1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364
|
-----------------------------------------------------------------------------
-- |
-- Module : Data.SBV.Core.Sized
-- Copyright : (c) Levent Erkok
-- License : BSD3
-- Maintainer: erkokl@gmail.com
-- Stability : experimental
--
-- Type-level sized floats.
-----------------------------------------------------------------------------
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE FlexibleInstances #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE TypeApplications #-}
{-# LANGUAGE TypeFamilies #-}
{-# LANGUAGE UndecidableInstances #-}
{-# OPTIONS_GHC -Wall -Werror #-}
module Data.SBV.Core.SizedFloats (
-- * Type-sized floats
FloatingPoint(..), FP(..), FPHalf, FPBFloat, FPSingle, FPDouble, FPQuad
-- * Constructing values
, fpFromRawRep, fpFromBigFloat, fpNaN, fpInf, fpZero
-- * Operations
, fpFromInteger, fpFromRational, fpFromFloat, fpFromDouble, fpEncodeFloat
-- * Internal operations
, fprCompareObject, fprToSMTLib2, mkBFOpts, bfToString, bfRemoveRedundantExp
) where
import Data.Char (intToDigit)
import Data.List (isSuffixOf)
import Data.Proxy
import GHC.TypeLits
import Data.Bits
import Data.Ratio
import Numeric
import Data.SBV.Core.Kind
import Data.SBV.Utils.Numeric (floatToWord)
import LibBF (BigFloat, BFOpts, RoundMode, Status)
import qualified LibBF as BF
-- | A floating point value, indexed by its exponent and significand sizes.
--
-- An IEEE SP is @FloatingPoint 8 24@
-- DP is @FloatingPoint 11 53@
-- etc.
newtype FloatingPoint (eb :: Nat) (sb :: Nat) = FloatingPoint FP
deriving (Eq, Ord)
-- | Abbreviation for IEEE half precision float, bit width 16 = 5 + 11.
type FPHalf = FloatingPoint 5 11
-- | Abbreviation for brain-float precision float, bit width 16 = 8 + 8.
type FPBFloat = FloatingPoint 8 8
-- | Abbreviation for IEEE single precision float, bit width 32 = 8 + 24.
type FPSingle = FloatingPoint 8 24
-- | Abbreviation for IEEE double precision float, bit width 64 = 11 + 53.
type FPDouble = FloatingPoint 11 53
-- | Abbreviation for IEEE quadruble precision float, bit width 128 = 15 + 113.
type FPQuad = FloatingPoint 15 113
-- | Show instance for Floats. By default we print in base 10, with standard scientific notation.
instance Show (FloatingPoint eb sb) where
show (FloatingPoint r) = show r
-- | Internal representation of a parameterized float.
--
-- A note on cardinality: If we have eb exponent bits, and sb significand bits,
-- then the total number of floats is 2^sb*(2^eb-1) + 3: All exponents except 11..11
-- is allowed. So we get, 2^eb-1, different combinations, each with a sign, giving
-- us 2^sb*(2^eb-1) totals. Then we have two infinities, and one NaN, adding 3 more.
data FP = FP { fpExponentSize :: Int
, fpSignificandSize :: Int
, fpValue :: BigFloat
}
deriving (Ord, Eq)
instance Show FP where
show = bfRemoveRedundantExp . bfToString 10 False False
-- | Remove redundant p+0 etc.
bfRemoveRedundantExp :: String -> String
bfRemoveRedundantExp v = walk useless
where walk [] = v
walk (s:ss)
| s `isSuffixOf` v = reverse . drop (length s) . reverse $ v
| True = walk ss
-- these suffixes are useless, drop them
useless = [c : s ++ "0" | c <- "pe@", s <- ["+", "-", ""]]
-- | Show a big float in the base given.
-- NB. Do not be tempted to use BF.showFreeMin below; it produces arguably correct
-- but very confusing results. See <https://github.com/GaloisInc/cryptol/issues/1089>
-- for a discussion of the issues.
