netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Stephen Hemminger <stephen.hemminger@vyatta.com>
To: Roland Dreier <rdreier@cisco.com>
Cc: Bill Fink <billfink@mindspring.com>,
	Evgeniy Polyakov <johnpol@2ka.mipt.ru>,
	David Miller <davem@davemloft.net>,
	aglo@citi.umich.edu, shemminger@vyatta.com,
	netdev@vger.kernel.org, rees@umich.edu, bfields@fieldses.org
Subject: Re: setsockopt()
Date: Tue, 8 Jul 2008 13:48:45 -0700	[thread overview]
Message-ID: <20080708134845.2372a483@speedy> (raw)
In-Reply-To: <adar6a4x510.fsf@cisco.com>

On Mon, 07 Jul 2008 23:29:31 -0700
Roland Dreier <rdreier@cisco.com> wrote:

> Interesting... I'd not tried nuttcp before, and on my testbed, which is
> a very high-bandwidth, low-RTT network (IP-over-InfiniBand with DDR IB,
> so the network is capable of 16 Gbps, and the RTT is ~25 microseconds),
> the difference between autotuning and not for nuttcp is huge (testing
> with 2.6.26-rc8 plus some pending 2.6.27 patches that add checksum
> offload, LSO and LRO to the IP-over-IB driver):
> 
> nuttcp -T30 -i1 ends up with:
> 
> 14465.0625 MB /  30.01 sec = 4043.6073 Mbps 82 %TX 2 %RX
> 
> while setting the window even to 128 KB with
> nuttcp -w128k -T30 -i1 ends up with:
> 
> 36416.8125 MB /  30.00 sec = 10182.8137 Mbps 90 %TX 96 %RX
> 
> so it's a factor of 2.5 with nuttcp.  I've never seen other apps behave
> like that -- for example NPtcp (netpipe) only gets slower when
> explicitly setting the window size.
> 
> Strange...

I suspect that the link is so fast that the window growth isn't happening
fast enough. With only a 30 second test, you probably barely made it
out of TCP slow start.

  parent reply	other threads:[~2008-07-08 20:48 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-07-07 18:18 setsockopt() Olga Kornievskaia
2008-07-07 21:24 ` setsockopt() Stephen Hemminger
2008-07-07 21:30   ` setsockopt() Olga Kornievskaia
2008-07-07 21:33     ` setsockopt() Stephen Hemminger
2008-07-07 21:49     ` setsockopt() David Miller
2008-07-08  4:54       ` setsockopt() Evgeniy Polyakov
2008-07-08  6:02         ` setsockopt() Bill Fink
2008-07-08  6:29           ` setsockopt() Roland Dreier
2008-07-08  6:43             ` setsockopt() Evgeniy Polyakov
2008-07-08  7:03               ` setsockopt() Roland Dreier
2008-07-08 18:48             ` setsockopt() Bill Fink
2008-07-09 18:10               ` setsockopt() Roland Dreier
2008-07-09 18:34                 ` setsockopt() Evgeniy Polyakov
2008-07-10  2:50                   ` setsockopt() Bill Fink
2008-07-10 17:26                     ` setsockopt() Rick Jones
2008-07-11  0:50                       ` setsockopt() Bill Fink
2008-07-08 20:48             ` Stephen Hemminger [this message]
2008-07-08 22:05               ` setsockopt() Bill Fink
2008-07-09  5:25                 ` setsockopt() Evgeniy Polyakov
2008-07-09  5:47                   ` setsockopt() Bill Fink
2008-07-09  6:03                     ` setsockopt() Evgeniy Polyakov
2008-07-09 18:11                       ` setsockopt() J. Bruce Fields
2008-07-09 18:43                         ` setsockopt() Evgeniy Polyakov
2008-07-09 22:28                           ` setsockopt() J. Bruce Fields
2008-07-10  1:06                             ` setsockopt() Evgeniy Polyakov
2008-07-10 20:05                               ` [PATCH] Documentation: clarify tcp_{r,w}mem sysctl docs J. Bruce Fields
2008-07-10 23:50                                 ` David Miller
2008-07-08 20:12       ` setsockopt() Jim Rees
2008-07-08 21:54         ` setsockopt() John Heffner
2008-07-08 23:51           ` setsockopt() Jim Rees
2008-07-09  0:07             ` setsockopt() John Heffner
2008-07-07 22:50     ` setsockopt() Rick Jones
2008-07-07 23:00       ` setsockopt() David Miller
2008-07-07 23:27         ` setsockopt() Rick Jones
2008-07-08  1:15           ` setsockopt() Rick Jones
2008-07-08  1:48             ` setsockopt() J. Bruce Fields
2008-07-08  1:44           ` setsockopt() David Miller
2008-07-08  3:33       ` setsockopt() John Heffner
2008-07-08 18:16         ` setsockopt() Rick Jones
2008-07-08 19:10           ` setsockopt() John Heffner
     [not found]         ` <349f35ee0807090255s58fd040bne265ee117d06d397@mail.gmail.com>
2008-07-09 10:38           ` setsockopt() Jerry Chu
2008-07-07 21:32   ` setsockopt() J. Bruce Fields
2008-07-08  1:17 ` setsockopt() John Heffner

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20080708134845.2372a483@speedy \
    --to=stephen.hemminger@vyatta.com \
    --cc=aglo@citi.umich.edu \
    --cc=bfields@fieldses.org \
    --cc=billfink@mindspring.com \
    --cc=davem@davemloft.net \
    --cc=johnpol@2ka.mipt.ru \
    --cc=netdev@vger.kernel.org \
    --cc=rdreier@cisco.com \
    --cc=rees@umich.edu \
    --cc=shemminger@vyatta.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).