[Haskell-cafe] How to interprete profile output

Sebastian Reese str at holumbus.org
Fri May 29 10:47:42 EDT 2009

Hi there,

i am currently working on the mapreduce framework  
(http://darcs2.fh-wedel.de/repos/holumbus/) which can also be found in  

I was doing the simple WordFrequency example, which does nothing but  
counting word frequencies and runs a few seconds.

When I run this program with ghc 6.10 and profiling option "+RTS -p  
-RTS" i get this (http://pastebin.com/f5163325) output.
The interesting line is the one with "MAIN MAIN" in it.

1. Where does this MAIN come from? My main function is right under it.  
So is it some GHC profiling wrapper?

2. Why does it use over 90% of cpu time and 25% of allocated memory?  
So i know you can't tell me exactly why, but maybe someone can give me  
a hint. The mapreduce framework does a lot with network, IO, forkIO,  
MVars etc.. Is that maybe a reason for it? My guess is that there is  
some action, that can't be assinged to a specific cost centre and/or  
results from an empty MVar.

thanks for your help

This message was sent using IMP, the Internet Messaging Program.

More information about the Haskell-Cafe mailing list