profiling,-O in 6.12.1

[Serge D. Mechveliani]

To my 

> >>>It looks like  ghc-6.12.1  reports erroneous time profiling --
> >>>when the Main module of the project is made under  -O.
> >>>[..]
> >>>This is for  ghc-6.12.1  made from source for Debian Linux and
> >>>i386-like.
> >>> []
> >The key combination
> >                        ghc $dmCpOpt -prof --make Main
> >shows  95% for  eLoop,
> >and adding  -O  to this line shows  0%  for  eLoop,  independently on
> >presence of  -auto-all  in this line
> >(the whole library is made under  -O -prof).
> 
> 
> >Yes, I recall that the effect may be of inlining.


Simon Marlow wrote  on Thu, Feb 04, 2010

> Inlining should not in general affect the shape of the profile.  I can't 
> say much else here without seeing the specific code.

Finally, by setting  -O0 -prof  I have found  "who of top-levels takes 
most of time" -- hopefully, the same functions will occur for -O.
I do not know, maybe, this is a known bug (Simon writes, there exists 
some report).
Anyway, it is important to debug profiling -- to making unnecessary any 
irregular tricks with  -inline-not.

Dear GHC team,
if I put my project to a public place, would you find out what is the 
effect there about  -O -prof ?

Although the whole program is large, but probably, the very example with 
`main' + Complete.complete + a single SCC,  is simple for you to detect
the effect source.

Regards, 

-----------------
Serge Mechveliani
mechvel at botik.ru