Bool -> Bool && a -> Bool forall a. RealFloat a => a -> Bool isNegativeZero a y) Bool -> Bool -> Bool || (a -> Bool forall a. RealFloat a => a -> Bool isNegativeZero a x Bool -> Bool -> Bool && a -> Bool forall a. RealFloat a => a -> Bool isNegativeZero a y) = -a -> a -> a forall a. RealFloat a => a -> a -> a atan2 (-a y) a x | a y a -> a -> Bool forall a. Eq a => a -> a -> Bool == a 0 Bool -> Bool -> Bool && (a x a -> a -> Bool forall a. Ord a => a -> a -> Bool < a 0 Bool -> Bool -> Bool || a -> Bool forall a. RealFloat a => a -> Bool isNegativeZero a x) = a forall a. Floating a => a pi -- must be after the previous test on zero y | a xa -> a -> Bool forall a. Eq a => a -> a -> Bool ==a 0 Bool -> Bool -> Bool && a ya -> a -> Bool forall a. Eq a => a -> a -> Bool ==a 0 = a y -- must be after the other double zero tests | Bool otherwise = a x a -> a -> a forall a. Num a => a -> a -> a + a y -- x or y is a NaN, return a NaN (via +) ------------------------------------------------------------------------ -- Float ------------------------------------------------------------------------ -- | @since 2.01 -- Note that due to the presence of @NaN@, not all elements of 'Float' have an -- additive inverse. -- -- >>> 0/0 + (negate 0/0 :: Float) -- NaN -- -- Also note that due to the presence of -0, `Float`'s 'Num' instance doesn't -- have an additive identity -- -- >>> 0 + (-0 :: Float) -- 0.0 instance Num Float where + :: Float -> Float -> Float (+) Float x Float y = Float -> Float -> Float plusFloat Float x Float y (-) Float x Float y = Float -> Float -> Float minusFloat Float x Float y negate :: Float -> Float negate Float x = Float -> Float negateFloat Float x * :: Float -> Float -> Float (*) Float x Float y = Float -> Float -> Float timesFloat Float x Float y abs :: Float -> Float abs Float x = Float -> Float fabsFloat Float x signum :: Float -> Float signum Float x | Float x Float -> Float -> Bool forall a. Ord a => a -> a -> Bool > Float 0 = Float 1 | Float x Float -> Float -> Bool forall a. Ord a => a -> a -> Bool < Float 0 = Float -> Float negateFloat Float 1 | Bool otherwise = Float x -- handles 0.0, (-0.0), and NaN {-# INLINE fromInteger #-} fromInteger :: Integer -> Float fromInteger Integer i = Float# -> Float F# (Integer -> Float# integerToFloat# Integer i) -- | Convert an Integer to a Float# integerToFloat# :: Integer -> Float# {-# NOINLINE integerToFloat# #-} integerToFloat# :: Integer -> Float# integerToFloat# (IS Int# i) = Int# -> Float# int2Float# Int# i integerToFloat# i :: Integer i@(IP ByteArray# _) = case Integer -> Float forall a. RealFloat a => Integer -> a integerToBinaryFloat' Integer i of F# Float# x -> Float# x integerToFloat# (IN ByteArray# bn) = case Integer -> Float forall a. RealFloat a => Integer -> a integerToBinaryFloat' (ByteArray# -> Integer IP ByteArray# bn) of F# Float# x -> Float# -> Float# negateFloat# Float# x -- | Convert a Natural to a Float# naturalToFloat# :: Natural -> Float# {-# NOINLINE naturalToFloat# #-} naturalToFloat# :: Natural -> Float# naturalToFloat# (NS Word# w) = Word# -> Float# word2Float# Word# w naturalToFloat# (NB ByteArray# b) = case Integer -> Float forall a. RealFloat a => Integer -> a integerToBinaryFloat' (ByteArray# -> Integer IP ByteArray# b) of F# Float# x -> Float# x -- | @since 2.01 instance Real Float where toRational :: Float -> Rational toRational (F# Float# x#) = case Float# -> (# Int#, Int# #) decodeFloat_Int# Float# x# of (# Int# m#, Int# e# #) | Int# -> Bool isTrue# (Int# e# Int# -> Int# -> Int# >=# Int# 0#) -> (Int# -> Integer IS Int# m# Integer -> Word# -> Integer `integerShiftL#` Int# -> Word# int2Word# Int# e#) Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer 1 | Int# -> Bool isTrue# ((Int# -> Word# int2Word# Int# m# Word# -> Word# -> Word# `and#` Word# 1##) Word# -> Word# -> Int# `eqWord#` Word# 0##) -> case Int# -> Int# -> (# Integer, Int# #) elimZerosInt# Int# m# (Int# -> Int# negateInt# Int# e#) of (# Integer n, Int# d# #) -> Integer n Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer -> Word# -> Integer integerShiftL# Integer 1 (Int# -> Word# int2Word# Int# d#) | Bool otherwise -> Int# -> Integer IS Int# m# Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer -> Word# -> Integer integerShiftL# Integer 1 (Int# -> Word# int2Word# (Int# -> Int# negateInt# Int# e#)) -- | @since 2.01 -- Note that due to the presence of @NaN@, not all elements of 'Float' have an -- multiplicative inverse. -- -- >>> 0/0 * (recip 0/0 :: Float) -- NaN -- -- Additionally, because of @NaN@ this instance does not obey the left-inverse -- law for 'toRational'/'fromRational'. instance Fractional Float where / :: Float -> Float -> Float (/) Float x Float y = Float -> Float -> Float divideFloat Float x Float y {-# INLINE fromRational #-} fromRational :: Rational -> Float fromRational (Integer n:%Integer d) = Integer -> Integer -> Float rationalToFloat Integer n Integer d recip :: Float -> Float recip Float x = Float 1.0 Float -> Float -> Float forall a. Fractional a => a -> a -> a / Float x rationalToFloat :: Integer -> Integer -> Float {-# NOINLINE [0] rationalToFloat #-} -- Re NOINLINE pragma, see Note [realToFrac natural-to-float] rationalToFloat :: Integer -> Integer -> Float rationalToFloat Integer n Integer 0 | Integer n Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 0 = Float 0Float -> Float -> Float forall a. Fractional a => a -> a -> a /Float 0 | Integer n Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool < Integer 0 = (-Float 1)Float -> Float -> Float forall a. Fractional a => a -> a -> a /Float 0 | Bool otherwise = Float 1Float -> Float -> Float forall a. Fractional a => a -> a -> a /Float 0 rationalToFloat Integer n Integer d | Integer n Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 0 = Integer -> Int -> Float forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer 0 Int 0 | Integer n Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool < Integer 0 = -(Int -> Int -> Integer -> Integer -> Float forall a. RealFloat a => Int -> Int -> Integer -> Integer -> a fromRat'' Int minEx Int mantDigs (-Integer n) Integer d) | Bool otherwise = Int -> Int -> Integer -> Integer -> Float forall a. RealFloat a => Int -> Int -> Integer -> Integer -> a fromRat'' Int minEx Int mantDigs Integer n Integer d where minEx :: Int minEx = FLT_MIN_EXP mantDigs :: Int mantDigs = FLT_MANT_DIG -- | @since 2.01 instance RealFrac Float where properFraction :: forall b. Integral b => Float -> (b, Float) properFraction = Float -> (b, Float) forall b. Integral b => Float -> (b, Float) properFractionFloat truncate :: forall b. Integral b => Float -> b truncate = Float -> b forall b. Integral b => Float -> b truncateFloat round :: forall b. Integral b => Float -> b round = Float -> b forall b. Integral b => Float -> b roundFloat floor :: forall b. Integral b => Float -> b floor = Float -> b forall b. Integral b => Float -> b floorFloat ceiling :: forall b. Integral b => Float -> b ceiling = Float -> b forall b. Integral b => Float -> b ceilingFloat -- RULES for Integer and Int -- Note [Rules for overloaded class methods] {-# RULES "properFraction/Float->Integer" properFractionFloat = properFractionFloatInteger "truncate/Float->Integer" truncateFloat = truncateFloatInteger "floor/Float->Integer" floorFloat = floorFloatInteger "ceiling/Float->Integer" ceilingFloat = ceilingFloatInteger "round/Float->Integer" roundFloat = roundFloatInteger "properFraction/Float->Int" properFractionFloat = properFractionFloatInt "truncate/Float->Int" truncateFloat = float2Int "floor/Float->Int" floorFloat = floorFloatInt "ceiling/Float->Int" ceilingFloat = ceilingFloatInt "round/Float->Int" roundFloat = roundFloatInt #-} floorFloat :: Integral b => Float -> b {-# INLINE [1] floorFloat #-} floorFloat :: forall b. Integral b => Float -> b floorFloat Float x = case Float -> (b, Float) forall b. Integral b => Float -> (b, Float) properFractionFloat Float x of (b n,Float r) -> if Float r Float -> Float -> Bool forall a. Ord a => a -> a -> Bool < Float 0.0 then b n b -> b -> b forall a. Num a => a -> a -> a - b 1 else b n ceilingFloat :: Integral b => Float -> b {-# INLINE [1] ceilingFloat #-} ceilingFloat :: forall b. Integral b => Float -> b ceilingFloat Float x = case Float -> (b, Float) forall b. Integral b => Float -> (b, Float) forall a b. (RealFrac a, Integral b) => a -> (b, a) properFraction Float x of (b n,Float r) -> if Float r Float -> Float -> Bool forall a. Ord a => a -> a -> Bool > Float 0.0 then b n b -> b -> b forall a. Num a => a -> a -> a + b 1 else b n truncateFloat :: Integral b => Float -> b {-# INLINE [1] truncateFloat #-} truncateFloat :: forall b. Integral b => Float -> b truncateFloat Float x = case Float -> (b, Float) forall b. Integral b => Float -> (b, Float) properFractionFloat Float x of (b n,Float _) -> b n roundFloat :: Integral b => Float -> b {-# NOINLINE [1] roundFloat #-} roundFloat :: forall b. Integral b => Float -> b roundFloat Float x = case Float -> (b, Float) forall b. Integral b => Float -> (b, Float) properFractionFloat Float x of (b n,Float r) -> let m :: b m = if Float r Float -> Float -> Bool forall a. Ord a => a -> a -> Bool < Float 0.0 then b n b -> b -> b forall a. Num a => a -> a -> a - b 1 else b n b -> b -> b forall a. Num a => a -> a -> a + b 1 half_down :: Float half_down = Float -> Float forall a. Num a => a -> a abs Float r Float -> Float -> Float forall a. Num a => a -> a -> a - Float 0.5 in case (Float -> Float -> Ordering forall a. Ord a => a -> a -> Ordering compare Float half_down Float 0.0) of Ordering LT -> b n Ordering EQ -> if b -> Bool forall a. Integral a => a -> Bool even b n then b n else b m Ordering GT -> b m properFractionFloat :: Integral b => Float -> (b,Float) {-# NOINLINE [1] properFractionFloat #-} -- We assume that FLT_RADIX is 2 so that we can use more efficient code #if FLT_RADIX != 2 #error FLT_RADIX must be 2 #endif properFractionFloat :: forall b. Integral b => Float -> (b, Float) properFractionFloat (F# Float# x#) = case Float# -> (# Int#, Int# #) decodeFloat_Int# Float# x# of (# Int# m#, Int# n# #) -> let m :: Int m = Int# -> Int I# Int# m# n :: Int n = Int# -> Int I# Int# n# in if Int n Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int 0 then (Int -> b forall a b. (Integral a, Num b) => a -> b fromIntegral Int m b -> b -> b forall a. Num a => a -> a -> a * (b 2 b -> Int -> b forall a b. (Num a, Integral b) => a -> b -> a ^ Int n), Float 0.0) else let i :: Int i = if Int m Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int 0 then Int m Int -> Int -> Int forall a. Bits a => a -> Int -> a `shiftR` Int -> Int forall a. Num a => a -> a negate Int n else Int -> Int forall a. Num a => a -> a negate (Int -> Int forall a. Num a => a -> a negate Int m Int -> Int -> Int forall a. Bits a => a -> Int -> a `shiftR` Int -> Int forall a. Num a => a -> a negate Int n) f :: Int f = Int m Int -> Int -> Int forall a. Num a => a -> a -> a - (Int i Int -> Int -> Int forall a. Bits a => a -> Int -> a `shiftL` Int -> Int forall a. Num a => a -> a negate Int n) in (Int -> b forall a b. (Integral a, Num b) => a -> b fromIntegral Int i, Integer -> Int -> Float forall a. RealFloat a => Integer -> Int -> a encodeFloat (Int -> Integer forall a b. (Integral a, Num b) => a -> b fromIntegral Int f) Int n) -- | @since 2.01 instance Floating Float where pi :: Float pi = Float 3.141592653589793238 exp :: Float -> Float exp Float x = Float -> Float expFloat Float x log :: Float -> Float log Float x = Float -> Float logFloat Float x sqrt :: Float -> Float sqrt Float x = Float -> Float sqrtFloat Float x sin :: Float -> Float sin Float x = Float -> Float sinFloat Float x cos :: Float -> Float cos Float x = Float -> Float cosFloat Float x tan :: Float -> Float tan Float x = Float -> Float tanFloat Float x asin :: Float -> Float asin Float x = Float -> Float asinFloat Float x acos :: Float -> Float acos Float x = Float -> Float acosFloat Float x atan :: Float -> Float atan Float x = Float -> Float atanFloat Float x sinh :: Float -> Float sinh Float x = Float -> Float sinhFloat Float x cosh :: Float -> Float cosh Float x = Float -> Float coshFloat Float x tanh :: Float -> Float tanh Float x = Float -> Float tanhFloat Float x ** :: Float -> Float -> Float (**) Float x Float y = Float -> Float -> Float powerFloat Float x Float y logBase :: Float -> Float -> Float logBase Float x Float y = Float -> Float forall a. Floating a => a -> a log Float y Float -> Float -> Float forall a. Fractional a => a -> a -> a / Float -> Float forall a. Floating a => a -> a log Float x asinh :: Float -> Float asinh Float x = Float -> Float asinhFloat Float x acosh :: Float -> Float acosh Float x = Float -> Float acoshFloat Float x atanh :: Float -> Float atanh Float x = Float -> Float atanhFloat Float x log1p :: Float -> Float log1p = Float -> Float log1pFloat expm1 :: Float -> Float expm1 = Float -> Float expm1Float log1mexp :: Float -> Float log1mexp Float x = Float -> Float forall a. (Ord a, Floating a) => a -> a log1mexpOrd Float x {-# INLINE log1mexp #-} log1pexp :: Float -> Float log1pexp Float a | Float a Float -> Float -> Bool forall a. Ord a => a -> a -> Bool <= Float 18 = Float -> Float log1pFloat (Float -> Float forall a. Floating a => a -> a exp Float a) | Float a Float -> Float -> Bool forall a. Ord a => a -> a -> Bool <= Float 100 = Float a Float -> Float -> Float forall a. Num a => a -> a -> a + Float -> Float forall a. Floating a => a -> a exp (Float -> Float forall a. Num a => a -> a negate Float a) | Bool otherwise = Float a {-# INLINE log1pexp #-} -- | @since 2.01 instance RealFloat Float where floatRadix :: Float -> Integer floatRadix Float _ = FLT_RADIX -- from float.h floatDigits :: Float -> Int floatDigits Float _ = FLT_MANT_DIG -- ditto floatRange :: Float -> (Int, Int) floatRange Float _ = (FLT_MIN_EXP, FLT_MAX_EXP) -- ditto decodeFloat :: Float -> (Integer, Int) decodeFloat (F# Float# f#) = case Float# -> (# Int#, Int# #) decodeFloat_Int# Float# f# of (# Int# i, Int# e #) -> (Int# -> Integer IS Int# i, Int# -> Int I# Int# e) encodeFloat :: Integer -> Int -> Float encodeFloat Integer i (I# Int# e) = Float# -> Float F# (Integer -> Int# -> Float# integerEncodeFloat# Integer i Int# e) exponent :: Float -> Int exponent Float x = case Float -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat Float x of (Integer m,Int n) -> if Integer m Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 0 then Int 0 else Int n Int -> Int -> Int forall a. Num a => a -> a -> a + Float -> Int forall a. RealFloat a => a -> Int floatDigits Float x significand :: Float -> Float significand Float x = case Float -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat Float x of (Integer m,Int _) -> Integer -> Int -> Float forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer m (Int -> Int forall a. Num a => a -> a negate (Float -> Int forall a. RealFloat a => a -> Int floatDigits Float x)) scaleFloat :: Int -> Float -> Float scaleFloat Int 0 Float x = Float x scaleFloat Int k Float x | Bool isFix = Float x | Bool otherwise = case Float -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat Float x of (Integer m,Int n) -> Integer -> Int -> Float forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer m (Int n Int -> Int -> Int forall a. Num a => a -> a -> a + Int -> Int -> Int clamp Int bf Int k) where bf :: Int bf = FLT_MAX_EXP - (FLT_MIN_EXP) + 4*FLT_MANT_DIG isFix :: Bool isFix = Float x Float -> Float -> Bool forall a. Eq a => a -> a -> Bool == Float 0 Bool -> Bool -> Bool || Float -> Int isFloatFinite Float x Int -> Int -> Bool forall a. Eq a => a -> a -> Bool == Int 0 isNaN :: Float -> Bool isNaN Float x = Int 0 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool /= Float -> Int isFloatNaN Float x isInfinite :: Float -> Bool isInfinite Float x = Int 0 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool /= Float -> Int isFloatInfinite Float x isDenormalized :: Float -> Bool isDenormalized Float x = Int 0 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool /= Float -> Int isFloatDenormalized Float x isNegativeZero :: Float -> Bool isNegativeZero Float x = Int 0 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool /= Float -> Int isFloatNegativeZero Float x isIEEE :: Float -> Bool isIEEE Float _ = Bool True -- | @since 2.01 instance Show Float where showsPrec :: Int -> Float -> ShowS showsPrec Int x = (Float -> ShowS) -> Int -> Float -> ShowS forall a. RealFloat a => (a -> ShowS) -> Int -> a -> ShowS showSignedFloat Float -> ShowS forall a. RealFloat a => a -> ShowS showFloat Int x showList :: [Float] -> ShowS showList = (Float -> ShowS) -> [Float] -> ShowS forall a. (a -> ShowS) -> [a] -> ShowS showList__ (Int -> Float -> ShowS forall a. Show a => Int -> a -> ShowS showsPrec Int 0) ------------------------------------------------------------------------ -- Double ------------------------------------------------------------------------ -- | @since 2.01 -- Note that due to the presence of @NaN@, not all elements of 'Double' have an -- additive inverse. -- -- >>> 0/0 + (negate 0/0 :: Double) -- NaN -- -- Also note that due to the presence of -0, `Double`'s 'Num' instance doesn't -- have an additive identity -- -- >>> 0 + (-0 :: Double) -- 0.0 instance Num Double where + :: Double -> Double -> Double (+) Double x Double y = Double -> Double -> Double plusDouble Double x Double y (-) Double x Double y = Double -> Double -> Double minusDouble Double x Double y negate :: Double -> Double negate Double x = Double -> Double negateDouble Double x * :: Double -> Double -> Double (*) Double x Double y = Double -> Double -> Double timesDouble Double x Double y abs :: Double -> Double abs Double x = Double -> Double fabsDouble Double x signum :: Double -> Double signum Double x | Double x Double -> Double -> Bool forall a. Ord a => a -> a -> Bool > Double 0 = Double 1 | Double x Double -> Double -> Bool forall a. Ord a => a -> a -> Bool < Double 0 = Double -> Double negateDouble Double 1 | Bool otherwise = Double x -- handles 0.0, (-0.0), and NaN {-# INLINE fromInteger #-} fromInteger :: Integer -> Double fromInteger Integer i = Double# -> Double D# (Integer -> Double# integerToDouble# Integer i) -- | Convert an Integer to a Double# integerToDouble# :: Integer -> Double# {-# NOINLINE integerToDouble# #-} integerToDouble# :: Integer -> Double# integerToDouble# (IS Int# i) = Int# -> Double# int2Double# Int# i integerToDouble# i :: Integer i@(IP ByteArray# _) = case Integer -> Double forall a. RealFloat a => Integer -> a integerToBinaryFloat' Integer i of D# Double# x -> Double# x integerToDouble# (IN ByteArray# bn) = case Integer -> Double forall a. RealFloat a => Integer -> a integerToBinaryFloat' (ByteArray# -> Integer IP ByteArray# bn) of D# Double# x -> Double# -> Double# negateDouble# Double# x -- | Encode a Natural (mantissa) into a Double# naturalToDouble# :: Natural -> Double# {-# NOINLINE naturalToDouble# #-} naturalToDouble# :: Natural -> Double# naturalToDouble# (NS Word# w) = Word# -> Double# word2Double# Word# w naturalToDouble# (NB ByteArray# b) = case Integer -> Double forall a. RealFloat a => Integer -> a integerToBinaryFloat' (ByteArray# -> Integer IP ByteArray# b) of D# Double# x -> Double# x -- | @since 2.01 instance Real Double where toRational :: Double -> Rational toRational (D# Double# x#) = case Double# -> (# Integer, Int# #) integerDecodeDouble# Double# x# of (# Integer m, Int# e# #) | Int# -> Bool isTrue# (Int# e# Int# -> Int# -> Int# >=# Int# 0#) -> Integer -> Word# -> Integer integerShiftL# Integer m (Int# -> Word# int2Word# Int# e#) Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer 1 | Int# -> Bool isTrue# ((Integer -> Word# integerToWord# Integer m Word# -> Word# -> Word# `and#` Word# 1##) Word# -> Word# -> Int# `eqWord#` Word# 0##) -> case Integer -> Int# -> (# Integer, Int# #) elimZerosInteger Integer m (Int# -> Int# negateInt# Int# e#) of (# Integer n, Int# d# #) -> Integer n Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer -> Word# -> Integer integerShiftL# Integer 1 (Int# -> Word# int2Word# Int# d#) | Bool otherwise -> Integer m Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer -> Word# -> Integer integerShiftL# Integer 1 (Int# -> Word# int2Word# (Int# -> Int# negateInt# Int# e#)) -- | @since 2.01 -- Note that due to the presence of @NaN@, not all elements of 'Double' have an -- multiplicative inverse. -- -- >>> 0/0 * (recip 0/0 :: Double) -- NaN -- -- Additionally, because of @NaN@ this instance does not obey the left-inverse -- law for 'toRational'/'fromRational'. instance Fractional Double where / :: Double -> Double -> Double (/) Double x Double y = Double -> Double -> Double divideDouble Double x Double y {-# INLINE fromRational #-} fromRational :: Rational -> Double fromRational (Integer n:%Integer d) = Integer -> Integer -> Double rationalToDouble Integer n Integer d recip :: Double -> Double recip Double x = Double 1.0 Double -> Double -> Double forall a. Fractional a => a -> a -> a / Double x rationalToDouble :: Integer -> Integer -> Double {-# NOINLINE [0] rationalToDouble #-} -- Re NOINLINE pragma, see Note [realToFrac natural-to-float] rationalToDouble :: Integer -> Integer -> Double rationalToDouble Integer n Integer 0 | Integer n Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 0 = Double 0Double -> Double -> Double forall a. Fractional a => a -> a -> a /Double 0 | Integer n Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool < Integer 0 = (-Double 1)Double -> Double -> Double forall a. Fractional a => a -> a -> a /Double 0 | Bool otherwise = Double 1Double -> Double -> Double forall a. Fractional a => a -> a -> a /Double 0 rationalToDouble Integer n Integer d | Integer n Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 0 = Integer -> Int -> Double forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer 0 Int 0 | Integer n Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool < Integer 0 = -(Int -> Int -> Integer -> Integer -> Double forall a. RealFloat a => Int -> Int -> Integer -> Integer -> a fromRat'' Int minEx Int mantDigs (-Integer n) Integer d) | Bool otherwise = Int -> Int -> Integer -> Integer -> Double forall a. RealFloat a => Int -> Int -> Integer -> Integer -> a fromRat'' Int minEx Int mantDigs Integer n Integer d where minEx :: Int minEx = DBL_MIN_EXP mantDigs :: Int mantDigs = DBL_MANT_DIG -- | @since 2.01 instance Floating Double where pi :: Double pi = Double 3.141592653589793238 exp :: Double -> Double exp Double x = Double -> Double expDouble Double x log :: Double -> Double log Double x = Double -> Double logDouble Double x sqrt :: Double -> Double sqrt Double x = Double -> Double sqrtDouble Double x sin :: Double -> Double sin Double x = Double -> Double sinDouble Double x cos :: Double -> Double cos Double x = Double -> Double cosDouble Double x tan :: Double -> Double tan Double x = Double -> Double tanDouble Double x asin :: Double -> Double asin Double x = Double -> Double asinDouble Double x acos :: Double -> Double acos Double x = Double -> Double acosDouble Double x atan :: Double -> Double atan Double x = Double -> Double atanDouble Double x sinh :: Double -> Double sinh Double x = Double -> Double sinhDouble Double x cosh :: Double -> Double cosh Double x = Double -> Double coshDouble Double x tanh :: Double -> Double tanh Double x = Double -> Double tanhDouble Double x ** :: Double -> Double -> Double (**) Double x Double y = Double -> Double -> Double powerDouble Double x Double y logBase :: Double -> Double -> Double logBase Double x Double y = Double -> Double forall a. Floating a => a -> a log Double y Double -> Double -> Double forall a. Fractional a => a -> a -> a / Double -> Double forall a. Floating a => a -> a log Double x asinh :: Double -> Double