From: Baruch Even <baruch@ev-en.org>
To: "David S. Miller" <davem@davemloft.net>
Cc: John Heffner <jheffner@psc.edu>,
ak@muc.de, shemminger@osdl.org, netdev@oss.sgi.com
Subject: Re: [RFC] TCP congestion schedulers
Date: Mon, 21 Mar 2005 22:30:03 +0000 [thread overview]
Message-ID: <423F4AEB.9040100@ev-en.org> (raw)
In-Reply-To: <20050321135154.0bbeae85.davem@davemloft.net>
David S. Miller wrote:
> On Mon, 21 Mar 2005 16:25:56 -0500 (EST)
> John Heffner <jheffner@psc.edu> wrote:
>
>
>>Would you really expect a single extra indirect call per ack to have a
>>significant performance impact? This is surprising to me. Where does the
>>cost come from? Replacing instruction cache lines?
>
> Maybe not for ACK processing (that's very thick already) but
> perhaps for a lighter fast path definitely so.
According to my tests (wrapping tcp_ack with rdtsc's) it takes about
3000 clocks to do tcp_ack() even for fast path, slow path is not much
slower in most cases and anyway most of the time is spent either
handling SACKs or remove packets from the transmit queue (clean_rtx).
I doubt if the extra calls by function are going to be that much of an
issue.
Now, if I knew how to improve performance of the clean_rtx case that
would give a boost to ack performance.
Baruch
next prev parent reply other threads:[~2005-03-21 22:30 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-02-23 21:30 [PATCH] select congestion control with one sysctl Baruch Even
2005-02-23 21:57 ` David S. Miller
2005-02-24 0:23 ` Stephen Hemminger
2005-02-24 0:33 ` David S. Miller
2005-02-26 9:41 ` Arnaldo Carvalho de Melo
[not found] ` <421D30FA.1060900@ev-en.org>
[not found] ` <20050225120814.5fa77b13@dxpl.pdx.osdl.net>
[not found] ` <20050309210442.3e9786a6.davem@davemloft.net>
[not found] ` <4230288F.1030202@ev-en.org>
[not found] ` <20050310182629.1eab09ec.davem@davemloft.net>
[not found] ` <20050311120054.4bbf675a@dxpl.pdx.osdl.net>
[not found] ` <20050311201011.360c00da.davem@davemloft.net>
2005-03-14 23:17 ` [RFC] TCP congestion schedulers Stephen Hemminger
2005-03-15 19:54 ` John Heffner
2005-03-15 22:16 ` John Heffner
2005-03-18 4:12 ` David S. Miller
2005-03-18 12:53 ` Arnaldo Carvalho de Melo
2005-03-18 13:43 ` jamal
2005-03-18 16:13 ` Arnaldo Carvalho de Melo
2005-03-18 16:45 ` Stephen Hemminger
2005-03-18 16:59 ` Arnaldo Carvalho de Melo
2005-03-19 20:19 ` Andi Kleen
2005-03-21 21:25 ` John Heffner
2005-03-21 21:51 ` David S. Miller
2005-03-21 22:30 ` Baruch Even [this message]
2005-03-22 0:10 ` Rick Jones
2005-03-22 1:41 ` Olaf Kirch
2005-03-22 7:41 ` Andi Kleen
2005-03-28 23:51 ` Stephen Hemminger
2005-03-29 15:25 ` Andi Kleen
2005-03-29 17:17 ` Stephen Hemminger
2005-03-29 18:58 ` Rick Jones
2005-03-30 9:41 ` Matt Mackall
2005-03-29 19:32 ` John Heffner
2005-03-29 20:03 ` David S. Miller
2005-03-29 20:09 ` Rick Jones
2005-04-08 19:33 ` John Heffner
2005-04-08 20:20 ` Rick Jones
2005-02-24 1:05 ` [PATCH] select congestion control with one sysctl Daniele Lacamera
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=423F4AEB.9040100@ev-en.org \
--to=baruch@ev-en.org \
--cc=ak@muc.de \
--cc=davem@davemloft.net \
--cc=jheffner@psc.edu \
--cc=netdev@oss.sgi.com \
--cc=shemminger@osdl.org \
/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).