X-Spam-Check-By:	sourceware.org
Message-ID:	<04b601c651b4$a1e9e020$b3db87d4@multiplay.co.uk>
From:	"Steven Hartland" <killing AT multiplay DOT co DOT uk>
To:	"Cygwin List" <cygwin AT cygwin DOT com>
Subject:	very poor cygwin scp performance in some situations
Date:	Mon, 27 Mar 2006 16:39:20 +0100
MIME-Version:	1.0
X-Spam-Processed:	multiplay.co.uk, Mon, 27 Mar 2006 16:39:25 +0100 (not processed: message from valid local sender)
X-MDRemoteIP:	212.135.219.179
X-Return-Path:	killing AT multiplay DOT co DOT uk
X-MDaemon-Deliver-To:	cygwin AT cygwin DOT com
X-MDAV-Processed:	multiplay.co.uk, Mon, 27 Mar 2006 16:39:42 +0100
X-IsSubscribed:	yes
Mailing-List:	contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe:	<mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive:	<http://sourceware.org/ml/cygwin/>
List-Post:	<mailto:cygwin AT cygwin DOT com>
List-Help:	<mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/ml/#faqs>
Sender:	cygwin-owner AT cygwin DOT com
Mail-Followup-To:	cygwin AT cygwin DOT com
Delivered-To:	mailing list cygwin AT cygwin DOT com
Note-from-DJ:	This may be spam

I've done quite a bit of digging and not found any real answers
on why cygwin's scp performance is so poor.

So I decided to do some tests, here the results:
== machines ==
* P4 2.8Ghz
* 1GB RAM
* IDE disk
* Running nothing but these test

=== test1 ===
* FreeBSD 5.4-RELEASE
* OpenSSH_3.8.1p1 FreeBSD-20040419, OpenSSL 0.9.7e 25 Oct 2004

=== test2 ===
* cygwin 1.5.4
* OpenSSH_3.7.1p2, SSH protocols 1.5/2.0, OpenSSL 0.9.7c 30 Sep 2003

=== test3 ===
* cygwin 1.5.19
* OpenSSH_4.3p2, OpenSSL 0.9.8a 11 Oct 2005

=== test4 ===
* FreeBSD 6.0-RELEASE
* OpenSSH_4.2p1 FreeBSD-20050903, OpenSSL 0.9.7e-p1 25 Oct 2004

== tests ==
=== test1 -> test2 ( from test1 ) ===
* scp cygwin.zip dedicated25:/tmp/ 
* cygwin.zip 192MB   3.2MB/s   01:00

=== test1 -> test3 ( from test1 ) ===
* scp cygwin.zip test3:/tmp/ 
* cygwin.zip 192MB   6.4MB/s   00:30

=== test1 -> test4 ( from test1 ) ===
* scp cygwin.zip dev4:/tmp/
* cygwin.zip 192MB  10.7MB/s   00:18 

=== test3 -> test1 ( from test1 ) ===
* scp test3:/tmp/cygwin.zip /tmp
* cygwin.zip 192MB 723.4KB/s   04:32 

=== test2 -> test1 ( from test1 ) ===
* scp test2:/tmp/cygwin.zip /tmp
* cygwin.zip 192MB 705.3KB/s   04:39 

=== test4 -> test1 ( from test1 ) ===
* scp cygwin.zip dev4:/tmp/
* cygwin.zip 192MB  11.3MB/s   00:17 

=== test1 -> test2 ( from test2 ) ===
* scp test1:/tmp/cygwin.zip /tmp
* cygwin.zip 192MB   4.6MB/s   00:38

=== test2 -> test3 ( from test2 ) ===
* scp /tmp/cygwin.zip test3:/tmp/
* cygwin.zip 192MB   3.1MB/s   00:58

=== test1 -> test3 ( from test3 ) ===
* scp test1:/tmp/cygwin.zip /tmp
* cygwin.zip 192MB   6.4MB/s   00:30

=== test3 -> test2 ( from test3 ) ===
* scp /tmp/cygwin.zip test2:/tmp/
* cygwin.zip 192MB   3.1MB/s   01:02

=== test3 -> test1 ( from test3 ) ===
* scp /tmp/cygwin.zip test1:/tmp/
* cygwin.zip 192MB   1.5MB/s   02:10

As you can see copying from a FreeBSD box -> cygwin box the
performance is not to bad at ~6MB/s. Its not as good as it could
be 10 -> 11MB/s as shown by the FreeBSD -> FreeBSD test
but it's respectable.

The problem is with the cygwin -> FreeBSD test. When instigated
from the FreeBSD box where the rate drops to ~ 700KB/s which
is terrible. Instigating the copy from the cygwin box doubles this
but IMO this still produces an unacceptably low rate being almost
8 times slower than the real life achievable result.

So what could be the reason for this terrible performance?

Notes:
* All transfers where done with aes128-cbc hmac-md5
* CPU usage was barely measurable on the tests producing low rates

    Steve


================================================
This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it. 

In the event of misdirection, illegible or incomplete transmission please telephone (023) 8024 3137
or return the E.mail to postmaster AT multiplay DOT co DOT uk.


--
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/

- Raw text -