From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [tbench regression fixes]: digging out smelly deadmen. Date: Thu, 30 Oct 2008 19:43:32 +0100 Message-ID: <490A0054.2030903@cosmosbay.com> References: <20081009231759.GA8664@tservice.net.ru> <20081010115518.GA3159@tservice.net.ru> <20081010115725.GD19487@elte.hu> <200810250025.35734.rjw@sisk.pl> <20081026112924.GA29258@ioremap.net> <20081026122300.GA30905@ioremap.net> <20081030111526.7d9bb0f8@extreme> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Evgeniy Polyakov , "Rafael J. Wysocki" , Ingo Molnar , Evgeniy Polyakov , Peter Zijlstra , linux-kernel@vger.kernel.org, netdev@vger.kernel.org, David Miller , Mike Galbraith , Andrew Morton To: Stephen Hemminger Return-path: Received: from gw1.cosmosbay.com ([86.65.150.130]:46776 "EHLO gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753304AbYJ3Sog convert rfc822-to-8bit (ORCPT ); Thu, 30 Oct 2008 14:44:36 -0400 In-Reply-To: <20081030111526.7d9bb0f8@extreme> Sender: netdev-owner@vger.kernel.org List-ID: Stephen Hemminger a =E9crit : > Has anyone looked into the impact of port randomization on this bench= mark. > If it is generating lots of sockets quickly there could be an impact: > * port randomization causes available port space to get filled non-= uniformly > and what was once a linear scan may have to walk over existing po= rts. > (This could be improved by a hint bitmap) >=20 > * port randomization adds at least one modulus operation per socket > creation. This could be optimized by using a loop instead. tbench setups one socket per client, then send/receive lot of messages = on this socket. Connection setup time can be ignored for the tbench regression analysis