From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Phil Yang (Arm Technology China)" Subject: Re: [PATCH v2 1/3] packet_ordering: add statistics for each worker thread Date: Sat, 30 Mar 2019 16:55:22 +0000 Message-ID: References: <1546508946-12552-1-git-send-email-phil.yang@arm.com> <1553856998-25394-2-git-send-email-phil.yang@arm.com> <3AEA2BF9852C6F48A459DA490692831F2A432EBB@irsmsx110.ger.corp.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Cc: "Hunt, David" , "Gavin Hu (Arm Technology China)" , Honnappa Nagarahalli , nd , nd To: "Pattan, Reshma" , "dev@dpdk.org" , "thomas@monjalon.net" Return-path: Received: from EUR04-HE1-obe.outbound.protection.outlook.com (mail-eopbgr70078.outbound.protection.outlook.com [40.107.7.78]) by dpdk.org (Postfix) with ESMTP id 22119239 for ; Sat, 30 Mar 2019 17:55:24 +0100 (CET) In-Reply-To: <3AEA2BF9852C6F48A459DA490692831F2A432EBB@irsmsx110.ger.corp.intel.com> Content-Language: en-US List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" > -----Original Message----- > From: Pattan, Reshma > Sent: Saturday, March 30, 2019 12:40 AM > To: Phil Yang (Arm Technology China) ; dev@dpdk.org; > thomas@monjalon.net > Cc: Hunt, David ; Gavin Hu (Arm Technology China) > ; Honnappa Nagarahalli > ; nd > Subject: RE: [PATCH v2 1/3] packet_ordering: add statistics for each work= er > thread >=20 >=20 >=20 > > -----Original Message----- > > From: Phil Yang [mailto:phil.yang@arm.com] > > > > The current implementation using '__sync' built-ins to synchronize > > statistics within worker threads. '__sync' built-ins functions are > > full barriers which will affect the performance, so add a per worker pa= ckets > statistics. > > > > Enable by option --insight-worker. > > >=20 > I don't feel the need of this new option to print per core stats. Any re= ason > for this? Hi Reshma, Thanks for your comment.=20 The per core stats aims at removing the '__sync' builtin full barrier in th= e worker thread.=20 It records the workload of each core (It shows the bottleneck core as well)= . Since the maximum core number may be more than 128, so disable the print= in default and add this new option for debugging use.=20 Anyway, if you insist there is no need to print that info out, I can remove= the option. But I think using per core stats will benefit performance. Thanks, Phil >=20 > Thanks, > Reshma