From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Mon, 15 Nov 2010 18:49:10 +0100 From: Martin Schwidefsky Subject: Re: [RFC][PATCH v2 5/7] taskstats: Improve cumulative CPU time accounting Message-ID: <20101115184910.549a828b@mschwide.boeblingen.de.ibm.com> In-Reply-To: <1289837036.2109.501.camel@laptop> References: <20101111170352.732381138@linux.vnet.ibm.com> <20101111170815.404670062@linux.vnet.ibm.com> <20101113183810.GA9021@redhat.com> <20101115165521.21baac60@mschwide.boeblingen.de.ibm.com> <1289837036.2109.501.camel@laptop> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-Archive: List-Post: To: Peter Zijlstra Cc: Oleg Nesterov , Michael Holzheu , Shailabh Nagar , Andrew Morton , Venkatesh Pallipadi , Suresh Siddha , Ingo Molnar , John stultz , Thomas Gleixner , Balbir Singh , Heiko Carstens , Roland McGrath , linux-kernel@vger.kernel.org, linux-s390@vger.kernel.org List-ID: On Mon, 15 Nov 2010 17:03:56 +0100 Peter Zijlstra wrote: > On Mon, 2010-11-15 at 16:55 +0100, Martin Schwidefsky wrote: > > What we want is a low-overhead tool that precisely shows > > where the cpu spent its time (or didn't because of steal time). The > > granularity target is tenths of microseconds, something that should be > > possible with decent hardware. > > To what purpose? Is that a trick question? Why do we have tools like "top"? Or process accounting? The point is that the quality of the numbers we get right now is rather bad, the overhead of scanning /proc is horrendous and the 10ms granularity is rather coarse. -- blue skies, Martin. "Reality continues to ruin my life." - Calvin.