Complete.Org: Mailing Lists: Archives: freeciv-dev: November 2003:
[Freeciv-Dev] Speed observations
Home

[Freeciv-Dev] Speed observations

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: Freeciv Developers <freeciv-dev@xxxxxxxxxxx>
Subject: [Freeciv-Dev] Speed observations
From: Christian Knoke <chrisk@xxxxxxxxx>
Date: Fri, 7 Nov 2003 19:59:40 +0100

Rather than doing exact profiling, I want to share some speed impressions
with you, most of them evolved since 1.14 is out. I'm in a good position to
do so, since I still use an ol' 200 MHz AMD box here for Freeciv.

If you get, by reading between the lines, the impression that playing
Freeciv is kind of a pain nowadays, that impression is not totally wrong.

This is about the GTK 1.2 client and the following game:

chris@max:~/freeciv/games> civparse -v Honecker+1680m.sav.gz 

Filename               Version   Ruleset   Year    Map       Generator
----------------------------------------------------------------------
Honecker+1680m.sav.gz  1.14.9    default   +1680   80 x 50     1

Player         Nation       Control  Government  Cities  AvSize  Techs
----------------------------------------------------------------------
Erich Honecker German       chris    Democracy      110     8.0     78
Karadorde Petr Serbian      AI       Democracy        3     4.0     64
José Miguel Ca Chilean      AI       Democracy        4     4.0     71
Gia Long       Vietnamese   dead     Democracy        0     0.0     70
Boadicea       Barbarian    dead     Despotism        0     0.0      4
Stilicho       Barbarian    dead     Anarchy          0     0.0      1
Primoz Trubar  Slovenian    AI       Democracy        2     4.0     64

Player         Units  Sum(A*FP*V)
---------------------------------
Erich Honecker   294       4298.0
Karadorde Petr     1          6.0
José Miguel Ca     3         12.0
Gia Long           0          0.0
Boadicea           0          0.0
Stilicho           0          0.0
Primoz Trubar      0          0.0

All times are rough CPU seconds (User time) given by 'ps aux'. The real time
is bigger.


1. Load and Start of a game

Thats the only case where civserver matters. In all other cases I leave
civserver out.

civserver 0:07
civclient 0:10


2. Moving a Mech.Inf. across the continent on railroads

Move animation off or on, doesn't matter

civclient 0:02

This is not caused by city descriptions (turned on). Note that I can see
that redraw after the move, and it is quick; with animation turned off there
is no redraw.

Also note, no difference with or without CMA active.


3. Moving a full loaded transport from one city into another

civlient 0:02

CMA on or off. Most time passes when the ship leaves and enters the city.
Redraw doesn't matter in my observations.


4. Switching global tax rate by 10 percent

CMA off:

civclient 0:04

CMA on:

civclient 0:16

No comment.


5. Switching what to build for all cities (110)

civclient 0:04

Some differences are likely caused by redraw of column headers in city
reports. These flicker a lot in some cases.


6. Switching CMA preset for all cities (110)

In this case the difference between civclient user time and real time is big;
the real time is hand-stopped.

civclient 0:07
real      0.31

-- 
Christian Knoke            * * *            http://cknoke.de
* * * * * * * * *  Ceterum censeo Microsoft esse dividendum.


[Prev in Thread] Current Thread [Next in Thread]
  • [Freeciv-Dev] Speed observations, Christian Knoke <=