Complete.Org: Mailing Lists: Archives: freeciv-dev: March 2002:
[Freeciv-Dev] Re: [RFC][Patch] Inline
Home

[Freeciv-Dev] Re: [RFC][Patch] Inline

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: Raahul Kumar <raahul_da_man@xxxxxxxxx>
Cc: freeciv development list <freeciv-dev@xxxxxxxxxxx>
Subject: [Freeciv-Dev] Re: [RFC][Patch] Inline
From: Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>
Date: Thu, 21 Mar 2002 13:48:57 +0100
Reply-to: rf13@xxxxxxxxxxxxxxxxxxxxxx

On Thu, Mar 21, 2002 at 04:14:02AM -0800, Raahul Kumar wrote:
> 
> <snip>
> 
> 
> > > 2.) gcc version
> > 
> > 3.0.3
> 
> I've only got 2.96. Should I d/l 3.03 to make it strictly comparable, or is
> using my version ok?

This is ok.

> > > 4.) CVS version
> > 
> > The one from yesterday.
> > 
> 
> March 20 it is.
> 
> 
> > 6.) Compile flags. The easiest solution would IMHO be if the patch
> > creates a file freeciv/make_options which the script passed to
> > make. Since these can be a lot:
> >  - -O2 vs -O3
> >  - -DUSE_INLINE or not
> >  - -fomit-frame-pointer or not
> 
> You used that? You didn't mention this before. 

I experimented with it. It showed a small increase.

> >  - -DNDEBUG or not
> > 
> > 7.) Enable profiling. Passed -pg to gcc and runs gprof later. Archive
> > the result.
> > 
> > > > Results:
> > > >  stock: 62.15s
> > > >  with this patch and no inlining: 80.97s (difference because of
> > > >  converting map_inx, dirstep,.. to functions)
> > > >  with this patch and inlining: 40.46s
> > > >  with this patch and inlining and NDEBUG: 36.83s
> > > >  with this patch and inlining and -O3: 39.48s
> > > > 
> > > 
> > > I seem to spot a missing
> > > 
> > > with this patch and inlining and -O3 and NDEBUG:
> > > 
> > > It would also be nice to have results for stock with -03 to compare with
> > > patch and -03.
> > 
> > I don't have these numbers.
> 
> I was requesting those numbers.

I do use you my computer today ;) Such usage _does_ (it shouldn't but
it does) change the user time. This is also a reason why a nightly
testing would be nice.

> It seems -03 is a fairly bad optimisation.

There is just nothing more to squeeze.

> Jason ran some tests on the stock freeciv, and he seemed to get large 
> increases
> in performance just by using NDEBUG. Jason, some comments on your previous
> tests please?
>
> With your patch and NDEBUG the speed increase is minor. Why?

Note that I disabled CHECK_MAP_POS in all tests. NDEBUG will also
disable CHECK_MAP_POS and so lead to an performance increase.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
  One nuclear bomb can ruin your whole day.


[Prev in Thread] Current Thread [Next in Thread]