netdev
[Top] [All Lists]

Re: Intel and TOE in the news

To: hadi@xxxxxxxxxx
Subject: Re: Intel and TOE in the news
From: Robert Olsson <Robert.Olsson@xxxxxxxxxxx>
Date: Mon, 21 Feb 2005 16:38:01 +0100
Cc: Thomas Graf <tgraf@xxxxxxx>, Andi Kleen <ak@xxxxxx>, Leonid Grossman <leonid.grossman@xxxxxxxxxxxx>, "'rick jones'" <rick.jones2@xxxxxx>, netdev@xxxxxxxxxxx, "'Alex Aizman'" <alex@xxxxxxxxxxxx>
In-reply-to: <1108996313.1090.178.camel@xxxxxxxxxxxxxxxx>
References: <20050220230713.GA62354@xxxxxx> <200502210332.j1L3WkDD014744@xxxxxxxxxxxxxxxxx> <20050221115006.GB87576@xxxxxx> <20050221132844.GU31837@xxxxxxxxxxxxxx> <1108994621.1089.158.camel@xxxxxxxxxxxxxxxx> <20050221141714.GV31837@xxxxxxxxxxxxxx> <1108996313.1090.178.camel@xxxxxxxxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
jamal writes:

 > Infact performance goes down when you batch in some cases depending on
 > the hardware used. My investigation shows that if you have a fast CPU
 > and a fast bus, theres always only one packet in flight within the
 > stack. Batching by definition requires more than one packet.

 Hello!

 Yes when queue length/batch increases you're risking to load the L2 
 twice for the same skb. Which is the most expensive operation.... 
 Forwarding profiles show most functions where cache misses occur.

                                        --ro

<Prev in Thread] Current Thread [Next in Thread>