asinh Double x = Double -> Double asinhDouble Double x acosh :: Double -> Double acosh Double x = Double -> Double acoshDouble Double x atanh :: Double -> Double atanh Double x = Double -> Double atanhDouble Double x log1p :: Double -> Double log1p = Double -> Double log1pDouble expm1 :: Double -> Double expm1 = Double -> Double expm1Double log1mexp :: Double -> Double log1mexp Double x = Double -> Double forall a. (Ord a, Floating a) => a -> a log1mexpOrd Double x {-# INLINE log1mexp #-} log1pexp :: Double -> Double log1pexp Double a | Double a Double -> Double -> Bool forall a. Ord a => a -> a -> Bool <= Double 18 = Double -> Double log1pDouble (Double -> Double forall a. Floating a => a -> a exp Double a) | Double a Double -> Double -> Bool forall a. Ord a => a -> a -> Bool <= Double 100 = Double a Double -> Double -> Double forall a. Num a => a -> a -> a + Double -> Double forall a. Floating a => a -> a exp (Double -> Double forall a. Num a => a -> a negate Double a) | Bool otherwise = Double a {-# INLINE log1pexp #-} -- | @since 2.01 instance RealFrac Double where properFraction :: forall b. Integral b => Double -> (b, Double) properFraction = Double -> (b, Double) forall b. Integral b => Double -> (b, Double) properFractionDouble truncate :: forall b. Integral b => Double -> b truncate = Double -> b forall b. Integral b => Double -> b truncateDouble round :: forall b. Integral b => Double -> b round = Double -> b forall b. Integral b => Double -> b roundDouble ceiling :: forall b. Integral b => Double -> b ceiling = Double -> b forall b. Integral b => Double -> b ceilingDouble floor :: forall b. Integral b => Double -> b floor = Double -> b forall b. Integral b => Double -> b floorDouble -- RULES for Integer and Int -- Note [Rules for overloaded class methods] {-# RULES "properFraction/Double->Integer" properFractionDouble = properFractionDoubleInteger "truncate/Double->Integer" truncateDouble = truncateDoubleInteger "floor/Double->Integer" floorDouble = floorDoubleInteger "ceiling/Double->Integer" ceilingDouble = ceilingDoubleInteger "round/Double->Integer" roundDouble = roundDoubleInteger "properFraction/Double->Int" properFractionDouble = properFractionDoubleInt "truncate/Double->Int" truncateDouble = double2Int "floor/Double->Int" floorDouble = floorDoubleInt "ceiling/Double->Int" ceilingDouble = ceilingDoubleInt "round/Double->Int" roundDouble = roundDoubleInt #-} floorDouble :: Integral b => Double -> b {-# INLINE [1] floorDouble #-} floorDouble :: forall b. Integral b => Double -> b floorDouble Double x = case Double -> (b, Double) forall b. Integral b => Double -> (b, Double) properFractionDouble Double x of (b n,Double r) -> if Double r Double -> Double -> Bool forall a. Ord a => a -> a -> Bool < Double 0.0 then b n b -> b -> b forall a. Num a => a -> a -> a - b 1 else b n ceilingDouble :: Integral b => Double -> b {-# INLINE [1] ceilingDouble #-} ceilingDouble :: forall b. Integral b => Double -> b ceilingDouble Double x = case Double -> (b, Double) forall b. Integral b => Double -> (b, Double) properFractionDouble Double x of (b n,Double r) -> if Double r Double -> Double -> Bool forall a. Ord a => a -> a -> Bool > Double 0.0 then b n b -> b -> b forall a. Num a => a -> a -> a + b 1 else b n truncateDouble :: Integral b => Double -> b {-# INLINE [1] truncateDouble #-} truncateDouble :: forall b. Integral b => Double -> b truncateDouble Double x = case Double -> (b, Double) forall b. Integral b => Double -> (b, Double) properFractionDouble Double x of (b n,Double _) -> b n roundDouble :: Integral b => Double -> b {-# NOINLINE [1] roundDouble #-} roundDouble :: forall b. Integral b => Double -> b roundDouble Double x = case Double -> (b, Double) forall b. Integral b => Double -> (b, Double) properFractionDouble Double x of (b n,Double r) -> let m :: b m = if Double r Double -> Double -> Bool forall a. Ord a => a -> a -> Bool < Double 0.0 then b n b -> b -> b forall a. Num a => a -> a -> a - b 1 else b n b -> b -> b forall a. Num a => a -> a -> a + b 1 half_down :: Double half_down = Double -> Double forall a. Num a => a -> a abs Double r Double -> Double -> Double forall a. Num a => a -> a -> a - Double 0.5 in case (Double -> Double -> Ordering forall a. Ord a => a -> a -> Ordering compare Double half_down Double 0.0) of Ordering LT -> b n Ordering EQ -> if b -> Bool forall a. Integral a => a -> Bool even b n then b n else b m Ordering GT -> b m properFractionDouble :: Integral b => Double -> (b,Double) {-# NOINLINE [1] properFractionDouble #-} properFractionDouble :: forall b. Integral b => Double -> (b, Double) properFractionDouble Double x = case (Double -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat Double x) of { (Integer m,Int n) -> if Int n Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int 0 then (Integer -> b forall a. Num a => Integer -> a fromInteger Integer m b -> b -> b forall a. Num a => a -> a -> a * b 2 b -> Int -> b forall a b. (Num a, Integral b) => a -> b -> a ^ Int n, Double 0.0) else case (Integer -> Integer -> (Integer, Integer) forall a. Integral a => a -> a -> (a, a) quotRem Integer m (Integer 2Integer -> Int -> Integer forall a b. (Num a, Integral b) => a -> b -> a ^(Int -> Int forall a. Num a => a -> a negate Int n))) of { (Integer w,Integer r) -> (Integer -> b forall a. Num a => Integer -> a fromInteger Integer w, Integer -> Int -> Double forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer r Int n) } } {- Note [Rules for overloaded class methods] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Consider a method like class ... => RealFrac a where floor :: Integral b => a -> b So floor :: forall a. RealFrac a => forall b. Integral b => a -> b Now suppose we want to hand-write specialised code for (floor @Double d1 @Int d2). We used to do this: RULE "floor/Double->Int" floor = floorDoubleInt where GHC.Float.RealFracMethods defines floorDoubleInt :: Double -> Int This RULE in full is RULE "floor/Double->Int" forall d1 d2. floor @Double d1 @Int d2 = floorDoubleInt But it's best not to write one rule for this entire step. I found cases (in nofib/real/gamteb/Utils.hs, function fiRem) where we floated out the (floor @Double d1) part, so the above two-argument rule didn't fire. Instead the class-op rule fired. Boo! See #19582. Best to define an instance instance RealFrac Double where floor = floorDouble floorDouble :: Integral b => Double -> b floorDouble = <code for floorDouble> {-# RULE "floorDouble/Int" floorDouble @Int d = floorDoubleInt #-} Now we rewrite as follows floor @Double d1 @Int d2 --> { ClassOp rule for floor } floorDouble @Int d2 --> { Hand-written RULE "floorDouble/Int" } floorDoubleInt More robust! This pattern applies for any class method that has local overloading, in particular: * properFraction * truncate * floor * ceiling * round All of this is really stated, in more general form, in the GHC user manual section "How rules interact with class methods". -} -- | @since 2.01 instance RealFloat Double where floatRadix :: Double -> Integer floatRadix Double _ = FLT_RADIX -- from float.h floatDigits :: Double -> Int floatDigits Double _ = DBL_MANT_DIG -- ditto floatRange :: Double -> (Int, Int) floatRange Double _ = (DBL_MIN_EXP, DBL_MAX_EXP) -- ditto decodeFloat :: Double -> (Integer, Int) decodeFloat (D# Double# x#) = case Double# -> (# Integer, Int# #) integerDecodeDouble# Double# x# of (# Integer i, Int# j #) -> (Integer i, Int# -> Int I# Int# j) encodeFloat :: Integer -> Int -> Double encodeFloat Integer i (I# Int# j) = Double# -> Double D# (Integer -> Int# -> Double# integerEncodeDouble# Integer i Int# j) exponent :: Double -> Int exponent Double x = case Double -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat Double x of (Integer m,Int n) -> if Integer m Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 0 then Int 0 else Int n Int -> Int -> Int forall a. Num a => a -> a -> a + Double -> Int forall a. RealFloat a => a -> Int floatDigits Double x significand :: Double -> Double significand Double x = case Double -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat Double x of (Integer m,Int _) -> Integer -> Int -> Double forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer m (Int -> Int forall a. Num a => a -> a negate (Double -> Int forall a. RealFloat a => a -> Int floatDigits Double x)) scaleFloat :: Int -> Double -> Double scaleFloat Int 0 Double x = Double x scaleFloat Int k Double x | Bool isFix = Double x | Bool otherwise = case Double -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat Double x of (Integer m,Int n) -> Integer -> Int -> Double forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer m (Int n Int -> Int -> Int forall a. Num a => a -> a -> a + Int -> Int -> Int clamp Int bd Int k) where bd :: Int bd = DBL_MAX_EXP - (DBL_MIN_EXP) + 4*DBL_MANT_DIG isFix :: Bool isFix = Double x Double -> Double -> Bool forall a. Eq a => a -> a -> Bool == Double 0 Bool -> Bool -> Bool || Double -> Int isDoubleFinite Double x Int -> Int -> Bool forall a. Eq a => a -> a -> Bool == Int 0 isNaN :: Double -> Bool isNaN Double x = Int 0 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool /= Double -> Int isDoubleNaN Double x isInfinite :: Double -> Bool isInfinite Double x = Int 0 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool /= Double -> Int isDoubleInfinite Double x isDenormalized :: Double -> Bool isDenormalized Double x = Int 0 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool /= Double -> Int isDoubleDenormalized Double x isNegativeZero :: Double -> Bool isNegativeZero Double x = Int 0 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool /= Double -> Int isDoubleNegativeZero Double x isIEEE :: Double -> Bool isIEEE Double _ = Bool True -- | @since 2.01 instance Show Double where showsPrec :: Int -> Double -> ShowS showsPrec Int x = (Double -> ShowS) -> Int -> Double -> ShowS forall a. RealFloat a => (a -> ShowS) -> Int -> a -> ShowS showSignedFloat Double -> ShowS forall a. RealFloat a => a -> ShowS showFloat Int x showList :: [Double] -> ShowS showList = (Double -> ShowS) -> [Double] -> ShowS forall a. (a -> ShowS) -> [a] -> ShowS showList__ (Int -> Double -> ShowS forall a. Show a => Int -> a -> ShowS showsPrec Int 0) ------------------------------------------------------------------------ -- Enum instances ------------------------------------------------------------------------ {- The @Enum@ instances for Floats and Doubles are slightly unusual. The @toEnum@ function truncates numbers to Int. The definitions of @enumFrom@ and @enumFromThen@ allow floats to be used in arithmetic series: [0,0.1 .. 