bfToString :: Int -> Bool -> Bool -> FP -> String
bfToString b withPrefix forceExponent (FP _ sb a)
| BF.bfIsNaN a = "NaN"
| BF.bfIsInf a = if BF.bfIsPos a then "Infinity" else "-Infinity"
| BF.bfIsZero a = if BF.bfIsPos a then "0.0" else "-0.0"
| True = trimZeros $ BF.bfToString b opts' a
where opts = BF.showRnd BF.NearEven <> BF.showFree (Just (fromIntegral sb))
opts' = case (withPrefix, forceExponent) of
(False, False) -> opts
(False, True ) -> BF.forceExp <> opts
(True, False) -> BF.addPrefix <> opts
(True, True ) -> BF.addPrefix <> BF.forceExp <> opts
-- In base 10, exponent starts with 'e'. Otherwise (2, 8, 16) it starts with 'p'
expChar = if b == 10 then 'e' else 'p'
trimZeros s
| '.' `elem` s = case span (/= expChar) s of
(pre, post) -> let pre' = reverse $ case dropWhile (== '0') $ reverse pre of
res@('.':_) -> '0' : res
res -> res
in pre' ++ post
| True = s
-- | Default options for BF options.
mkBFOpts :: Integral a => a -> a -> RoundMode -> BFOpts
mkBFOpts eb sb rm = BF.allowSubnormal <> BF.rnd rm <> BF.expBits (fromIntegral eb) <> BF.precBits (fromIntegral sb)
-- | Construct a float, by appropriately rounding
fpFromBigFloat :: Int -> Int -> BigFloat -> FP
fpFromBigFloat eb sb r = FP eb sb $ fst $ BF.bfRoundFloat (mkBFOpts eb sb BF.NearEven) r
-- | Convert from an sign/exponent/mantissa representation to a float. The values are the integers
-- representing the bit-patterns of these values, i.e., the raw representation. We assume that these
-- integers fit into the ranges given, i.e., no overflow checking is done here.
fpFromRawRep :: Bool -> (Integer, Int) -> (Integer, Int) -> FP
fpFromRawRep sign (e, eb) (s, sb) = FP eb sb $ BF.bfFromBits (mkBFOpts eb sb BF.NearEven) val
where es, val :: Integer
es = (e `shiftL` (sb - 1)) .|. s
val | sign = (1 `shiftL` (eb + sb - 1)) .|. es
| True = es
-- | Make NaN. Exponent is all 1s. Significand is non-zero. The sign is irrelevant.
fpNaN :: Int -> Int -> FP
fpNaN eb sb = fpFromBigFloat eb sb BF.bfNaN
-- | Make Infinity. Exponent is all 1s. Significand is 0.
fpInf :: Bool -> Int -> Int -> FP
fpInf sign eb sb = fpFromBigFloat eb sb $ if sign then BF.bfNegInf else BF.bfPosInf
-- | Make a signed zero.
fpZero :: Bool -> Int -> Int -> FP
fpZero sign eb sb = fpFromBigFloat eb sb $ if sign then BF.bfNegZero else BF.bfPosZero
-- | Make from an integer value.
fpFromInteger :: Int -> Int -> Integer -> FP
fpFromInteger eb sb iv = fpFromBigFloat eb sb $ BF.bfFromInteger iv
-- | Make a generalized floating-point value from a 'Rational'.
fpFromRational :: Int -> Int -> Rational -> FP
fpFromRational eb sb r = FP eb sb $ fst $ BF.bfDiv (mkBFOpts eb sb BF.NearEven) (BF.bfFromInteger (numerator r))
(BF.bfFromInteger (denominator r))
-- | Represent the FP in SMTLib2 format
fprToSMTLib2 :: FP -> String
fprToSMTLib2 (FP eb sb r)
| BF.bfIsNaN r = as "NaN"
| BF.bfIsInf r = as $ if BF.bfIsPos r then "+oo" else "-oo"
| BF.bfIsZero r = as $ if BF.bfIsPos r then "+zero" else "-zero"
| True = generic
where e = show eb
s = show sb
bits = BF.bfToBits (mkBFOpts eb sb BF.NearEven) r
significandMask = (1 :: Integer) `shiftL` (sb - 1) - 1
exponentMask = (1 :: Integer) `shiftL` eb - 1
fpSign = bits `testBit` (eb + sb - 1)
fpExponent = (bits `shiftR` (sb - 1)) .&. exponentMask
fpSignificand = bits .&. significandMask
generic = "(fp " ++ unwords [if fpSign then "#b1" else "#b0", mkB eb fpExponent, mkB (sb - 1) fpSignificand] ++ ")"
as x = "(_ " ++ x ++ " " ++ e ++ " " ++ s ++ ")"
