[Haskell-cafe] Re: Haskell performance

[Simon Marlow]

Simon Marlow wrote:

>> Nobench does already collect code size, but does not yet display it in
>> the results table.  I specifically want to collect compile time as well.
>> Not sure what the best way to measure allocation and peak memory use
>> are?
> 
> With GHC you need to use "+RTS -s" and then slurp in the <prog>.stat 
> file.  You can also get allocations, peak memory use, and separate 
> mutator/GC times this way.

Oh, and one more thing.  We have this program called nofib-analyse in GHC's 
source tree:

  http://darcs.haskell.org/ghc/utils/nofib-analyse

which takes the output from a couple of nofib runs and generates nice 
tables, in ASCII or LaTeX (for including in papers, see e.g. our 
pointer-tagging paper from ICFP'07).  The only reason we haven't switched 
to using nobench for GHC is the existence of this tool.  Unfortuantely it 
relies on specifics of the output generated by a nofib run, and uses a Perl 
script, etc. etc.  The point is, it needs some non-trivial porting.

I'm pointing this out just in case you or anyone else felt enthusiastic 
enough to port this to nobench, and to hopefully head off any duplication 
of effort.  Failing that, I'll probably get around to porting it myself at 
some point.

Cheers,
	Simon