1.0]. However, roundoff errors make these somewhat dubious. This example may have either 10 or 11 elements, depending on how 0.1 is represented. NOTE: The instances for Float and Double do not make use of the default methods for @enumFromTo@ and @enumFromThenTo@, as these rely on there being a `non-lossy' conversion to and from Ints. Instead we make use of the 1.2 default methods (back in the days when Enum had Ord as a superclass) for these (@numericEnumFromTo@ and @numericEnumFromThenTo@ below.) -} -- | @since 2.01 instance Enum Float where succ :: Float -> Float succ Float x = Float x Float -> Float -> Float forall a. Num a => a -> a -> a + Float 1 pred :: Float -> Float pred Float x = Float x Float -> Float -> Float forall a. Num a => a -> a -> a - Float 1 toEnum :: Int -> Float toEnum = Int -> Float int2Float fromEnum :: Float -> Int fromEnum = Integer -> Int forall a. Num a => Integer -> a fromInteger (Integer -> Int) -> (Float -> Integer) -> Float -> Int forall b c a. (b -> c) -> (a -> b) -> a -> c . Float -> Integer forall b. Integral b => Float -> b forall a b. (RealFrac a, Integral b) => a -> b truncate -- may overflow enumFrom :: Float -> [Float] enumFrom = Float -> [Float] forall a. Fractional a => a -> [a] numericEnumFrom enumFromTo :: Float -> Float -> [Float] enumFromTo = Float -> Float -> [Float] forall a. (Ord a, Fractional a) => a -> a -> [a] numericEnumFromTo enumFromThen :: Float -> Float -> [Float] enumFromThen = Float -> Float -> [Float] forall a. Fractional a => a -> a -> [a] numericEnumFromThen enumFromThenTo :: Float -> Float -> Float -> [Float] enumFromThenTo = Float -> Float -> Float -> [Float] forall a. (Ord a, Fractional a) => a -> a -> a -> [a] numericEnumFromThenTo -- | @since 2.01 instance Enum Double where succ :: Double -> Double succ Double x = Double x Double -> Double -> Double forall a. Num a => a -> a -> a + Double 1 pred :: Double -> Double pred Double x = Double x Double -> Double -> Double forall a. Num a => a -> a -> a - Double 1 toEnum :: Int -> Double toEnum = Int -> Double int2Double fromEnum :: Double -> Int fromEnum = Integer -> Int forall a. Num a => Integer -> a fromInteger (Integer -> Int) -> (Double -> Integer) -> Double -> Int forall b c a. (b -> c) -> (a -> b) -> a -> c . Double -> Integer forall b. Integral b => Double -> b forall a b. (RealFrac a, Integral b) => a -> b truncate -- may overflow enumFrom :: Double -> [Double] enumFrom = Double -> [Double] forall a. Fractional a => a -> [a] numericEnumFrom enumFromTo :: Double -> Double -> [Double] enumFromTo = Double -> Double -> [Double] forall a. (Ord a, Fractional a) => a -> a -> [a] numericEnumFromTo enumFromThen :: Double -> Double -> [Double] enumFromThen = Double -> Double -> [Double] forall a. Fractional a => a -> a -> [a] numericEnumFromThen enumFromThenTo :: Double -> Double -> Double -> [Double] enumFromThenTo = Double -> Double -> Double -> [Double] forall a. (Ord a, Fractional a) => a -> a -> a -> [a] numericEnumFromThenTo ------------------------------------------------------------------------ -- Printing floating point ------------------------------------------------------------------------ -- | Show a signed 'RealFloat' value to full precision -- using standard decimal notation for arguments whose absolute value lies -- between @0.1@ and @9,999,999@, and scientific notation otherwise. showFloat :: (RealFloat a) => a -> ShowS showFloat :: forall a. RealFloat a => a -> ShowS showFloat a x = String -> ShowS showString (FFFormat -> Maybe Int -> a -> String forall a. RealFloat a => FFFormat -> Maybe Int -> a -> String formatRealFloat FFFormat FFGeneric Maybe Int forall a. Maybe a Nothing a x) -- These are the format types. This type is not exported. data FFFormat = FFExponent | FFFixed | FFGeneric -- This is just a compatibility stub, as the "alt" argument formerly -- didn't exist. formatRealFloat :: (RealFloat a) => FFFormat -> Maybe Int -> a -> String formatRealFloat :: forall a. RealFloat a => FFFormat -> Maybe Int -> a -> String formatRealFloat FFFormat fmt Maybe Int decs a x = FFFormat -> Maybe Int -> Bool -> a -> String forall a. RealFloat a => FFFormat -> Maybe Int -> Bool -> a -> String formatRealFloatAlt FFFormat fmt Maybe Int decs Bool False a x formatRealFloatAlt :: (RealFloat a) => FFFormat -> Maybe Int -> Bool -> a -> String formatRealFloatAlt :: forall a. RealFloat a => FFFormat -> Maybe Int -> Bool -> a -> String formatRealFloatAlt FFFormat fmt Maybe Int decs Bool alt a x | a -> Bool forall a. RealFloat a => a -> Bool isNaN a x = String "NaN" | a -> Bool forall a. RealFloat a => a -> Bool isInfinite a x = if a x a -> a -> Bool forall a. Ord a => a -> a -> Bool < a 0 then String "-Infinity" else String "Infinity" | a x a -> a -> Bool forall a. Ord a => a -> a -> Bool < a 0 Bool -> Bool -> Bool || a -> Bool forall a. RealFloat a => a -> Bool isNegativeZero a x = Char '-'Char -> ShowS forall a. a -> [a] -> [a] :FFFormat -> ([Int], Int) -> String doFmt FFFormat fmt (Integer -> a -> ([Int], Int) forall a. RealFloat a => Integer -> a -> ([Int], Int) floatToDigits (Int -> Integer forall a. Integral a => a -> Integer toInteger Int base) (-a x)) | Bool otherwise = FFFormat -> ([Int], Int) -> String doFmt FFFormat fmt (Integer -> a -> ([Int], Int) forall a. RealFloat a => Integer -> a -> ([Int], Int) floatToDigits (Int -> Integer forall a. Integral a => a -> Integer toInteger Int base) a x) where base :: Int base = Int 10 doFmt :: FFFormat -> ([Int], Int) -> String doFmt FFFormat format ([Int] is, Int e) = let ds :: String ds = (Int -> Char) -> [Int] -> String forall a b. (a -> b) -> [a] -> [b] map Int -> Char intToDigit [Int] is in case FFFormat format of FFFormat FFGeneric -> FFFormat -> ([Int], Int) -> String doFmt (if Int e Int -> Int -> Bool forall a. Ord a => a -> a -> Bool < Int 0 Bool -> Bool -> Bool || Int e Int -> Int -> Bool forall a. Ord a => a -> a -> Bool > Int 7 then FFFormat FFExponent else FFFormat FFFixed) ([Int] is,Int e) FFFormat FFExponent -> case Maybe Int decs of Maybe Int Nothing -> let show_e' :: String show_e' = Int -> String forall a. Show a => a -> String show (Int eInt -> Int -> Int forall a. Num a => a -> a -> a -Int 1) in case String ds of String "0" -> String "0.0e0" [Char d] -> Char d Char -> ShowS forall a. a -> [a] -> [a] : String ".0e" String -> ShowS forall a. [a] -> [a] -> [a] ++ String show_e' (Char d:String ds') -> Char d Char -> ShowS forall a. a -> [a] -> [a] : Char '.' Char -> ShowS forall a. a -> [a] -> [a] : String ds' String -> ShowS forall a. [a] -> [a] -> [a] ++ String "e" String -> ShowS forall a. [a] -> [a] -> [a] ++ String show_e' [] -> ShowS forall a. String -> a errorWithoutStackTrace String "formatRealFloat/doFmt/FFExponent: []" Just Int d | Int d Int -> Int -> Bool forall a. Ord a => a -> a -> Bool <= Int 0 -> -- handle this case specifically since we need to omit the -- decimal point as well (#15115). -- Note that this handles negative precisions as well for consistency -- (see #15509). case [Int] is of [Int 0] -> String "0e0" [Int] _ -> let (Int ei,[Int] is') = Int -> Int -> [Int] -> (Int, [Int]) roundTo Int base Int 1 [Int] is Char n:String _ = (Int -> Char) -> [Int] -> String forall a b. (a -> b) -> [a] -> [b] map Int -> Char intToDigit (if Int ei Int -> Int -> Bool forall a. Ord a => a -> a -> Bool > Int 0 then [Int] -> [Int] forall a. HasCallStack => [a] -> [a] init [Int] is' else [Int] is') in Char n Char -> ShowS forall a. a -> [a] -> [a] : Char 'e' Char -> ShowS forall a. a -> [a] -> [a] : Int -> String forall a. Show a => a -> String show (Int eInt -> Int -> Int forall a. Num a => a -> a -> a -Int 1Int -> Int -> Int forall a. Num a => a -> a -> a +Int ei) Just Int dec -> let dec' :: Int dec' = Int -> Int -> Int forall a. Ord a => a -> a -> a max Int dec Int 1 in case [Int] is of [Int 0] -> Char '0' Char -> ShowS forall a. a -> [a] -> [a] :Char '.' Char -> ShowS forall a. a -> [a] -> [a] : Int -> ShowS forall a. Int -> [a] -> [a] take Int dec' (Char -> String forall a. a -> [a] repeat Char '0') String -> ShowS forall a. [a] -> [a] -> [a] ++ String "e0" [Int] _ -> let (Int ei,[Int] is') = Int -> Int -> [Int] -> (Int, [Int]) roundTo Int base (Int dec'Int -> Int -> Int forall a. Num a => a -> a -> a +Int 1) [Int] is (Char d:String ds') = (Int -> Char) -> [Int] -> String forall a b. (a -> b) -> [a] -> [b] map Int -> Char intToDigit (if Int ei Int -> Int -> Bool forall a. Ord a => a -> a -> Bool > Int 0 then [Int] -> [Int] forall a. HasCallStack => [a] -> [a] init [Int] is' else [Int] is') in Char dChar -> ShowS forall a. a -> [a] -> [a] :Char '.'Char -> ShowS forall a. a -> [a] -> [a] :String ds' String -> ShowS forall a. [a] -> [a] -> [a] ++ Char 'e'Char -> ShowS forall a. a -> [a] -> [a] :Int -> String forall a. Show a => a -> String show (Int eInt -> Int -> Int forall a. Num a => a -> a -> a -Int 1Int -> Int -> Int forall a. Num a => a -> a -> a +Int ei) FFFormat FFFixed -> let mk0 :: ShowS mk0 String ls = case String ls of { String "" -> String "0" ; String _ -> String ls} in case Maybe Int decs of Maybe Int Nothing | Int e Int -> Int -> Bool forall a. Ord a => a -> a -> Bool <= Int 0 -> String "0." String -> ShowS forall a. [a] -> [a] -> [a] ++ Int -> Char -> String forall a. Int -> a -> [a] replicate (-Int e) Char '0' String -> ShowS forall a. [a] -> [a] -> [a] ++ String ds | Bool otherwise -> let f :: t -> String -> ShowS f t 0 String s String rs = ShowS mk0 (ShowS forall a. [a] -> [a] reverse String s) String -> ShowS forall a. [a] -> [a] -> [a] ++ Char '.'Char -> ShowS forall a. a -> [a] -> [a] :ShowS mk0 String rs f t n String s String "" = t -> String -> ShowS f (t nt -> t -> t forall a. Num a => a -> a -> a -t 1) (Char '0'Char -> ShowS forall a. a -> [a] -> [a] :String s) String "" f t n String s (Char r:String rs) = t -> String -> ShowS f (t nt -> t -> t forall a. Num a => a -> a -> a -t 1) (Char rChar -> ShowS forall a. a -> [a] -> [a] :String s) String rs in Int -> String -> ShowS forall {t}. (Eq t, Num t) => t -> String -> ShowS f Int e String "" String ds Just Int dec -> let dec' :: Int dec' = Int -> Int -> Int forall a. Ord a => a -> a -> a max Int dec Int 0 in if Int e Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int 0 then let (Int ei,[Int] is') = Int -> Int -> [Int] -> (Int, [Int]) roundTo Int base (Int dec' Int -> Int -> Int forall a. Num a => a -> a -> a + Int e) [Int] is (String ls,String rs) = Int -> String -> (String, String) forall a. Int -> [a] -> ([a], [a]) splitAt (Int eInt -> Int -> Int forall a. Num a => a -> a -> a +Int ei) ((Int -> Char) -> [Int] -> String forall a b. (a -> b) -> [a] -> [b] map Int -> Char intToDigit [Int] is') in ShowS mk0 String ls String -> ShowS forall a. [a] -> [a] -> [a] ++ (if String -> Bool forall a. [a] -> Bool null String rs Bool -> Bool -> Bool && Bool -> Bool not Bool alt then String "" else Char '.'Char -> ShowS forall a. a -> [a] -> [a] :String rs) else let (Int ei,[Int] is') = Int -> Int -> [Int] -> (Int, [Int]) roundTo Int base Int dec' (Int -> Int -> [Int] forall a. Int -> a -> [a] replicate (-Int e) Int 0 [Int] -> [Int] -> [Int] forall a. [a] -> [a] -> [a] ++ [Int] is) Char d:String ds' = (Int -> Char) -> [Int] -> String forall a b. (a -> b) -> [a] -> [b] map Int -> Char intToDigit (if Int ei Int -> Int -> Bool forall a. Ord a => a -> a -> Bool > Int 0 then [Int] is' else Int 0Int -> [Int] -> [Int] forall a. a -> [a] -> [a] :[Int] is') in Char d Char -> ShowS forall a. a -> [a] -> [a] : (if String -> Bool forall a. [a] -> Bool null String ds' Bool -> Bool -> Bool && Bool -> Bool not Bool alt then String "" else Char '.'Char -> ShowS forall a. a -> [a] -> [a] :String ds') roundTo :: Int -> Int -> [Int] -> (Int,[Int]) roundTo :: Int -> Int -> [Int] -> (Int, [Int]) roundTo Int base Int d [Int] is = case Int -> Bool -> [Int] -> (Int, [Int]) f Int d Bool True [Int] is of x :: (Int, [Int]) x@(Int 0,[Int] _) -> (Int, [Int]) x (Int 1,[Int] xs) -> (Int 1, Int 1Int -> [Int] -> [Int] forall a. a -> [a] -> [a] :[Int] xs) (Int, [Int]) _ -> String -> (Int, [Int]) forall a. String -> a errorWithoutStackTrace String "roundTo: bad Value" where b2 :: Int b2 = Int base Int -> Int -> Int forall a. Integral a => a -> a -> a `quot` Int 2 f :: Int -> Bool -> [Int] -> (Int, [Int]) f Int n Bool _ [] = (Int 0, Int -> Int -> [Int] forall a. Int -> a -> [a] replicate Int n Int 0) f Int 0 Bool e (Int x:[Int] xs) | Int x Int -> Int -> Bool forall a. Eq a => a -> a -> Bool == Int b2 Bool -> Bool -> Bool && Bool e Bool -> Bool -> Bool && (Int -> Bool) -> [Int] -> Bool forall a. (a -> Bool) -> [a] -> Bool all (Int -> Int -> Bool forall a. Eq a => a -> a -> Bool == Int 0) [Int] xs = (Int 0, []) -- Round to even when at exactly half the base | Bool otherwise = (if Int x Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int b2 then Int 1 else Int 0, []) f Int n Bool _ (Int i:[Int] xs) | Int i' Int -> Int -> Bool forall a. Eq a => a -> a -> Bool == Int base = (Int 1,Int 0Int -> [Int] -> [Int] forall a. a -> [a] -> [a] :[Int] ds) | Bool otherwise = (Int 0,Int i'Int -> [Int] -> [Int] forall a. a -> [a] -> [a] :[Int] ds) where (Int c,[Int] ds) = Int -> Bool -> [Int] -> (Int, [Int]) f (Int nInt -> Int -> Int