From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750723AbWDISca (ORCPT ); Sun, 9 Apr 2006 14:32:30 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750734AbWDISca (ORCPT ); Sun, 9 Apr 2006 14:32:30 -0400 Received: from mail.gmx.net ([213.165.64.20]:21463 "HELO mail.gmx.net") by vger.kernel.org with SMTP id S1750723AbWDISca (ORCPT ); Sun, 9 Apr 2006 14:32:30 -0400 X-Authenticated: #14349625 Subject: Re: [patch][rfc] quell interactive feeding frenzy From: Mike Galbraith To: Al Boldi Cc: linux-kernel@vger.kernel.org In-Reply-To: <200604091944.28954.a1426z@gawab.com> References: <200604091944.28954.a1426z@gawab.com> Content-Type: text/plain Date: Sun, 09 Apr 2006 20:33:16 +0200 Message-Id: <1144607596.7408.34.camel@homer> Mime-Version: 1.0 X-Mailer: Evolution 2.4.0 Content-Transfer-Encoding: 7bit X-Y-GMX-Trusted: 0 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sun, 2006-04-09 at 19:44 +0300, Al Boldi wrote: > bert hubert wrote: > > On Sun, Apr 09, 2006 at 01:39:38PM +0200, Mike Galbraith wrote: > > > Ok, unusable may be overstated. Nonetheless, that bit of code causes > > > serious problems. It makes my little PIII/500 test box trying to fill > > > one 100Mbit local network unusable. That is not overstated. > > > > If you try to make a PIII/500 fill 100mbit of TCP/IP using lots of > > different processes, that IS a corner load. > > > > I'm sure you can fix this (rare) workload but are you very sure you are > > not killing off performance for other situations? > > This really has nothing to do w/ workload but rather w/ multi-user processing > /tasking /threading. And the mere fact that the 2.6 kernel prefers > kernel-threads should imply an overall performance increase (think pdflush). > > The reason why not many have noticed this scheduler problem(s) is because > most hackers nowadays work w/ the latest fastest hw available which does not > allow them to see these problems (think Windows, where most problems are > resolved by buying the latest hw). > > Real Hackers never miss out on making their work run on the smallest common > denominator (think i386dx :). Please don't trim the cc list. I almost didn't see this, and I really do want to hear each and every opinion. -Mike