From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 20979 invoked by alias); 13 Dec 2003 19:55:24 -0000 Mailing-List: contact xconq7-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: xconq7-owner@sources.redhat.com Received: (qmail 20972 invoked from network); 13 Dec 2003 19:55:23 -0000 Received: from unknown (HELO smtp14.fre.skanova.net) (195.67.227.31) by sources.redhat.com with SMTP; 13 Dec 2003 19:55:23 -0000 Received: from [212.181.162.155] (h155n1fls24o1048.bredband.comhem.se [212.181.162.155]) by smtp14.fre.skanova.net (8.12.10/8.12.10) with ESMTP id hBDJtBMw002685; Sat, 13 Dec 2003 20:55:17 +0100 (CET) X-Sender: u22611592@m1.226.comhem.se Message-Id: In-Reply-To: <200312131621.hBDGLR809132@panix5.panix.com> References: (message from Hans Ronne on Sat, 13 Dec 2003 12:07:36 +0100) <20031207181259.GG1378@adlp.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Sun, 14 Dec 2003 04:33:00 -0000 To: Jim Kingdon From: Hans Ronne Subject: Re: fighters fighting without ammo Cc: xconq7@sources.redhat.com X-SW-Source: 2003/txt/msg01029.txt.bz2 >> I think there would be a negative effect on performance, but I haven't >> tested this rigourously. > >I wonder if it would be a good idea to put in some instrumentation for >performance (like reporting times for "AI action-reaction code", "UI", >and other such categories). Of course the tricky part is that the hot >spots might vary by game, by whether one is early or late in the game, >and any number of other variables. But subjectively it seems like I'm >waiting more and more for xconq while the AI's move (I use sequential >mode, because the game is easier if the AI's move first and use up >their ACP's before I move). I usually do profiling when I want to look at performance. I improved the xconq profiling support on the Mac a long time ago, but perhaps something should be done to facilitate profiling on other platforms as well? I haven't looked into this since I mostly use the Mac. As for the recent slowdowns in the AI code, they are caused by the new path code. Since I disabled the path cache in order to stop the network sync errors, it has to recalculate the paths every time, which takes a lot of time. Hopefully, when Peter has fixed the code, it will be faster again. Hans