From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by dpdk.org (Postfix) with ESMTP id B385636E for ; Tue, 29 Sep 2015 04:41:24 +0200 (CEST) Received: from fmsmga002.fm.intel.com ([10.253.24.26]) by fmsmga103.fm.intel.com with ESMTP; 28 Sep 2015 19:41:13 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.17,605,1437462000"; d="scan'208";a="814955159" Received: from fmsmsx108.amr.corp.intel.com ([10.18.124.206]) by fmsmga002.fm.intel.com with ESMTP; 28 Sep 2015 19:41:12 -0700 Received: from fmsmsx117.amr.corp.intel.com (10.18.116.17) by FMSMSX108.amr.corp.intel.com (10.18.124.206) with Microsoft SMTP Server (TLS) id 14.3.248.2; Mon, 28 Sep 2015 19:41:12 -0700 Received: from shsmsx101.ccr.corp.intel.com (10.239.4.153) by fmsmsx117.amr.corp.intel.com (10.18.116.17) with Microsoft SMTP Server (TLS) id 14.3.248.2; Mon, 28 Sep 2015 19:41:11 -0700 Received: from shsmsx102.ccr.corp.intel.com ([169.254.2.179]) by SHSMSX101.ccr.corp.intel.com ([169.254.1.75]) with mapi id 14.03.0248.002; Tue, 29 Sep 2015 10:41:10 +0800 From: "Xu, Qian Q" To: Thomas Monjalon Thread-Topic: [dpdk-dev] [PATCH v3] doc: add performance test guide to the linux gsg Thread-Index: AQHQ95wvuv6XHyrkIUW/6h2eAPNqOp5SzTyQ Date: Tue, 29 Sep 2015 02:41:09 +0000 Message-ID: <82F45D86ADE5454A95A89742C8D1410E01DE2309@shsmsx102.ccr.corp.intel.com> References: <1443095316-31699-1-git-send-email-qian.q.xu@intel.com> <1584750.6FoDJtCokB@xps13> In-Reply-To: <1584750.6FoDJtCokB@xps13> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.239.127.40] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Cc: "dev@dpdk.org" Subject: Re: [dpdk-dev] [PATCH v3] doc: add performance test guide to the linux gsg X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 29 Sep 2015 02:41:25 -0000 Thomas Thanks for the comments. See my comments inline below. Let me know if you h= ave objections, if no, I will send a v4 patch soon.=20 Thanks Qian -----Original Message----- From: Thomas Monjalon [mailto:thomas.monjalon@6wind.com]=20 Sent: Friday, September 25, 2015 10:11 PM To: Xu, Qian Q Cc: dev@dpdk.org Subject: Re: [dpdk-dev] [PATCH v3] doc: add performance test guide to the l= inux gsg 2015-09-24 19:48, Qian Xu: > --- a/doc/guides/linux_gsg/index.rst > +++ b/doc/guides/linux_gsg/index.rst > @@ -47,3 +47,4 @@ Contents > build_sample_apps > enable_func > quick_start > + perf_test_intel_platform_nic Why "test"? What about nic_perf_intel_platform? [Qian] OK.=20 > --- /dev/null > +++ b/doc/guides/linux_gsg/perf_test_intel_platform_nic.rst [...] > +Grub Boot Parameters > +~~~~~~~~~~~~~~~~~~~~ > + > +The following are some recommendations on GRUB boot settings: > + > +#. Use the default grub file as a starting point. > + > +#. Reserve 1G huge pages via grub configurations. For example to reserve= 8 huge pages of 1G size:: > + > + default_hugepagesz=3D1G hugepagesz=3D1G hugepages=3D8 > + > +#. Isolate CPU cores which will be used for DPDK. For example:: > + > + isolcpus=3D2,3,4,5,6,7,8 > + > +#. If it wants to use VFIO, use the following additional grub parameters= :: > + > + iommu=3Dpt intel_iommu=3Don These paramters can be set with GRUB, yes. But they are Linux command line parameters. I suggest to change the title. [Qian] OK.=20 [...] > +4. Bind the test ports to igb_uio. For example bind two ports to a DPDK = compatible driver and check the status: Not speaking about VFIO here suggests that igb_uio gives better performance= . By the way, UIO/VFIO is not used by mlx4. [Qian] In fact not suggests a better performance with igb_uio, but as to VF= IO, we need to turn on VT-d in the BIOS, it's an additional step. So here I= want to change it as "UIO(e.g: igb_uio) or VFIO(turn on VT-d in the BIOS)"= , not sure mlx4 uses which module for dpdk running.=20 [...] > + Note: Don't use one 2 lcores on the same core one 2 lcores? looks to be a typo [Qian] Will correct it. [...] > +5. As it is known that a XL710 40G port need at least two queue pairs to= achieve best performance, Is it well known? At least it is said here. [Qian] Maybe well known only in Intel. I will change it as : A XL710 40G po= rt need at least two queue pairs to achieve best performance.=20 > +8. Configure the traffic to a traffic generator such as IXIA or Spirent. > + > + * Start creating a stream on packet generator, e.g. IXIA. Not sure mentioning IXIA or Spirent helps here. Same comment for the (nice) diagram. [Qian] *Start creating a stream on packet generator, usually hardware traff= ic generator can provide line rate traffic, so the NIC's best performance w= ill not be limited by the traffic generator. Hardware traffic generator is = like IXIA or Spirent. You can also use high speed software traffic generato= r to check the NIC's performance.=20 Thanks for the documentation effort.