From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga06.intel.com (mga06.intel.com [134.134.136.31]) by dpdk.org (Postfix) with ESMTP id 340AB1B773 for ; Fri, 9 Feb 2018 16:55:31 +0100 (CET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga003.jf.intel.com ([10.7.209.27]) by orsmga104.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 09 Feb 2018 07:55:29 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.46,483,1511856000"; d="scan'208";a="26127226" Received: from fmsmsx103.amr.corp.intel.com ([10.18.124.201]) by orsmga003.jf.intel.com with ESMTP; 09 Feb 2018 07:55:29 -0800 Received: from fmsmsx157.amr.corp.intel.com (10.18.116.73) by FMSMSX103.amr.corp.intel.com (10.18.124.201) with Microsoft SMTP Server (TLS) id 14.3.319.2; Fri, 9 Feb 2018 07:55:28 -0800 Received: from fmsmsx117.amr.corp.intel.com ([169.254.3.243]) by FMSMSX157.amr.corp.intel.com ([169.254.14.60]) with mapi id 14.03.0319.002; Fri, 9 Feb 2018 07:55:28 -0800 From: "Wiles, Keith" To: Theano CC: "users@dpdk.org" , Christopher Hodsdon Thread-Topic: [dpdk-users] Throughput dropping by half with 1B increase in packet size Thread-Index: AQHTobw5Xuf0+DZaS0KjF/6dfjHGM6Ocv3KA Date: Fri, 9 Feb 2018 15:55:28 +0000 Message-ID: <006F812B-472C-492F-9E0A-30D1E430933E@intel.com> References: In-Reply-To: Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.252.140.227] Content-Type: text/plain; charset="us-ascii" Content-ID: <18C71DC27A1087418E7F2A8F7F715F4E@intel.com> Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Subject: Re: [dpdk-users] Throughput dropping by half with 1B increase in packet size X-BeenThere: users@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK usage discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 09 Feb 2018 15:55:32 -0000 > On Feb 9, 2018, at 9:39 AM, Theano wrote: >=20 > Hi All, >=20 > We are eventually going to build an application over DPDK, and we've > started out with a very simple pktgen-like application that sends and > receives packets to a different machine running l2fwd. >=20 > With packets up to 64B (60B headers and payload + 4B of FCS), l2fwd shows > line rate packet forwarding of this application's packets. As soon as the > packet size increases to 65B, the throughput (in pps) halves. With > dpdk-pktgen we observe a much more graceful decline in packet throughput = as > packet size increases. >=20 > I examined the dpdk-pktgen code and can't find any significant difference= s > between how it constructs its packets vs. how we do it, but of course I'm > missing something. >=20 > Has anyone seen this before and/or has any ideas on how to identify the > problem? Adding one byte will push you to have two cache lines used and that can red= uce performance. Make sure the packets are aligned to a 64 byte boundary. Other then that I see no reason for the 1/2 performance without comparing t= he code. Pktgen-dpdk I spent a fair amount of time looking at the performan= ce, but I also had to play with burst size and the number of packet sent to= the NIC they too need to be cached aligned or multiples of a cache line. >=20 > We are using DPDK version 17.02.1; please let me know if any other > information would be helpful. > Thanks, > Theano. Regards, Keith