[Haskell-cafe] Performance question

[Ketil Malde]

haskell at kudling.de writes:

> Profiling says that the majority of the time is spend in "main". But i have no idea where.
> Can someone give me a hint?

Yes.  Lots of them, but somehow, I suspect nobody tried your code.

> COST CENTRE              MODULE                                               no.    entries  %time %alloc   %time %alloc
>
> MAIN                     MAIN                                                   1           0   0.0    0.0   100.0  100.0
>  main                    Main                                                 254           1  88.1   90.8   100.0  100.0
>   monteCarloPi           Main                                                 255           1   0.6    1.1    11.9    9.2
>    pairs                 Main                                                 257    10000000   0.7    1.4     0.7    1.4
>    countHits             Main                                                 256    10000001   4.2    2.9    10.6    6.7
>     accumulateHit        Main                                                 258    27852236   3.0    2.3     6.4    3.8
>      isInCircle          Main                                                 259    30000000   3.3    1.5     3.3    1.5
>  CAF:lit_r1A7            Main                                                 248           1   0.0    0.0     0.0    0.0
>   isInCircle             Main                                                 260           0   0.0    0.0     0.0    0.0

Thomas van Noort:

> First thing I noticed, how about removing the sqrt in isInCircle:

I did this.  The result was the same - the sqrt doesn't matter, and
perhaps it even gets optimized away?

> The first thing I would do is replace your
> isInCircle :: (Floating a, Ord a)  => (a,a) -> Bool
> with
> isInCircle :: (Double, Double) -> Bool

Then I did that.  The result was still the same - I bet GHC is smart
enought to specialize this on its own.

The intresting thing about your profile is that all the time is spent
in 'main'.  Now, why would that be?  I refactored a bit, and in
partiuclar wrote this function:

  mkRandoms :: IO [Double]
  mkRandoms = do
	randomNumberGenerator <- getStdGen
        return $ randoms randomNumberGenerator

Here's the new profile (10^6 iterations):

  COST CENTRE                    MODULE               %time %alloc

  mkRandoms                      Main                  96.1   96.0
  accumulateHit                  Main                   1.6    1.7
  pairs                          Main                   1.2    1.3
  monteCarloPi                   Main                   0.2    1.0

So it seems we're just tremendously lousy at generating random
Doubles.  Eliminating this bottleneck, the remaining 3.9% would put it
at about 2x the C++ performance.

-k
-- 
If I haven't seen further, it is by standing in the footprints of giants