From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pf0-x243.google.com (mail-pf0-x243.google.com [IPv6:2607:f8b0:400e:c00::243]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id 3yBBwl28hDzDqlv for ; Tue, 10 Oct 2017 20:42:30 +1100 (AEDT) Received: by mail-pf0-x243.google.com with SMTP id b85so13967591pfj.1 for ; Tue, 10 Oct 2017 02:42:30 -0700 (PDT) Date: Tue, 10 Oct 2017 19:42:14 +1000 From: Nicholas Piggin To: linuxppc-dev@lists.ozlabs.org Cc: kvm-ppc@vger.kernel.org, linux-pm@vger.kernel.org Subject: Re: [PATCH] powerpc/pseries/cpuidle: add polling idle for shared processor guests Message-ID: <20171010194140.5aec88d9@roar.ozlabs.ibm.com> In-Reply-To: <20171010071109.31207-1-npiggin@gmail.com> References: <20171010071109.31207-1-npiggin@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Tue, 10 Oct 2017 17:11:09 +1000 Nicholas Piggin wrote: > For shared processor guests (e.g., KVM), add an idle polling mode rather > than immediately returning to the hypervisor when the guest CPU goes > idle. > > Test setup is a 2 socket POWER9 with 4 guests running, each with vCPUs > equal to 1/2 of real of CPUs. Saturated each guest with tbench. Using > polling idle gives about 1.4x throughput. Actually it's even more noticeable when the host is not over subscribed by the looks. 2.5x-3x increase in throughput for just a single guest. Thanks, Nick