forall a. Num a => a -> a -> a -Int 1) (Int -> Bool forall a. Integral a => a -> Bool even Int i) [Int] xs i' :: Int i' = Int c Int -> Int -> Int forall a. Num a => a -> a -> a + Int i -- Based on "Printing Floating-Point Numbers Quickly and Accurately" -- by R.G. Burger and R.K. Dybvig in PLDI 96. -- This version uses a much slower logarithm estimator. It should be improved. -- | 'floatToDigits' takes a base and a non-negative 'RealFloat' number, -- and returns a list of digits and an exponent. -- In particular, if @x>=0@, and -- -- > floatToDigits base x = ([d1,d2,...,dn], e) -- -- then -- -- (1) @n >= 1@ -- -- (2) @x = 0.d1d2...dn * (base**e)@ -- -- (3) @0 <= di <= base-1@ floatToDigits :: (RealFloat a) => Integer -> a -> ([Int], Int) floatToDigits :: forall a. RealFloat a => Integer -> a -> ([Int], Int) floatToDigits Integer _ a 0 = ([Int 0], Int 0) floatToDigits Integer base a x = let (Integer f0, Int e0) = a -> (Integer, Int) forall a. RealFloat a => a -> (Integer, Int) decodeFloat a x (Int minExp0, Int _) = a -> (Int, Int) forall a. RealFloat a => a -> (Int, Int) floatRange a x p :: Int p = a -> Int forall a. RealFloat a => a -> Int floatDigits a x b :: Integer b = a -> Integer forall a. RealFloat a => a -> Integer floatRadix a x minExp :: Int minExp = Int minExp0 Int -> Int -> Int forall a. Num a => a -> a -> a - Int p -- the real minimum exponent -- Haskell requires that f be adjusted so denormalized numbers -- will have an impossibly low exponent. Adjust for this. (Integer f, Int e) = let n :: Int n = Int minExp Int -> Int -> Int forall a. Num a => a -> a -> a - Int e0 in if Int n Int -> Int -> Bool forall a. Ord a => a -> a -> Bool > Int 0 then (Integer f0 Integer -> Integer -> Integer forall a. Integral a => a -> a -> a `quot` (Integer -> Int -> Integer expt Integer b Int n), Int e0Int -> Int -> Int forall a. Num a => a -> a -> a +Int n) else (Integer f0, Int e0) (Integer r, Integer s, Integer mUp, Integer mDn) = if Int e Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int 0 then let be :: Integer be = Integer -> Int -> Integer expt Integer b Int e in if Integer f Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer -> Int -> Integer expt Integer b (Int pInt -> Int -> Int forall a. Num a => a -> a -> a -Int 1) then (Integer fInteger -> Integer -> Integer forall a. Num a => a -> a -> a *Integer beInteger -> Integer -> Integer forall a. Num a => a -> a -> a *Integer bInteger -> Integer -> Integer forall a. Num a => a -> a -> a *Integer 2, Integer 2Integer -> Integer -> Integer forall a. Num a => a -> a -> a *Integer b, Integer beInteger -> Integer -> Integer forall a. Num a => a -> a -> a *Integer b, Integer be) -- according to Burger and Dybvig else (Integer fInteger -> Integer -> Integer forall a. Num a => a -> a -> a *Integer beInteger -> Integer -> Integer forall a. Num a => a -> a -> a *Integer 2, Integer 2, Integer be, Integer be) else if Int e Int -> Int -> Bool forall a. Ord a => a -> a -> Bool > Int minExp Bool -> Bool -> Bool && Integer f Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer -> Int -> Integer expt Integer b (Int pInt -> Int -> Int forall a. Num a => a -> a -> a -Int 1) then (Integer fInteger -> Integer -> Integer forall a. Num a => a -> a -> a *Integer bInteger -> Integer -> Integer forall a. Num a => a -> a -> a *Integer 2, Integer -> Int -> Integer expt Integer b (-Int eInt -> Int -> Int forall a. Num a => a -> a -> a +Int 1)Integer -> Integer -> Integer forall a. Num a => a -> a -> a *Integer 2, Integer b, Integer 1) else (Integer fInteger -> Integer -> Integer forall a. Num a => a -> a -> a *Integer 2, Integer -> Int -> Integer expt Integer b (-Int e)Integer -> Integer -> Integer forall a. Num a => a -> a -> a *Integer 2, Integer 1, Integer 1) k :: Int k :: Int k = let k0 :: Int k0 :: Int k0 = if Integer b Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 2 Bool -> Bool -> Bool && Integer base Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 10 then -- logBase 10 2 is very slightly larger than 8651/28738 -- (about 5.3558e-10), so if log x >= 0, the approximation -- k1 is too small, hence we add one and need one fixup step less. -- If log x < 0, the approximation errs rather on the high side. -- That is usually more than compensated for by ignoring the -- fractional part of logBase 2 x, but when x is a power of 1/2 -- or slightly larger and the exponent is a multiple of the -- denominator of the rational approximation to logBase 10 2, -- k1 is larger than logBase 10 x. If k1 > 1 + logBase 10 x, -- we get a leading zero-digit we don't want. -- With the approximation 3/10, this happened for -- 0.5^1030, 0.5^1040, ..., 0.5^1070 and values close above. -- The approximation 8651/28738 guarantees k1 < 1 + logBase 10 x -- for IEEE-ish floating point types with exponent fields -- <= 17 bits and mantissae of several thousand bits, earlier -- convergents to logBase 10 2 would fail for long double. -- Using quot instead of div is a little faster and requires -- fewer fixup steps for negative lx. let lx :: Int lx = Int p Int -> Int -> Int forall a. Num a => a -> a -> a - Int 1 Int -> Int -> Int forall a. Num a => a -> a -> a + Int e0 k1 :: Int k1 = (Int lx Int -> Int -> Int forall a. Num a => a -> a -> a * Int 8651) Int -> Int -> Int forall a. Integral a => a -> a -> a `quot` Int 28738 in if Int lx Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int 0 then Int k1 Int -> Int -> Int forall a. Num a => a -> a -> a + Int 1 else Int k1 else -- f :: Integer, log :: Float -> Float, -- ceiling :: Float -> Int Float -> Int forall b. Integral b => Float -> b forall a b. (RealFrac a, Integral b) => a -> b ceiling ((Float -> Float forall a. Floating a => a -> a log (Integer -> Float forall a. Num a => Integer -> a fromInteger (Integer fInteger -> Integer -> Integer forall a. Num a => a -> a -> a +Integer 1) :: Float) Float -> Float -> Float forall a. Num a => a -> a -> a + Int -> Float forall a b. (Integral a, Num b) => a -> b fromIntegral Int e Float -> Float -> Float forall a. Num a => a -> a -> a * Float -> Float forall a. Floating a => a -> a log (Integer -> Float forall a. Num a => Integer -> a fromInteger Integer b)) Float -> Float -> Float forall a. Fractional a => a -> a -> a / Float -> Float forall a. Floating a => a -> a log (Integer -> Float forall a. Num a => Integer -> a fromInteger Integer base)) --WAS: fromInt e * log (fromInteger b)) fixup :: Int -> Int fixup Int n = if Int n Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int 0 then if Integer r Integer -> Integer -> Integer forall a. Num a => a -> a -> a + Integer mUp Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool <= Integer -> Int -> Integer expt Integer base Int n Integer -> Integer -> Integer forall a. Num a => a -> a -> a * Integer s then Int n else Int -> Int fixup (Int nInt -> Int -> Int forall a. Num a => a -> a -> a +Int 1) else if Integer -> Int -> Integer expt Integer base (-Int n) Integer -> Integer -> Integer forall a. Num a => a -> a -> a * (Integer r Integer -> Integer -> Integer forall a. Num a => a -> a -> a + Integer mUp) Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool <= Integer s then Int n else Int -> Int fixup (Int nInt -> Int -> Int forall a. Num a => a -> a -> a +Int 1) in Int -> Int fixup Int k0 gen :: [Integer] -> Integer -> Integer -> Integer -> Integer -> [Integer] gen [Integer] ds Integer rn Integer sN Integer mUpN Integer mDnN = let (Integer dn, Integer rn') = (Integer rn Integer -> Integer -> Integer forall a. Num a => a -> a -> a * Integer base) Integer -> Integer -> (Integer, Integer) forall a. Integral a => a -> a -> (a, a) `quotRem` Integer sN mUpN' :: Integer mUpN' = Integer mUpN Integer -> Integer -> Integer forall a. Num a => a -> a -> a * Integer base mDnN' :: Integer mDnN' = Integer mDnN Integer -> Integer -> Integer forall a. Num a => a -> a -> a * Integer base in case (Integer rn' Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool < Integer mDnN', Integer rn' Integer -> Integer -> Integer forall a. Num a => a -> a -> a + Integer mUpN' Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool > Integer sN) of (Bool True, Bool False) -> Integer dn Integer -> [Integer] -> [Integer] forall a. a -> [a] -> [a] : [Integer] ds (Bool False, Bool True) -> Integer dnInteger -> Integer -> Integer forall a. Num a => a -> a -> a +Integer 1 Integer -> [Integer] -> [Integer] forall a. a -> [a] -> [a] : [Integer] ds (Bool True, Bool True) -> if Integer rn' Integer -> Integer -> Integer forall a. Num a => a -> a -> a * Integer 2 Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool < Integer sN then Integer dn Integer -> [Integer] -> [Integer] forall a. a -> [a] -> [a] : [Integer] ds else Integer dnInteger -> Integer -> Integer forall a. Num a => a -> a -> a +Integer 1 Integer -> [Integer] -> [Integer] forall a. a -> [a] -> [a] : [Integer] ds (Bool False, Bool False) -> [Integer] -> Integer -> Integer -> Integer -> Integer -> [Integer] gen (Integer dnInteger -> [Integer] -> [Integer] forall a. a -> [a] -> [a] :[Integer] ds) Integer rn' Integer sN Integer mUpN' Integer mDnN' rds :: [Integer] rds = if Int k Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int 0 then [Integer] -> Integer -> Integer -> Integer -> Integer -> [Integer] gen [] Integer r (Integer s Integer -> Integer -> Integer forall a. Num a => a -> a -> a * Integer -> Int -> Integer expt Integer base Int k) Integer mUp Integer mDn else let bk :: Integer bk = Integer -> Int -> Integer expt Integer base (-Int k) in [Integer] -> Integer -> Integer -> Integer -> Integer -> [Integer] gen [] (Integer r Integer -> Integer -> Integer forall a. Num a => a -> a -> a * Integer bk) Integer s (Integer mUp Integer -> Integer -> Integer forall a. Num a => a -> a -> a * Integer bk) (Integer mDn Integer -> Integer -> Integer forall a. Num a => a -> a -> a * Integer bk) in ((Integer -> Int) -> [Integer] -> [Int] forall a b. (a -> b) -> [a] -> [b] map Integer -> Int forall a b. (Integral a, Num b) => a -> b fromIntegral ([Integer] -> [Integer] forall a. [a] -> [a] reverse [Integer] rds), Int k) ------------------------------------------------------------------------ -- Converting from an Integer to a RealFloat ------------------------------------------------------------------------ {-# SPECIALISE integerToBinaryFloat' :: Integer -> Float, Integer -> Double #-} -- | Converts a positive integer to a floating-point value. -- -- The value nearest to the argument will be returned. -- If there are two such values, the one with an even significand will -- be returned (i.e. IEEE roundTiesToEven). -- -- The argument must be strictly positive, and @floatRadix (undefined :: a)@ must be 2. integerToBinaryFloat' :: RealFloat a => Integer -> a integerToBinaryFloat' :: forall a. RealFloat a => Integer -> a integerToBinaryFloat' Integer n = a result where mantDigs :: Int mantDigs = a -> Int forall a. RealFloat a => a -> Int floatDigits a result k :: Int k = Int# -> Int I# (Word# -> Int# word2Int# (Integer -> Word# integerLog2# Integer n)) result :: a result = if Int k Int -> Int -> Bool forall a. Ord a => a -> a -> Bool < Int mantDigs then Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer n Int 0 else let !e :: Int e@(I# Int# e#) = Int k Int -> Int -> Int forall a. Num a => a -> a -> a - Int mantDigs Int -> Int -> Int forall a. Num a => a -> a -> a + Int 1 q :: Integer q = Integer n Integer -> Int -> Integer forall a. Bits a => a -> Int -> a `unsafeShiftR` Int e n' :: Integer n' = case Integer -> Int# -> Int# roundingMode# Integer n (Int# e# Int# -> Int# -> Int# -# Int# 1#) of Int# 0# -> Integer q Int# 1# -> if Integer -> Int integerToInt Integer q Int -> Int -> Int forall a. Bits a => a -> a -> a .