mkB sz val = "#b" ++ pad sz (showIntAtBase 2 intToDigit val "")
pad l str = replicate (l - length str) '0' ++ str
-- | Structural comparison only, for internal map indexes
fprCompareObject :: FP -> FP -> Ordering
fprCompareObject (FP eb sb a) (FP eb' sb' b) = case (eb, sb) `compare` (eb', sb') of
LT -> LT
GT -> GT
EQ -> a `BF.bfCompare` b
-- | Compute the signum of a big float
bfSignum :: BigFloat -> BigFloat
bfSignum r | BF.bfIsNaN r = r
| BF.bfIsZero r = r
| BF.bfIsPos r = BF.bfFromInteger 1
| True = BF.bfFromInteger (-1)
-- | Num instance for big-floats
instance Num FP where
(+) = lift2 BF.bfAdd
(-) = lift2 BF.bfSub
(*) = lift2 BF.bfMul
abs = lift1 BF.bfAbs
signum = lift1 bfSignum
fromInteger = error "FP.fromInteger: Not supported for arbitrary floats. Use fpFromInteger instead, specifying the precision"
negate = lift1 BF.bfNeg
-- | Fractional instance for big-floats
instance Fractional FP where
fromRational = error "FP.fromRational: Not supported for arbitrary floats. Use fpFromRational instead, specifying the precision"
(/) = lift2 BF.bfDiv
-- | Floating instance for big-floats
instance Floating FP where
sqrt (FP eb sb a) = FP eb sb $ fst $ BF.bfSqrt (mkBFOpts eb sb BF.NearEven) a
FP eb sb a ** FP _ _ b = FP eb sb $ fst $ BF.bfPow (mkBFOpts eb sb BF.NearEven) a b
pi = unsupported "Floating.FP.pi"
exp = unsupported "Floating.FP.exp"
log = unsupported "Floating.FP.log"
sin = unsupported "Floating.FP.sin"
cos = unsupported "Floating.FP.cos"
tan = unsupported "Floating.FP.tan"
asin = unsupported "Floating.FP.asin"
acos = unsupported "Floating.FP.acos"
atan = unsupported "Floating.FP.atan"
sinh = unsupported "Floating.FP.sinh"
cosh = unsupported "Floating.FP.cosh"
tanh = unsupported "Floating.FP.tanh"
asinh = unsupported "Floating.FP.asinh"
acosh = unsupported "Floating.FP.acosh"
atanh = unsupported "Floating.FP.atanh"
-- | Real-float instance for big-floats. Beware! Some of these aren't really all that well tested.
instance RealFloat FP where
floatRadix _ = 2
floatDigits (FP _ sb _) = sb
floatRange (FP eb _ _) = (fromIntegral (-v+3), fromIntegral v)
where v :: Integer
v = 2 ^ ((fromIntegral eb :: Integer) - 1)
isNaN (FP _ _ r) = BF.bfIsNaN r
isInfinite (FP _ _ r) = BF.bfIsInf r
isDenormalized (FP eb sb r) = BF.bfIsSubnormal (mkBFOpts eb sb BF.NearEven) r
isNegativeZero (FP _ _ r) = BF.bfIsZero r && BF.bfIsNeg r
isIEEE _ = True
decodeFloat i@(FP _ _ r) = case BF.bfToRep r of
BF.BFNaN -> decodeFloat (0/0 :: Double)
BF.BFRep s n -> case n of
BF.Zero -> (0, 0)
BF.Inf -> let (_, m) = floatRange i
x = (2 :: Integer) ^ toInteger (m+1)
in (if s == BF.Neg then -x else x, 0)
BF.Num x y -> -- The value here is x * 2^y
(if s == BF.Neg then -x else x, fromIntegral y)
encodeFloat = error "FP.encodeFloat: Not supported for arbitrary floats. Use fpEncodeFloat instead, specifying the precision"
-- | Encode from exponent/mantissa form to a float representation. Corresponds to 'encodeFloat' in Haskell.
fpEncodeFloat :: Int -> Int -> Integer -> Int -> FP
fpEncodeFloat eb sb m n | n < 0 = fpFromRational eb sb (m % n')
| True = fpFromRational eb sb (m * n' % 1)
where n' :: Integer
n' = (2 :: Integer) ^ abs (fromIntegral n :: Integer)
-- | Real instance for big-floats. Beware, not that well tested!
instance Real FP where
toRational i
| n >= 0 = m * 2 ^ n % 1
| True = m % 2 ^ abs n
where (m, n) = decodeFloat i
-- | Real-frac instance for big-floats. Beware, not that well tested!
