X-Spam-Check-By: sourceware.org
X-AntiVirus: PTMail-AV 0.3.88
X-Virus-Status: Clean (0.16738 seconds)
Mime-Version: 1.0 (Apple Message framework v746.3)
In-Reply-To: <001401c651d0$872c9090$b3db87d4@multiplay.co.uk>
References: <048501c651b8$0862c1c0$a501a8c0@CAM.ARTIMI.COM> <001401c651d0$872c9090$b3db87d4@multiplay.co.uk>
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <89F9E813-1E27-4FEE-9359-6C73A8CD871A@ptnix.com>
Content-Transfer-Encoding: 7bit
From: Pedro Inacio <pedro.inacio@ptnix.com>
Subject: Re: very poor cygwin scp performance in some situations
Date: Mon, 27 Mar 2006 23:22:52 +0100
To: Cygwin List <cygwin@cygwin.com>
X-IsSubscribed: yes
Mailing-List: contact cygwin-help@cygwin.com; run by ezmlm
Precedence: bulk
List-Unsubscribe: <mailto:cygwin-unsubscribe-archive-cygwin=delorie.com@cygwin.com>
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

Hello,

I've initiated the post "select() too slow".
The example I've done was a non-blocking tcp echo_server.c and how  
slow it was on linux compared with cygwin.
In fact after activation oh tcp_nodelay option the times were similar.

Meanwhile I've done another test, I've added ssl on top of that  
echo_server.c with similar results than on linux, with the  
tcp_nodelay set, and the results are again against cygwin binary.
On cygwin is too slow. I just don't know where the problem is.



On 2006/03/27, at 19:59, Steven Hartland wrote:

> ----- Original Message ----- From: "Dave Korn"
>
>> On 27 March 2006 16:39, Steven Hartland wrote:
>>> I've done quite a bit of digging and not found any real answers
>>> on why cygwin's scp performance is so poor.
>>  See the thread about "select() too slow" last week for an  
>> explanation of the
>> Nagle algorithm, how windoze's implementation of it appears to be  
>> crippling,
>> and how using TCP_NODELAY can fix everything.  (Thread ran from 11  
>> Mar to 19
>> Mar).
>
> As a quick and dirty hack I channed set_nodelay into set_nonblock  
> which
> seems to be set on all the sockets but unfortunately it doesn't  
> seem to help.
> Anyone more familiar with the internals of openssh know where this  
> might
> done?
>
>    Steve
>

--
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/

