Profiling parsers written using parser combinator libraries in Haskell

Question

What are the standard methods to profile parser written in parser combinator libraries in Haskell?

I'm currently using uu-parsinglib but I would be very interested in knowing the methods of profiling also other parser combinator libraries like Parsec.

Right now I have written my parser and it is slow and eats a lot of ram (for 600 lines input text it takes about 5 seconds to parse with more than 1Gb of RAM and I want to investigate how I can improve it)

Answer 1

Try heap profiling:

$ ./prog +RTS -K128M -hc -p 
$ hp2ps -c prog.hp

If the profile looks like a mountain and is in the megabytes, probably you construct a large data structure first and then reduce it (and then could look into using accumulators or memoization).

More detailed info: http://book.realworldhaskell.org/read/profiling-and-optimization.html