Hello everyone!

I am trying to find out the execution time of mergesort for a list size of 1 million integers. I have done the same in Erlang as well and the execution time in Erlang was around 0.93 seconds. I have implemented the program in almost the same way in Haskell as well but for some reason the Haskell implementation is taking around 12 seconds which doesn't seem right.

Here is the implementation in Haskell:

{-# LANGUAGE OverloadedStrings #-}

{-# LANGUAGE BangPatterns #-}

import Control.Exception

import Formatting

import Formatting.Clock

import System.Clock

import Control.DeepSeq

mergesort [] = []

mergesort [x] = [x]

mergesort xs = let (lhalf, rhalf) = splitAt (length xs `div` 2) xs

in merge' (mergesort lhalf) (mergesort rhalf)

merge' lhalf rhalf = merge lhalf rhalf []

merge [] [] acc = reverse acc

merge [] y acc = reverse acc ++ y

merge x [] acc = reverse acc ++ x

merge (l:ls) (r:rs) acc

| l < r = merge ls (r:rs) (l:acc)

| otherwise = merge rs (l:ls) (r:acc)

toList :: String -> [Integer]

toList input = read ("[" ++ input ++ "]")

main = do

file <- getLine

contents <- readFile file

let !unsortedlist = (toList contents)

start <- getTime Monotonic

evaluate(force (mergesort unsortedlist))

end <- getTime Monotonic

fprint (timeSpecs % "\n") start end

What am I doing wrong?