From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thibault VINCENT Subject: Exceed 1GB/s with virtio-net ? Date: Mon, 13 Sep 2010 12:32:23 +0200 Message-ID: <4C8DFDB7.9000406@smartjog.com> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="------------enigB5E479E98361CEF6331A3F46" To: kvm@vger.kernel.org Return-path: Received: from mx.fr.smartjog.net ([95.81.144.3]:40607 "EHLO mx.fr.smartjog.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751018Ab0IMKh4 (ORCPT ); Mon, 13 Sep 2010 06:37:56 -0400 Received: from mail.dmz-ext.fr.lan (mail.dmz-ext.fr.lan [192.168.11.102]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx.fr.smartjog.net (Postfix [SmartJog]) with ESMTPS id 37F357283B for ; Mon, 13 Sep 2010 10:30:32 +0000 (UTC) Received: from localhost (localhost [127.0.0.1]) by mail.dmz-ext.fr.lan (Postfix [SmartJog]) with ESMTP id 19059572C1 for ; Mon, 13 Sep 2010 10:30:32 +0000 (GMT) Received: from mail.dmz-ext.fr.lan ([127.0.0.1]) by localhost (mail.dmz-ext.fr.lan [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 2mKQV69xElqw for ; Mon, 13 Sep 2010 10:30:30 +0000 (GMT) Received: from [192.168.0.53] (tvincent.user.fr.lan [192.168.0.53]) by mail.dmz-ext.fr.lan (Postfix [SmartJog]) with ESMTP id A0D6D572BC for ; Mon, 13 Sep 2010 10:30:30 +0000 (GMT) Sender: kvm-owner@vger.kernel.org List-ID: This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --------------enigB5E479E98361CEF6331A3F46 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Hello I'm trying to achieve higher than gigabit transferts over a virtio NIC with no success, and I can't find a recent bug or discussion about such an issue. The simpler test consist of two VM running on a high-end blade server with 4 cores and 4GB RAM each, and a virtio NIC dedicated to the inter-VM communication. On the host, the two vnet interfaces are enslaved into a bridge. I use a combination of 2.6.35 on the host and 2.6.32 in the VMs. Running iperf or netperf on these VMs, with TCP or UDP, result in ~900Mbits/s transferts. This is what could be expected of a 1G interface, and indeed the e1000 emulation performs similar. Changing the txqueuelen, MTU, and offloading settings on every interface (bridge/tap/virtio_net) didn't improve the speed, nor did the installation of irqbalance and the increase in CPU and RAM. Is this normal ? Is the multiple queue patch intended to address this ? It's quite possible I missed something :) Thank you dear list ! --=20 Thibault VINCENT SmartJog S.A.S. - Groupe TDF - P=F4le multim=E9dia 27 Bd Hippolyte Marques, 94200 Ivry sur Seine, France Phone : +33.1.58.68.62.38 Fax : +33.1.58.68.60.97 --------------enigB5E479E98361CEF6331A3F46 Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) iQIcBAEBAgAGBQJMjf28AAoJEEtbpw2b6p8jm5wP/2BDiaLpt4wPgqq79yPPOteY HtoTm7zQB8RvF+0Gejyn7PFM32eGWU+ZHZjHwm8NhCvsDfFUaQ8qNgp4SfznESf8 nBJQhUaCCPFC7m8sCnfgwVbo8WZEpE/3pD+FPbZxFIc8qDWShqWFhyYU3JM3w61d n3NQTErNsOOt+1cSb21DWCzLNmBlZpX1atNfFH5fU+ndI/hlh4a0xhzquQvAEG1Q Lg9U5PdhcSR70l6cGBLGjPYzwR9Ej33kKideIvz+EMhwGki9NLmeI3Ml8L0wC+Jy Z50Z0MVBg7XBMDxooBwyBMqGjtNDvpszEcV+xq5chH09+FvSqGMyU2C69xtZGwD7 NVoU2nvkWa65/XRmKVeGm0jYoxNy5D7ILwbm2JX8ZRb6d0oTj8q4ZAr9zVMjVRob v6PAYSkdZ0GkbrU3sASj5HL58w8yCUi3MyyidoN1FUY/Pt270T40sZtLHUoZ59/t aJW+X6vW9GHLqswBCcELxH8p0dNDG1cu4oyCk5+uTYnZIstFYc6iMh7szczP0NJs hxDgsQoIxOW4d5ciWZiFWW1caU98wSU0l/LB5sx6LigAdvIgVsv4Hhw+5iiTuEir 2MSzudY/4iYP+Io4gPABx3+1xZoANJqOIvgNq/kaiZ7xVXUJnMpuI1IAFDsKH7Ln DI8sGKDFlT6Q6MxGU7no =AKGI -----END PGP SIGNATURE----- --------------enigB5E479E98361CEF6331A3F46--