Complete.Org: Mailing Lists: Archives: freeciv-dev: September 2001:
[Freeciv-Dev] Re: Unused field in move_cost_map removed.
Home

[Freeciv-Dev] Re: Unused field in move_cost_map removed.

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: Gregory Berkolaiko <gberkolaiko@xxxxxxxxxxx>
Cc: freeciv development list <freeciv-dev@xxxxxxxxxxx>
Subject: [Freeciv-Dev] Re: Unused field in move_cost_map removed.
From: Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>
Date: Wed, 19 Sep 2001 14:24:08 +0200
Reply-to: rf13@xxxxxxxxxxxxxxxxxxxxxx

On Wed, Sep 19, 2001 at 12:13:37PM +0100, Gregory Berkolaiko wrote:
>  --- Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx> wrote: 
> > On Tue, Sep 18, 2001 at 01:51:21PM +0100, Gregory Berkolaiko wrote:
> [..]
> > 
> > Thue confirmed that this was the original intention. Still the number
> > I get from you aren't clear. How many times is this shortcut taken? 
> 
> I feel quite reluctant calculating the number, because the evidence I got
> from run-times was quite convincing.  Instead, I will find the places in
> the code where danger of returning the same tile twice still exists and
> minimize such danger.
> 
> > From how many times in total? Also can you try the attached shell
> > script? This gives at least for me constant user times.
> 
> I tried the script (minor edited), the results of the runs are attached.
> In one of the corrected runs the time is quite high but I think it is due
> to the system doing something at the time.
> 
> The difference in time is smaller than in my previous results because you
> set endyear to 500AD and I did it to 1800AD.  Also in my previous runs I
> used three different seeds:
> (rand)seed = 42, 43, 44
> 
> What exactly gives you constant user times?

I experimented a bit:

no extra load (just the normal seti@home, mp3 player):
96.73user 0.54system 2:05.21elapsed 77%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (807major+3715minor)pagefaults 0swaps
96.49user 1.32system 2:12.10elapsed 74%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (733major+3715minor)pagefaults 0swaps

extra load:
111.55user 1.63system 4:41.64elapsed 40%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (733major+3715minor)pagefaults 0swaps

So the user time increase but not as much as the real time. So I would
suspect CPU cache misses from rescheduling and a not very fine
measurement by the kernel overall.

To answer your question: don't change the workload.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 "It is not yet possible to change operating system by writing
  to /proc/sys/kernel/ostype."              sysctl(2) man page


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