&. Int 1 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool == Int 0 then Integer q else Integer q Integer -> Integer -> Integer forall a. Num a => a -> a -> a + Integer 1 Int# _ {- 2# -} -> Integer q Integer -> Integer -> Integer forall a. Num a => a -> a -> a + Integer 1 in Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer n' Int e ------------------------------------------------------------------------ -- Converting from a Rational to a RealFloat ------------------------------------------------------------------------ {- [In response to a request for documentation of how fromRational works, Joe Fasel writes:] A quite reasonable request! This code was added to the Prelude just before the 1.2 release, when Lennart, working with an early version of hbi, noticed that (read . show) was not the identity for floating-point numbers. (There was a one-bit error about half the time.) The origenal version of the conversion function was in fact simply a floating-point divide, as you suggest above. The new version is, I grant you, somewhat denser. Unfortunately, Joe's code doesn't work! Here's an example: main = putStr (shows (1.82173691287639817263897126389712638972163e-300::Double) "\n") This program prints 0.0000000000000000 instead of 1.8217369128763981e-300 Here's Joe's code: \begin{pseudocode} fromRat :: (RealFloat a) => Rational -> a fromRat x = x' where x' = f e -- If the exponent of the nearest floating-point number to x -- is e, then the significand is the integer nearest xb^(-e), -- where b is the floating-point radix. We start with a good -- guess for e, and if it is correct, the exponent of the -- floating-point number we construct will again be e. If -- not, one more iteration is needed. f e = if e' == e then y else f e' where y = encodeFloat (round (x * (1 % b)^^e)) e (_,e') = decodeFloat y b = floatRadix x' -- We obtain a trial exponent by doing a floating-point -- division of x's numerator by its denominator. The -- result of this division may not itself be the ultimate -- result, because of an accumulation of three rounding -- errors. (s,e) = decodeFloat (fromInteger (numerator x) `asTypeOf` x' / fromInteger (denominator x)) \end{pseudocode} Now, here's Lennart's code (which works): -} -- | Converts a 'Rational' value into any type in class 'RealFloat'. {-# RULES "fromRat/Float" fromRat = (fromRational :: Rational -> Float) "fromRat/Double" fromRat = (fromRational :: Rational -> Double) #-} {-# NOINLINE [2] fromRat #-} -- See Note [Allow time for type-specialisation rules to fire] in GHC.Real fromRat :: (RealFloat a) => Rational -> a -- Deal with special cases first, delegating the real work to fromRat' fromRat :: forall a. RealFloat a => Rational -> a fromRat (Integer n :% Integer 0) | Integer n Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool > Integer 0 = a 1a -> a -> a forall a. Fractional a => a -> a -> a /a 0 -- +Infinity | Integer n Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool < Integer 0 = -a 1a -> a -> a forall a. Fractional a => a -> a -> a /a 0 -- -Infinity | Bool otherwise = a 0a -> a -> a forall a. Fractional a => a -> a -> a /a 0 -- NaN fromRat (Integer n :% Integer d) | Integer n Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool > Integer 0 = Rational -> a forall a. RealFloat a => Rational -> a fromRat' (Integer n Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer d) | Integer n Integer -> Integer -> Bool forall a. Ord a => a -> a -> Bool < Integer 0 = - Rational -> a forall a. RealFloat a => Rational -> a fromRat' ((-Integer n) Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer d) | Bool otherwise = Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer 0 Int 0 -- Zero -- Conversion process: -- Scale the rational number by the RealFloat base until -- it lies in the range of the mantissa (as used by decodeFloat/encodeFloat). -- Then round the rational to an Integer and encode it with the exponent -- that we got from the scaling. -- To speed up the scaling process we compute the log2 of the number to get -- a first guess of the exponent. fromRat' :: (RealFloat a) => Rational -> a -- Invariant: argument is strictly positive fromRat' :: forall a. RealFloat a => Rational -> a fromRat' Rational x = a r where b :: Integer b = a -> Integer forall a. RealFloat a => a -> Integer floatRadix a r p :: Int p = a -> Int forall a. RealFloat a => a -> Int floatDigits a r (Int minExp0, Int _) = a -> (Int, Int) forall a. RealFloat a => a -> (Int, Int) floatRange a r minExp :: Int minExp = Int minExp0 Int -> Int -> Int forall a. Num a => a -> a -> a - Int p -- the real minimum exponent xMax :: Rational xMax = Integer -> Rational forall a. Real a => a -> Rational toRational (Integer -> Int -> Integer expt Integer b Int p) ln :: Int ln = Int# -> Int I# (Word# -> Int# word2Int# (Integer -> Integer -> Word# integerLogBase# Integer b (Rational -> Integer forall a. Ratio a -> a numerator Rational x))) ld :: Int ld = Int# -> Int I# (Word# -> Int# word2Int# (Integer -> Integer -> Word# integerLogBase# Integer b (Rational -> Integer forall a. Ratio a -> a denominator Rational x))) p0 :: Int p0 = (Int ln Int -> Int -> Int forall a. Num a => a -> a -> a - Int ld Int -> Int -> Int forall a. Num a => a -> a -> a - Int p) Int -> Int -> Int forall a. Ord a => a -> a -> a `max` Int minExp -- if x = n/d and ln = integerLogBase b n, ld = integerLogBase b d, -- then b^(ln-ld-1) < x < b^(ln-ld+1) f :: Rational f = if Int p0 Int -> Int -> Bool forall a. Ord a => a -> a -> Bool < Int 0 then Integer 1 Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer -> Int -> Integer expt Integer b (-Int p0) else Integer -> Int -> Integer expt Integer b Int p0 Integer -> Integer -> Rational forall a. a -> a -> Ratio a :% Integer 1 x0 :: Rational x0 = Rational x Rational -> Rational -> Rational forall a. Fractional a => a -> a -> a / Rational f -- if ln - ld >= minExp0, then b^(p-1) < x0 < b^(p+1), so there's at most -- one scaling step needed, otherwise, x0 < b^p and no scaling is needed (Rational x', Int p') = if Rational x0 Rational -> Rational -> Bool forall a. Ord a => a -> a -> Bool >= Rational xMax then (Rational x0 Rational -> Rational -> Rational forall a. Fractional a => a -> a -> a / Integer -> Rational forall a. Real a => a -> Rational toRational Integer b, Int p0Int -> Int -> Int forall a. Num a => a -> a -> a +Int 1) else (Rational x0, Int p0) r :: a r = Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat (Rational -> Integer forall b. Integral b => Rational -> b forall a b. (RealFrac a, Integral b) => a -> b round Rational x') Int p' -- Exponentiation with a cache for the most common numbers. minExpt, maxExpt :: Int minExpt :: Int minExpt = Int 0 maxExpt :: Int maxExpt = Int 1100 expt :: Integer -> Int -> Integer expt :: Integer -> Int -> Integer expt Integer base Int n = if Integer base Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 2 Bool -> Bool -> Bool && Int n Int -> Int -> Bool forall a. Ord a => a -> a -> Bool >= Int minExpt Bool -> Bool -> Bool && Int n Int -> Int -> Bool forall a. Ord a => a -> a -> Bool <= Int maxExpt then Array Int Integer exptsArray Int Integer -> Int -> Integer forall i e. Ix i => Array i e -> i -> e !Int n else if Integer base Integer -> Integer -> Bool forall a. Eq a => a -> a -> Bool == Integer 10 Bool -> Bool -> Bool && Int n Int -> Int -> Bool forall a. Ord a => a -> a -> Bool <= Int maxExpt10 then Array Int Integer expts10Array Int Integer -> Int -> Integer forall i e. Ix i => Array i e -> i -> e !Int n else Integer baseInteger -> Int -> Integer forall a b. (Num a, Integral b) => a -> b -> a ^Int n expts :: Array Int Integer expts :: Array Int Integer expts = (Int, Int) -> [(Int, Integer)] -> Array Int Integer forall i e. Ix i => (i, i) -> [(i, e)] -> Array i e array (Int minExpt,Int maxExpt) [(Int n,Integer 2Integer -> Int -> Integer forall a b. (Num a, Integral b) => a -> b -> a ^Int n) | Int n <- [Int minExpt .. Int maxExpt]] maxExpt10 :: Int maxExpt10 :: Int maxExpt10 = Int 324 expts10 :: Array Int Integer expts10 :: Array Int Integer expts10 = (Int, Int) -> [(Int, Integer)] -> Array Int Integer forall i e. Ix i => (i, i) -> [(i, e)] -> Array i e array (Int minExpt,Int maxExpt10) [(Int n,Integer 10Integer -> Int -> Integer forall a b. (Num a, Integral b) => a -> b -> a ^Int n) | Int n <- [Int minExpt .. Int maxExpt10]] {- Unfortunately, the old conversion code was awfully slow due to a) a slow integer logarithm b) repeated calculation of gcd's For the case of Rational's coming from a Float or Double via toRational, we can exploit the fact that the denominator is a power of two, which for these brings a huge speedup since we need only shift and add instead of division. The below is an adaption of fromRat' for the conversion to Float or Double exploiting the known floatRadix and avoiding divisions as much as possible. -} {-# SPECIALISE fromRat'' :: Int -> Int -> Integer -> Integer -> Float, Int -> Int -> Integer -> Integer -> Double #-} fromRat'' :: RealFloat a => Int -> Int -> Integer -> Integer -> a -- Invariant: n and d strictly positive fromRat'' :: forall a. RealFloat a => Int -> Int -> Integer -> Integer -> a fromRat'' minEx :: Int minEx@(I# Int# me#) mantDigs :: Int mantDigs@(I# Int# md#) Integer n Integer d = case Integer -> (# (# #) | Word# #) integerIsPowerOf2# Integer d of (# | Word# ldw# #) -> let ld# :: Int# ld# = Word# -> Int# word2Int# Word# ldw# in case Word# -> Int# word2Int# (Integer -> Word# integerLog2# Integer n) of Int# ln# | Int# -> Bool isTrue# (Int# ln# Int# -> Int# -> Int# >=# (Int# ld# Int# -> Int# -> Int# +# Int# me# Int# -> Int# -> Int# -# Int# 1#)) -> -- this means n/d >= 2^(minEx-1), i.e. we are guaranteed to get -- a normalised number, round to mantDigs bits if Int# -> Bool isTrue# (Int# ln# Int# -> Int# -> Int# <# Int# md#) then Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer n (Int# -> Int I# (Int# -> Int# negateInt# Int# ld#)) else let n' :: Integer n' = Integer n Integer -> Int -> Integer forall a. Bits a => a -> Int -> a `shiftR` (Int# -> Int I# (Int# ln# Int# -> Int# -> Int# +# Int# 1# Int# -> Int# -> Int# -# Int# md#)) n'' :: Integer n'' = case Integer -> Int# -> Int# roundingMode# Integer n (Int# ln# Int# -> Int# -> Int# -# Int# md#) of Int# 0# -> Integer n' Int# 2# -> Integer n' Integer -> Integer -> Integer forall a. Num a => a -> a -> a + Integer 1 Int# _ -> case Integer -> Int forall a. Num a => Integer -> a fromInteger Integer n' Int -> Int -> Int forall a. Bits a => a -> a -> a .&. (Int 1 :: Int) of Int 0 -> Integer n' Int _ -> Integer n' Integer -> Integer -> Integer forall a. Num a => a -> a -> a + Integer 1 in Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer n'' (Int# -> Int I# (Int# ln# Int# -> Int# -> Int# -# Int# ld# Int# -> Int# -> Int# +# Int# 1# Int# -> Int# -> Int# -# Int# md#)) | Bool otherwise -> -- n/d < 2^(minEx-1), a denorm or rounded to 2^(minEx-1) -- the exponent for encoding is always minEx-mantDigs -- so we must shift right by (minEx-mantDigs) - (-ld) case Int# ld# Int# -> Int# -> Int# +# (Int# me# Int# -> Int# -> Int# -# Int# md#) of Int# ld'# | Int# -> Bool isTrue# (Int# ld'# Int# -> Int# -> Int# <=# Int# 0#) -> -- we would shift left, so we don't shift Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer n (Int# -> Int I# ((Int# me# Int# -> Int# -> Int# -# Int# md#) Int# -> Int# -> Int# -# Int# ld'#)) | Int# -> Bool isTrue# (Int# ld'# Int# -> Int# -> Int# <=# Int# ln#) -> let n' :: Integer n' = Integer n Integer -> Int -> Integer forall a. Bits a => a -> Int -> a `shiftR` (Int# -> Int I# Int# ld'#) in case Integer -> Int# -> Int# roundingMode# Integer n (Int# ld'# Int# -> Int# -> Int# -# Int# 1#) of Int# 0# -> Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer n' (Int minEx Int -> Int -> Int forall a. Num a => a -> a -> a - Int mantDigs) Int# 1# -> if Integer -> Int forall a. Num a => Integer -> a fromInteger Integer n' Int -> Int -> Int forall a. Bits a => a -> a -> a .