Mailing-List: contact cygwin-help@cygwin.com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe@cygwin.com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin@cygwin.com>
List-Help: <mailto:cygwin-help@cygwin.com>, <http://sourceware.org/ml/#faqs>
Sender: cygwin-owner@cygwin.com
Mail-Followup-To: cygwin@cygwin.com
Delivered-To: mailing list cygwin@cygwin.com
Message-ID: <BAY102-DAV12D4850AC850F49280B6A7A9040@phx.gbl>
X-Originating-Email: [unknown_kev_cat@hotmail.com]
X-Sender: unknown_kev_cat@hotmail.com
From: "Tacvek" <unknown_kev_cat@hotmail.com>
To: "Gerrit P. Haase" <gerrit@familiehaase.de>
Cc: <cygwin@cygwin.com>
References: <000b01c563c7$f4456f00$976d65da@DANNY> <4298F195.9090600@familiehaase.de> <20050529000152.GE3094@trixie.casa.cgf.cx> <20050529001846.GG3094@trixie.casa.cgf.cx> <20050529030523.GA10263@trixie.casa.cgf.cx> <20050529171324.GA26603@trixie.casa.cgf.cx> <429A19EC.9080905@familiehaase.de> <20050529201431.GB28340@trixie.casa.cgf.cx> <d7fo8b$uda$1@sea.gmane.org> <429B78A5.2050107@familiehaase.de>
Subject: Re: Serious performance problems (new snapshot has some performance  improvement)
Date: Mon, 30 May 2005 20:00:35 -0400
MIME-Version: 1.0
Content-Type: text/plain; 	format=flowed; 	charset="iso-8859-1"; 	reply-type=response
Content-Transfer-Encoding: 7bit


> Interesting, why is it faster when running a binary that doesn't
> depend on cygwin1.dll after swapping the DLL?  Some Win caching mechanism?
>
Repeating many times should minimize any caching:

#New DLL
$ /bin/time -f "%E %S %U" cygspd-mingw-ne-o7 cygspd.dat
0:04.48 0.01 0.02
0:01.40 0.00 0.02
0:01.46 0.01 0.01
0:01.39 0.01 0.01
0:01.43 0.01 0.01
#old DLL
$ /bin/time -f "%E %S %U" cygspd-mingw-ne-o7 cygspd.dat
0:01.45 0.01 0.01
0:01.41 0.00 0.01
0:01.40 0.01 0.01
0:01.41 0.00 0.02
0:01.41 0.00 0.02
0:01.43 0.01 0.01


Ok there is obviously some caching, which means that the original values I 
gave inidacte that the improvement is somewhat less than they indicate. They 
did not indicate very spectacular performance anyway. But fractional seconds 
do not mean much.
Based on some additioal experimentation, you should assume that the values 
are acurate to +/- 1 second. 

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

