# Difference between revisions of "Introduction/Direct Translation"

(+faster qsort) |
|||

Line 72: | Line 72: | ||

This uses various extensions to make the types ridiculously general, but the actual algorithm (quickSort) is plain Haskell. |
This uses various extensions to make the types ridiculously general, but the actual algorithm (quickSort) is plain Haskell. |
||

+ | |||

+ | A more specific/direct translation (neither this nor the C version is [[polymorphic]]) is offered by [http://www.haskell.org/pipermail/haskell-cafe/2009-August/065269.html Daniel Fischer], who reports that this version runs within 2x of the C version: |
||

+ | |||

+ | <haskell> |
||

+ | import Data.Array.Base (unsafeRead, unsafeWrite) |
||

+ | import Data.Array.ST |
||

+ | import Control.Monad.ST |
||

+ | |||

+ | myqsort :: STUArray s Int Int -> Int -> Int -> ST s () |
||

+ | myqsort a lo hi |
||

+ | | lo < hi = do |
||

+ | let lscan p h i |
||

+ | | i < h = do |
||

+ | v <- unsafeRead a i |
||

+ | if p < v then return i else lscan p h (i+1) |
||

+ | | otherwise = return i |
||

+ | rscan p l i |
||

+ | | l < i = do |
||

+ | v <- unsafeRead a i |
||

+ | if v < p then return i else rscan p l (i-1) |
||

+ | | otherwise = return i |
||

+ | swap i j = do |
||

+ | v <- unsafeRead a i |
||

+ | unsafeRead a j >>= unsafeWrite a i |
||

+ | unsafeWrite a j v |
||

+ | sloop p l h |
||

+ | | l < h = do |
||

+ | l1 <- lscan p h l |
||

+ | h1 <- rscan p l1 h |
||

+ | if (l1 < h1) then (swap l1 h1 >> sloop p l1 h1) else return l1 |
||

+ | | otherwise = return l |
||

+ | piv <- unsafeRead a hi |
||

+ | i <- sloop piv lo hi |
||

+ | swap i hi |
||

+ | myqsort a lo (i-1) |
||

+ | myqsort a (i+1) hi |
||

+ | | otherwise = return () |
||

+ | </haskell> |

## Latest revision as of 08:07, 16 August 2009

The quicksort quoted in Introduction isn't the "real" quicksort and doesn't scale for longer lists like the c code does.

http://programming.reddit.com/info/5yutf/comments/

Here are some points to how the "real" quicksort would look in haskell.

Lennart Augustsson has a quicksort entry on his blog which is pure (no unsafe):

http://augustss.blogspot.com/2007/08/quicksort-in-haskell-quicksort-is.html

Another version (uses System.IO.Unsafe), is below.

There is also a "parallel" quicksort at

http://www.macs.hw.ac.uk/~dsg/gph/nofib/

roconnor claims that in haskell the "real" quicksort is really a treesort:

http://programming.reddit.com/info/2h0j2/comments

Unfortunately none of the above "real" quicksorts seems to compile as given, when copy/pasted into ghci. Can someone fix? The "parallel" quicksort gave error "unknown package concurrent" when I ran make in quicksort/gransim.

Has anyone got a functioning "real" quicksort that works on copy/paste?

The program below is working very very slowly. It's probably slowsort... :o)

```
import Control.Monad (when)
import Control.Monad.ST
import Data.Array.ST
import Data.Array.IArray
import Data.Array.MArray
qsort :: (IArray a e, Ix i, Enum i, Ord e) => a i e -> a i e
qsort arr = processArray quickSort arr
processArray :: (IArray a e, IArray b e, Ix i)
=> (forall s. (STArray s) i e -> ST s ()) -> a i e -> b i e
processArray f (arr :: a i e) = runST $ do
arr' <- thaw arr :: ST s (STArray s i e)
f arr'
unsafeFreeze arr'
quickSort :: (MArray a e m, Ix i, Enum i, Ord e) => a i e -> m ()
quickSort arr = qsort' =<< getBounds arr
where
qsort' (lo, hi) | lo >= hi = return ()
| otherwise = do
p <- readArray arr hi
l <- mainLoop p lo hi
swap l hi
qsort' (lo, pred l)
qsort' (succ l, hi)
mainLoop p l h | l >= h = return l
| otherwise = do
l' <- doTil (\l' b -> l' < h && b <= p) succ l
h' <- doTil (\h' b -> h' > l' && b >= p) pred h
when (l' < h') $ do
swap l' h'
mainLoop p l' h'
doTil p op ix = do
b <- readArray arr ix
if p ix b then doTil p op (op ix) else return ix
swap xi yi = do
x <- readArray arr xi
readArray arr yi >>= writeArray arr xi
writeArray arr yi x
```

This uses various extensions to make the types ridiculously general, but the actual algorithm (quickSort) is plain Haskell.

A more specific/direct translation (neither this nor the C version is polymorphic) is offered by Daniel Fischer, who reports that this version runs within 2x of the C version:

```
import Data.Array.Base (unsafeRead, unsafeWrite)
import Data.Array.ST
import Control.Monad.ST
myqsort :: STUArray s Int Int -> Int -> Int -> ST s ()
myqsort a lo hi
| lo < hi = do
let lscan p h i
| i < h = do
v <- unsafeRead a i
if p < v then return i else lscan p h (i+1)
| otherwise = return i
rscan p l i
| l < i = do
v <- unsafeRead a i
if v < p then return i else rscan p l (i-1)
| otherwise = return i
swap i j = do
v <- unsafeRead a i
unsafeRead a j >>= unsafeWrite a i
unsafeWrite a j v
sloop p l h
| l < h = do
l1 <- lscan p h l
h1 <- rscan p l1 h
if (l1 < h1) then (swap l1 h1 >> sloop p l1 h1) else return l1
| otherwise = return l
piv <- unsafeRead a hi
i <- sloop piv lo hi
swap i hi
myqsort a lo (i-1)
myqsort a (i+1) hi
| otherwise = return ()
```