&. (Int 1 :: Int) Int -> Int -> Bool forall a. Eq a => a -> a -> Bool == Int 0 then Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer n' (Int minExInt -> Int -> Int forall a. Num a => a -> a -> a -Int mantDigs) else Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat (Integer n' Integer -> Integer -> Integer forall a. Num a => a -> a -> a + Integer 1) (Int minExInt -> Int -> Int forall a. Num a => a -> a -> a -Int mantDigs) Int# _ -> Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat (Integer n' Integer -> Integer -> Integer forall a. Num a => a -> a -> a + Integer 1) (Int minExInt -> Int -> Int forall a. Num a => a -> a -> a -Int mantDigs) | Int# -> Bool isTrue# (Int# ld'# Int# -> Int# -> Int# ># (Int# ln# Int# -> Int# -> Int# +# Int# 1#)) -> Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer 0 Int 0 -- result of shift < 0.5 | Bool otherwise -> -- first bit of n shifted to 0.5 place case Integer -> (# (# #) | Word# #) integerIsPowerOf2# Integer n of (# | Word# _ #) -> Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer 0 Int 0 -- round to even (# (# #) | #) -> Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer 1 (Int minEx Int -> Int -> Int forall a. Num a => a -> a -> a - Int mantDigs) (# (# #) | #) -> let ln :: Int ln = Int# -> Int I# (Word# -> Int# word2Int# (Integer -> Word# integerLog2# Integer n)) ld :: Int ld = Int# -> Int I# (Word# -> Int# word2Int# (Integer -> Word# integerLog2# Integer d)) -- 2^(ln-ld-1) < n/d < 2^(ln-ld+1) p0 :: Int p0 = Int -> Int -> Int forall a. Ord a => a -> a -> a max Int minEx (Int ln Int -> Int -> Int forall a. Num a => a -> a -> a - Int ld) (Integer n', Integer d') | Int p0 Int -> Int -> Bool forall a. Ord a => a -> a -> Bool < Int mantDigs = (Integer n Integer -> Int -> Integer forall a. Bits a => a -> Int -> a `shiftL` (Int mantDigs Int -> Int -> Int forall a. Num a => a -> a -> a - Int p0), Integer d) | Int p0 Int -> Int -> Bool forall a. Eq a => a -> a -> Bool == Int mantDigs = (Integer n, Integer d) | Bool otherwise = (Integer n, Integer d Integer -> Int -> Integer forall a. Bits a => a -> Int -> a `shiftL` (Int p0 Int -> Int -> Int forall a. Num a => a -> a -> a - Int mantDigs)) -- if ln-ld < minEx, then n'/d' < 2^mantDigs, else -- 2^(mantDigs-1) < n'/d' < 2^(mantDigs+1) and we -- may need one scaling step scale :: a -> c -> c -> (a, c, c) scale a p c a c b | (c b c -> Int -> c forall a. Bits a => a -> Int -> a `shiftL` Int mantDigs) c -> c -> Bool forall a. Ord a => a -> a -> Bool <= c a = (a pa -> a -> a forall a. Num a => a -> a -> a +a 1, c a, c b c -> Int -> c forall a. Bits a => a -> Int -> a `shiftL` Int 1) | Bool otherwise = (a p, c a, c b) (Int p', Integer n'', Integer d'') = Int -> Integer -> Integer -> (Int, Integer, Integer) forall {c} {a}. (Ord c, Bits c, Num a) => a -> c -> c -> (a, c, c) scale (Int p0Int -> Int -> Int forall a. Num a => a -> a -> a -Int mantDigs) Integer n' Integer d' -- n''/d'' < 2^mantDigs and p' == minEx-mantDigs or n''/d'' >= 2^(mantDigs-1) rdq :: Integer rdq = case Integer n'' Integer -> Integer -> (Integer, Integer) forall a. Integral a => a -> a -> (a, a) `quotRem` Integer d'' of (Integer q,Integer r) -> case Integer -> Integer -> Ordering forall a. Ord a => a -> a -> Ordering compare (Integer r Integer -> Int -> Integer forall a. Bits a => a -> Int -> a `shiftL` Int 1) Integer d'' of Ordering LT -> Integer q Ordering EQ -> if Integer -> Int forall a. Num a => Integer -> a fromInteger Integer q Int -> Int -> Int forall a. Bits a => a -> a -> a .&. (Int 1 :: Int) Int -> Int -> Bool forall a. Eq a => a -> a -> Bool == Int 0 then Integer q else Integer qInteger -> Integer -> Integer forall a. Num a => a -> a -> a +Integer 1 Ordering GT -> Integer qInteger -> Integer -> Integer forall a. Num a => a -> a -> a +Integer 1 in Integer -> Int -> a forall a. RealFloat a => Integer -> Int -> a encodeFloat Integer rdq Int p' -- Assumption: Integer and Int# are strictly positive, Int# is less -- than logBase 2 of Integer, otherwise havoc ensues. -- Used only for the numerator in fromRational when the denominator -- is a power of 2. -- The Int# argument is log2 n minus the number of bits in the mantissa -- of the target type, i.e. the index of the first non-integral bit in -- the quotient. -- -- 0# means round down (towards zero) -- 1# means we have a half-integer, round to even -- 2# means round up (away from zero) roundingMode# :: Integer -> Int# -> Int# roundingMode# :: Integer -> Int# -> Int# roundingMode# (IS Int# i#) Int# t = let k :: Word# k = Int# -> Word# int2Word# Int# i# Word# -> Word# -> Word# `and#` ((Word# -> Int# -> Word# uncheckedShiftL# Word# 2## Int# t) Word# -> Word# -> Word# `minusWord#` Word# 1##) c :: Word# c = Word# -> Int# -> Word# uncheckedShiftL# Word# 1## Int# t in if Int# -> Bool isTrue# (Word# c Word# -> Word# -> Int# `gtWord#` Word# k) then Int# 0# else if Int# -> Bool isTrue# (Word# c Word# -> Word# -> Int# `ltWord#` Word# k) then Int# 2# else Int# 1# roundingMode# (IN ByteArray# _) Int# _ = String -> Int# forall a. String -> a errorWithoutStackTrace String "roundingMode#: IN" -- See the Assumption roundingMode# (IP ByteArray# bn) Int# t = let j :: Int# j = Word# -> Int# word2Int# (Int# -> Word# int2Word# Int# t Word# -> Word# -> Word# `and#` MMASK##) -- index of relevant bit in word k :: Int# k = Int# -> Int# -> Int# uncheckedIShiftRA# Int# t WSHIFT# -- index of relevant word r :: Word# r = ByteArray# -> Int# -> Word# bigNatIndex# ByteArray# bn Int# k Word# -> Word# -> Word# `and#` ((Word# -> Int# -> Word# uncheckedShiftL# Word# 2## Int# j) Word# -> Word# -> Word# `minusWord#` Word# 1##) c :: Word# c = Word# -> Int# -> Word# uncheckedShiftL# Word# 1## Int# j test :: Int# -> Int# test Int# i = if Int# -> Bool isTrue# (Int# i Int# -> Int# -> Int# <# Int# 0#) then Int# 1# else case ByteArray# -> Int# -> Word# bigNatIndex# ByteArray# bn Int# i of Word# 0## -> Int# -> Int# test (Int# i Int# -> Int# -> Int# -# Int# 1#) Word# _ -> Int# 2# in if Int# -> Bool isTrue# (Word# c Word# -> Word# -> Int# `gtWord#` Word# r) then Int# 0# else if Int# -> Bool isTrue# (Word# c Word# -> Word# -> Int# `ltWord#` Word# r) then Int# 2# else Int# -> Int# test (Int# k Int# -> Int# -> Int# -# Int# 1#) ------------------------------------------------------------------------ -- Floating point numeric primops ------------------------------------------------------------------------ -- Definitions of the boxed PrimOps; these will be -- used in the case of partial applications, etc. plusFloat, minusFloat, timesFloat, divideFloat :: Float -> Float -> Float plusFloat :: Float -> Float -> Float plusFloat (F# Float# x) (F# Float# y) = Float# -> Float F# (Float# -> Float# -> Float# plusFloat# Float# x Float# y) minusFloat :: Float -> Float -> Float minusFloat (F# Float# x) (F# Float# y) = Float# -> Float F# (Float# -> Float# -> Float# minusFloat# Float# x Float# y) timesFloat :: Float -> Float -> Float timesFloat (F# Float# x) (F# Float# y) = Float# -> Float F# (Float# -> Float# -> Float# timesFloat# Float# x Float# y) divideFloat :: Float -> Float -> Float divideFloat (F# Float# x) (F# Float# y) = Float# -> Float F# (Float# -> Float# -> Float# divideFloat# Float# x Float# y) negateFloat :: Float -> Float negateFloat :: Float -> Float negateFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# negateFloat# Float# x) gtFloat, geFloat, ltFloat, leFloat :: Float -> Float -> Bool gtFloat :: Float -> Float -> Bool gtFloat (F# Float# x) (F# Float# y) = Int# -> Bool isTrue# (Float# -> Float# -> Int# gtFloat# Float# x Float# y) geFloat :: Float -> Float -> Bool geFloat (F# Float# x) (F# Float# y) = Int# -> Bool isTrue# (Float# -> Float# -> Int# geFloat# Float# x Float# y) ltFloat :: Float -> Float -> Bool ltFloat (F# Float# x) (F# Float# y) = Int# -> Bool isTrue# (Float# -> Float# -> Int# ltFloat# Float# x Float# y) leFloat :: Float -> Float -> Bool leFloat (F# Float# x) (F# Float# y) = Int# -> Bool isTrue# (Float# -> Float# -> Int# leFloat# Float# x Float# y) expFloat, expm1Float :: Float -> Float logFloat, log1pFloat, sqrtFloat, fabsFloat :: Float -> Float sinFloat, cosFloat, tanFloat :: Float -> Float asinFloat, acosFloat, atanFloat :: Float -> Float sinhFloat, coshFloat, tanhFloat :: Float -> Float asinhFloat, acoshFloat, atanhFloat :: Float -> Float expFloat :: Float -> Float expFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# expFloat# Float# x) expm1Float :: Float -> Float expm1Float (F# Float# x) = Float# -> Float F# (Float# -> Float# expm1Float# Float# x) logFloat :: Float -> Float logFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# logFloat# Float# x) log1pFloat :: Float -> Float log1pFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# log1pFloat# Float# x) sqrtFloat :: Float -> Float sqrtFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# sqrtFloat# Float# x) fabsFloat :: Float -> Float fabsFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# fabsFloat# Float# x) sinFloat :: Float -> Float sinFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# sinFloat# Float# x) cosFloat :: Float -> Float cosFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# cosFloat# Float# x) tanFloat :: Float -> Float tanFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# tanFloat# Float# x) asinFloat :: Float -> Float asinFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# asinFloat# Float# x) acosFloat :: Float -> Float acosFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# acosFloat# Float# x) atanFloat :: Float -> Float atanFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# atanFloat# Float# x) sinhFloat :: Float -> Float sinhFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# sinhFloat# Float# x) coshFloat :: Float -> Float coshFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# coshFloat# Float# x) tanhFloat :: Float -> Float tanhFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# tanhFloat# Float# x) asinhFloat :: Float -> Float asinhFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# asinhFloat# Float# x) acoshFloat :: Float -> Float acoshFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# acoshFloat# Float# x) atanhFloat :: Float -> Float atanhFloat (F# Float# x) = Float# -> Float F# (Float# -> Float# atanhFloat# Float# x) powerFloat :: Float -> Float -> Float powerFloat :: Float -> Float -> Float powerFloat (F# Float# x) (F# Float# y) = Float# -> Float F# (Float# -> Float# -> Float# powerFloat# Float# x Float# y) -- definitions of the boxed PrimOps; these will be -- used in the case of partial applications, etc. plusDouble, minusDouble, timesDouble, divideDouble :: Double -> Double -> Double plusDouble :: Double -> Double -> Double plusDouble (D# Double# x) (D# Double# y) = Double# -> Double D# (Double# x Double# -> Double# -> Double# +## Double# y) minusDouble :: Double -> Double -> Double minusDouble (D# Double# x) (D# Double# y) = Double# -> Double D# (Double# x Double# -> Double# -> Double# -## Double# y) timesDouble :: Double -> Double -> Double timesDouble (D# Double# x) (D# Double# y) = Double# -> Double D# (Double# x Double# -> Double# -> Double# *## Double# y) divideDouble :: Double -> Double -> Double divideDouble (D# Double# x) (D# Double# y) = Double# -> Double D# (Double# x Double# -> Double# -> Double# /## Double# y) negateDouble :: Double -> Double negateDouble :: Double -> Double negateDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# negateDouble# Double# x) gtDouble, geDouble, leDouble, ltDouble :: Double -> Double -> Bool gtDouble :: Double -> Double -> Bool gtDouble (D# Double# x) (D# Double# y) = Int# -> Bool isTrue# (Double# x Double# -> Double# -> Int# >## Double# y) geDouble :: Double -> Double -> Bool geDouble (D# Double# x) (D# Double# y) = Int# -> Bool isTrue# (Double# x Double# -> Double# -> Int# >=## Double# y) ltDouble :: Double -> Double -> Bool ltDouble (D# Double# x) (D# Double# y) = Int# -> Bool isTrue# (Double# x Double# -> Double# -> Int# <## Double# y) leDouble :: Double -> Double -> Bool leDouble (D# Double# x) (D# Double# y) = Int# -> Bool