instance RealFrac FP where
properFraction (FP eb sb r) = case BF.bfRoundInt BF.ToNegInf r of
(r', BF.Ok) | BF.bfSign r == BF.bfSign r' -> (getInt r', FP eb sb r - FP eb sb r')
x -> error $ "RealFrac.FP.properFraction: Failed to convert: " ++ show (r, x)
where getInt x = case BF.bfToRep x of
BF.BFNaN -> error $ "Data.SBV.FloatingPoint.properFraction: Failed to convert: " ++ show (r, x)
BF.BFRep s n -> case n of
BF.Zero -> 0
BF.Inf -> error $ "Data.SBV.FloatingPoint.properFraction: Failed to convert: " ++ show (r, x)
BF.Num v y -> -- The value here is x * 2^y, and is integer if y >= 0
let e :: Integer
e = 2 ^ (fromIntegral y :: Integer)
sgn = if s == BF.Neg then ((-1) *) else id
in if y > 0
then fromIntegral $ sgn $ v * e
else fromIntegral $ sgn v
-- | Num instance for FloatingPoint
instance ValidFloat eb sb => Num (FloatingPoint eb sb) where
FloatingPoint a + FloatingPoint b = FloatingPoint $ a + b
FloatingPoint a * FloatingPoint b = FloatingPoint $ a * b
abs (FloatingPoint fp) = FloatingPoint (abs fp)
signum (FloatingPoint fp) = FloatingPoint (signum fp)
negate (FloatingPoint fp) = FloatingPoint (negate fp)
fromInteger = FloatingPoint . fpFromInteger (intOfProxy (Proxy @eb)) (intOfProxy (Proxy @sb))
instance ValidFloat eb sb => Fractional (FloatingPoint eb sb) where
fromRational = FloatingPoint . fpFromRational (intOfProxy (Proxy @eb)) (intOfProxy (Proxy @sb))
FloatingPoint a / FloatingPoint b = FloatingPoint (a / b)
unsupported :: String -> a
unsupported w = error $ "Data.SBV.FloatingPoint: Unsupported operation: " ++ w ++ ". Please request this as a feature!"
-- Float instance. Most methods are left unimplemented.
instance ValidFloat eb sb => Floating (FloatingPoint eb sb) where
pi = FloatingPoint pi
exp (FloatingPoint i) = FloatingPoint (exp i)
sqrt (FloatingPoint i) = FloatingPoint (sqrt i)
FloatingPoint a ** FloatingPoint b = FloatingPoint $ a ** b
log (FloatingPoint i) = FloatingPoint (log i)
sin (FloatingPoint i) = FloatingPoint (sin i)
cos (FloatingPoint i) = FloatingPoint (cos i)
tan (FloatingPoint i) = FloatingPoint (tan i)
asin (FloatingPoint i) = FloatingPoint (asin i)
acos (FloatingPoint i) = FloatingPoint (acos i)
atan (FloatingPoint i) = FloatingPoint (atan i)
sinh (FloatingPoint i) = FloatingPoint (sinh i)
cosh (FloatingPoint i) = FloatingPoint (cosh i)
tanh (FloatingPoint i) = FloatingPoint (tanh i)
asinh (FloatingPoint i) = FloatingPoint (asinh i)
acosh (FloatingPoint i) = FloatingPoint (acosh i)
atanh (FloatingPoint i) = FloatingPoint (atanh i)
-- | Lift a unary operation, simple case of function with no status. Here, we call fpFromBigFloat since the big-float isn't size aware.
lift1 :: (BigFloat -> BigFloat) -> FP -> FP
lift1 f (FP eb sb a) = fpFromBigFloat eb sb $ f a
-- Lift a binary operation. Here we don't call fpFromBigFloat, because the result is correctly rounded.
lift2 :: (BFOpts -> BigFloat -> BigFloat -> (BigFloat, Status)) -> FP -> FP -> FP
lift2 f (FP eb sb a) (FP _ _ b) = FP eb sb $ fst $ f (mkBFOpts eb sb BF.NearEven) a b
-- | Convert from a IEEE float.
fpFromFloat :: Int -> Int -> Float -> FP
fpFromFloat 8 24 f = let fw = floatToWord f
(sgn, e, s) = (fw `testBit` 31, fromIntegral (fw `shiftR` 23) .&. 0xFF, fromIntegral fw .&. 0x7FFFFF)
in fpFromRawRep sgn (e, 8) (s, 24)
fpFromFloat eb sb f = error $ "SBV.fprFromFloat: Unexpected input: " ++ show (eb, sb, f)
-- | Convert from a IEEE double.
fpFromDouble :: Int -> Int -> Double -> FP
fpFromDouble 11 53 d = FP 11 54 $ BF.bfFromDouble d
fpFromDouble eb sb d = error $ "SBV.fprFromDouble: Unexpected input: " ++ show (eb, sb, d)
|