From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?ISO-8859-1?Q?Roger_Pau_Monn=E9?= Subject: Re: [Xen-devel] [PATCH 0/4] Implement persistent grant in xen-netfront/netback Date: Thu, 15 Nov 2012 11:56:17 +0100 Message-ID: <50A4CA51.8080208@citrix.com> References: <1352962987-541-1-git-send-email-annie.li@oracle.com> <20121115074057.GO8912@reaktio.net> <50A4AA06.8080900@oracle.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: =?ISO-8859-1?Q?Pasi_K=E4rkk=E4inen?= , "netdev@vger.kernel.org" , "xen-devel@lists.xensource.com" , Ian Campbell , "konrad.wilk@oracle.com" To: ANNIE LI Return-path: Received: from smtp.ctxuk.citrix.com ([62.200.22.115]:3005 "EHLO SMTP.EU.CITRIX.COM" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S964838Ab2KOK4U (ORCPT ); Thu, 15 Nov 2012 05:56:20 -0500 In-Reply-To: <50A4AA06.8080900@oracle.com> Sender: netdev-owner@vger.kernel.org List-ID: On 15/11/12 09:38, ANNIE LI wrote: >=20 >=20 > On 2012-11-15 15:40, Pasi K=E4rkk=E4inen wrote: >> Hello, >> >> On Thu, Nov 15, 2012 at 03:03:07PM +0800, Annie Li wrote: >>> This patch implements persistent grants for xen-netfront/netback. T= his >>> mechanism maintains page pools in netback/netfront, these page pool= s is used to >>> save grant pages which are mapped. This way improve performance whi= ch is wasted >>> when doing grant operations. >>> >>> Current netback/netfront does map/unmap grant operations frequently= when >>> transmitting/receiving packets, and grant operations costs much cpu= clock. In >>> this patch, netfront/netback maps grant pages when needed and then = saves them >>> into a page pool for future use. All these pages will be unmapped w= hen >>> removing/releasing the net device. >>> >> Do you have performance numbers available already? with/without pers= istent grants? > I have some simple netperf/netserver test result with/without persist= ent=20 > grants, >=20 > Following is result of with persistent grant patch, >=20 > Guests, Sum, Avg, Min, Max > 1, 15106.4, 15106.4, 15106.36, 15106.36 > 2, 13052.7, 6526.34, 6261.81, 6790.86 > 3, 12675.1, 6337.53, 6220.24, 6454.83 > 4, 13194, 6596.98, 6274.70, 6919.25 >=20 >=20 > Following are result of without persistent patch >=20 > Guests, Sum, Avg, Min, Max > 1, 10864.1, 10864.1, 10864.10, 10864.10 > 2, 10898.5, 5449.24, 4862.08, 6036.40 > 3, 10734.5, 5367.26, 5261.43, 5473.08 > 4, 10924, 5461.99, 5314.84, 5609.14 In the block case, performance improvement is seen when using a large number of guests, could you perform the same benchmark increasing the number of guests to 15?