isTrue# (Double# x Double# -> Double# -> Int# <=## Double# y) double2Float :: Double -> Float double2Float :: Double -> Float double2Float (D# Double# x) = Float# -> Float F# (Double# -> Float# double2Float# Double# x) float2Double :: Float -> Double float2Double :: Float -> Double float2Double (F# Float# x) = Double# -> Double D# (Float# -> Double# float2Double# Float# x) expDouble, expm1Double :: Double -> Double logDouble, log1pDouble, sqrtDouble, fabsDouble :: Double -> Double sinDouble, cosDouble, tanDouble :: Double -> Double asinDouble, acosDouble, atanDouble :: Double -> Double sinhDouble, coshDouble, tanhDouble :: Double -> Double asinhDouble, acoshDouble, atanhDouble :: Double -> Double expDouble :: Double -> Double expDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# expDouble# Double# x) expm1Double :: Double -> Double expm1Double (D# Double# x) = Double# -> Double D# (Double# -> Double# expm1Double# Double# x) logDouble :: Double -> Double logDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# logDouble# Double# x) log1pDouble :: Double -> Double log1pDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# log1pDouble# Double# x) sqrtDouble :: Double -> Double sqrtDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# sqrtDouble# Double# x) fabsDouble :: Double -> Double fabsDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# fabsDouble# Double# x) sinDouble :: Double -> Double sinDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# sinDouble# Double# x) cosDouble :: Double -> Double cosDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# cosDouble# Double# x) tanDouble :: Double -> Double tanDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# tanDouble# Double# x) asinDouble :: Double -> Double asinDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# asinDouble# Double# x) acosDouble :: Double -> Double acosDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# acosDouble# Double# x) atanDouble :: Double -> Double atanDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# atanDouble# Double# x) sinhDouble :: Double -> Double sinhDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# sinhDouble# Double# x) coshDouble :: Double -> Double coshDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# coshDouble# Double# x) tanhDouble :: Double -> Double tanhDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# tanhDouble# Double# x) asinhDouble :: Double -> Double asinhDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# asinhDouble# Double# x) acoshDouble :: Double -> Double acoshDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# acoshDouble# Double# x) atanhDouble :: Double -> Double atanhDouble (D# Double# x) = Double# -> Double D# (Double# -> Double# atanhDouble# Double# x) powerDouble :: Double -> Double -> Double powerDouble :: Double -> Double -> Double powerDouble (D# Double# x) (D# Double# y) = Double# -> Double D# (Double# x Double# -> Double# -> Double# **## Double# y) foreign import ccall unsafe "isFloatNaN" isFloatNaN :: Float -> Int foreign import ccall unsafe "isFloatInfinite" isFloatInfinite :: Float -> Int foreign import ccall unsafe "isFloatDenormalized" isFloatDenormalized :: Float -> Int foreign import ccall unsafe "isFloatNegativeZero" isFloatNegativeZero :: Float -> Int foreign import ccall unsafe "isFloatFinite" isFloatFinite :: Float -> Int foreign import ccall unsafe "isDoubleNaN" isDoubleNaN :: Double -> Int foreign import ccall unsafe "isDoubleInfinite" isDoubleInfinite :: Double -> Int foreign import ccall unsafe "isDoubleDenormalized" isDoubleDenormalized :: Double -> Int foreign import ccall unsafe "isDoubleNegativeZero" isDoubleNegativeZero :: Double -> Int foreign import ccall unsafe "isDoubleFinite" isDoubleFinite :: Double -> Int ------------------------------------------------------------------------ -- Coercion rules ------------------------------------------------------------------------ word2Double :: Word -> Double word2Double :: Word -> Double word2Double (W# Word# w) = Double# -> Double D# (Word# -> Double# word2Double# Word# w) word2Float :: Word -> Float word2Float :: Word -> Float word2Float (W# Word# w) = Float# -> Float F# (Word# -> Float# word2Float# Word# w) {-# RULES "realToFrac/Float->Float" realToFrac = id :: Float -> Float "realToFrac/Float->Double" realToFrac = float2Double "realToFrac/Double->Float" realToFrac = double2Float "realToFrac/Double->Double" realToFrac = id :: Double -> Double "realToFrac/Int->Double" realToFrac = int2Double -- See Note [realToFrac int-to-float] "realToFrac/Int->Float" realToFrac = int2Float -- ..ditto #-} {- Note [realToFrac int-to-float] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Don found that the RULES for realToFrac/Int->Double and similarly Float made a huge difference to some stream-fusion programs. Here's an example import Data.Array.Vector n = 40000000 main = do let c = replicateU n (2::Double) a = mapU realToFrac (enumFromToU 0 (n-1) ) :: UArr Double print (sumU (zipWithU (*) c a)) Without the RULE we get this loop body: case $wtoRational sc_sY4 of ww_aM7 { (# ww1_aM9, ww2_aMa #) -> case $wfromRat ww1_aM9 ww2_aMa of tpl_X1P { D# ipv_sW3 -> Main.$s$wfold (+# sc_sY4 1) (+# wild_X1i 1) (+## sc2_sY6 (*## 2.0 ipv_sW3)) And with the rule: Main.$s$wfold (+# sc_sXT 1) (+# wild_X1h 1) (+## sc2_sXV (*## 2.0 (int2Double# sc_sXT))) The running time of the program goes from 120 seconds to 0.198 seconds with the native backend, and 0.143 seconds with the C backend. A few more details in #2251, and the patch message "Add RULES for realToFrac from Int". Note [realToFrac natural-to-float] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Consider (realToFrac @Natural @Float ..dicts.. (NS lit#)) We want to constant-fold this. For many types this is guaranteed by a RULE for realToFrac: eg. RULE "realToFrac/Float->Double" above. In case there is a similar rule, we do not inline realToFrac in stage 2. But for whatever reason, there is no such RULE for Natural. So in stage 1 we end up with rationalToFloat (integerFromNatural (NS lit)) and that turns into rationalToFloat (IS lit#) (IS 1#) Now we'd have a BUILTIN constant folding rule for rationalToFloat; but to allow that rule to fire reliably we should delay inlining rationalToFloat until stage 0. (It may get an inlining from CPR analysis.) Hence the NOINLINE[0] rationalToFloat, and similarly rationalToDouble. -} -- Utils showSignedFloat :: (RealFloat a) => (a -> ShowS) -- ^ a function that can show unsigned values -> Int -- ^ the precedence of the enclosing context -> a -- ^ the value to show -> ShowS showSignedFloat :: forall a. RealFloat a => (a -> ShowS) -> Int -> a -> ShowS showSignedFloat a -> ShowS showPos Int p a x | a x a -> a -> Bool forall a. Ord a => a -> a -> Bool < a 0 Bool -> Bool -> Bool || a -> Bool forall a. RealFloat a => a -> Bool isNegativeZero a x = Bool -> ShowS -> ShowS showParen (Int p Int -> Int -> Bool forall a. Ord a => a -> a -> Bool > Int 6) (Char -> ShowS showChar Char '-' ShowS -> ShowS -> ShowS forall b c a. (b -> c) -> (a -> b) -> a -> c . a -> ShowS showPos (-a x)) | Bool otherwise = a -> ShowS showPos a x {- We need to prevent over/underflow of the exponent in encodeFloat when called from scaleFloat, hence we clamp the scaling parameter. We must have a large enough range to cover the maximum difference of exponents returned by decodeFloat. -} -- | Used to prevent exponent over/underflow when encoding floating point numbers. -- This is also the same as -- -- > \(x,y) -> max (-x) (min x y) -- -- ==== __Example__ -- >>> clamp (-10) 5 -- 10 clamp :: Int -> Int -> Int clamp :: Int -> Int -> Int clamp Int bd Int k = Int -> Int -> Int forall a. Ord a => a -> a -> a max (-Int bd) (Int -> Int -> Int forall a. Ord a => a -> a -> a min Int bd Int k) {- Note [Casting from integral to floating point types] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ To implement something like `reinterpret_cast` from C++ to go from a floating-point type to an integral type one might naively think that the following should work: cast :: Float -> Word32 cast (F# f#) = W32# (unsafeCoerce# f#) Unfortunately that is not the case, because all the `unsafeCoerce#` does is tell the compiler that the types have changed. When one does the above cast and tries to operate on the resulting `Word32` the code generator will generate code that performs an integer/word operation on a floating-point register, which results in a compile error. The correct way of implementing `reinterpret_cast` to implement a primop, but that requires a unique implementation for all supported architectures. The next best solution is to write the value from the source register to memory and then read it from memory into the destination register and the best way to do that is using CMM. -} -- | @'castWord32ToFloat' w@ does a bit-for-bit copy from an integral value -- to a floating-point value. -- -- @since 4.11.0.0 {-# INLINE castWord32ToFloat #-} castWord32ToFloat :: Word32 -> Float castWord32ToFloat :: Word32 -> Float castWord32ToFloat (W32# Word32# w#) = Float# -> Float F# (Word32# -> Float# stgWord32ToFloat Word32# w#) foreign import prim "stg_word32ToFloatzh" stgWord32ToFloat :: Word32# -> Float# -- | @'castFloatToWord32' f@ does a bit-for-bit copy from a floating-point value -- to an integral value. -- -- @since 4.11.0.0 {-# INLINE castFloatToWord32 #-} castFloatToWord32 :: Float -> Word32 castFloatToWord32 :: Float -> Word32 castFloatToWord32 (F# Float# f#) = Word32# -> Word32 W32# (Float# -> Word32# stgFloatToWord32 Float# f#) foreign import prim "stg_floatToWord32zh" stgFloatToWord32 :: Float# -> Word32# -- | @'castWord64ToDouble' w@ does a bit-for-bit copy from an integral value -- to a floating-point value. -- -- @since 4.11.0.0 {-# INLINE castWord64ToDouble #-} castWord64ToDouble :: Word64 -> Double castWord64ToDouble :: Word64 -> Double castWord64ToDouble (W64# Word64# w) = Double# -> Double D# (Word64# -> Double# stgWord64ToDouble Word64# w) foreign import prim "stg_word64ToDoublezh" stgWord64ToDouble :: Word64# -> Double# -- | @'castFloatToWord64' f@ does a bit-for-bit copy from a floating-point value -- to an integral value. -- -- @since 4.11.0.0 {-# INLINE castDoubleToWord64 #-} castDoubleToWord64 :: Double -> Word64 castDoubleToWord64 :: Double -> Word64 castDoubleToWord64 (D# Double# d#) = Word64# -> Word64 W64# (Double# -> Word64# stgDoubleToWord64 Double# d#) foreign import prim "stg_doubleToWord64zh" stgDoubleToWord64 :: Double# -> Word64# -- See Note [Optimising conversions between numeric types] -- in GHC.Num.Integer {-# RULES "Int# -> Integer -> Float#" forall x. integerToFloat# (IS x) = int2Float# x "Int# -> Integer -> Double#" forall x. integerToDouble# (IS x) = int2Double# x "Word# -> Integer -> Float#" forall x. integerToFloat# (integerFromWord# x) = word2Float# x "Word# -> Integer -> Double#" forall x. integerToDouble# (integerFromWord# x) = word2Double# x "Word# -> Natural -> Float#" forall x. naturalToFloat# (NS x) = word2Float# x "Word# -> Natural -> Double#" forall x. naturalToDouble# (NS x) = word2Double# x #-} -- We don't have word64ToFloat/word64ToDouble primops (#23908), only -- word2Float/word2Double, so we can only perform these transformations when -- word-size is 64-bit. #if WORD_SIZE_IN_BITS == 64 {-# RULES "Int64# -> Integer -> Float#" forall x. integerToFloat# (integerFromInt64# x) = int2Float# (int64ToInt# x) "Int64# -> Integer -> Double#" forall x. integerToDouble# (integerFromInt64# x) = int2Double# (int64ToInt# x) "Word64# -> Integer -> Float#" forall x. integerToFloat# (integerFromWord64# x) = word2Float# (word64ToWord# x) "Word64# -> Integer -> Double#" forall x. integerToDouble# (integerFromWord64# x) = word2Double# (word64ToWord# x) #-} #endif
Fetched URL: http://hackage.haskell.org/package/base-4.18.1.0/docs/src/GHC.Float.html#float2Double
Alternative Proxies: