From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga02.intel.com (mga02.intel.com [134.134.136.20]) by dpdk.org (Postfix) with ESMTP id 29DB41B724 for ; Thu, 8 Feb 2018 12:24:35 +0100 (CET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga006.fm.intel.com ([10.253.24.20]) by orsmga101.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 08 Feb 2018 03:24:34 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.46,478,1511856000"; d="scan'208,217";a="202525654" Received: from irsmsx153.ger.corp.intel.com ([163.33.192.75]) by fmsmga006.fm.intel.com with ESMTP; 08 Feb 2018 03:24:29 -0800 Received: from irsmsx108.ger.corp.intel.com ([169.254.11.9]) by IRSMSX153.ger.corp.intel.com ([169.254.9.213]) with mapi id 14.03.0319.002; Thu, 8 Feb 2018 11:24:29 +0000 From: "O'Driscoll, Tim" To: "'ci@dpdk.org'" CC: 'Bob Noseworthy' , "Mcnamara, John" , 'Shepard Siegel' , 'Thomas Monjalon' , 'Erez Scop' , 'Shreyansh Jain' , "Xu, Qian Q" , "'pmacarth@iol.unh.edu'" , "'Matt Spencer'" , Gema Gomez , 'George Zhao' , "Mishra, Shishir" , 'Lixuming' , "Tkachuk, Georgii" , 'Trishan de Lanerolle' , Sean Campbell , Ali Alnubani , May Chen Thread-Topic: Minutes of DPDK Lab Meeting, February 6th Thread-Index: AdOgzay8dIpqexHkRwGTaBamYNFtNw== Date: Thu, 8 Feb 2018 11:24:29 +0000 Message-ID: <26FA93C7ED1EAA44AB77D62FBE1D27BAA4AC1F5A@IRSMSX108.ger.corp.intel.com> Accept-Language: en-IE, en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-titus-metadata-40: eyJDYXRlZ29yeUxhYmVscyI6IiIsIk1ldGFkYXRhIjp7Im5zIjoiaHR0cDpcL1wvd3d3LnRpdHVzLmNvbVwvbnNcL0ludGVsMyIsImlkIjoiZWZiMzk4YWQtMmI4YS00NGM5LTkyMDUtYzk2Njk1ZmI2Mjg4IiwicHJvcHMiOlt7Im4iOiJDVFBDbGFzc2lmaWNhdGlvbiIsInZhbHMiOlt7InZhbHVlIjoiQ1RQX05UIn1dfV19LCJTdWJqZWN0TGFiZWxzIjpbXSwiVE1DVmVyc2lvbiI6IjE3LjIuNS4xOCIsIlRydXN0ZWRMYWJlbEhhc2giOiJ1c1wvOFpiYWcydDBtS2ZOV0c3ZWhMc0VQWVwvZ0RmelQ0XC8zQUtBVlFDS3IzdUNtOE8rbDFMeWRmaDlscGowa3Y0In0= x-ctpclassification: CTP_NT dlp-product: dlpe-windows dlp-version: 11.0.0.116 dlp-reaction: no-action x-originating-ip: [163.33.239.182] Content-Type: multipart/alternative; boundary="_000_26FA93C7ED1EAA44AB77D62FBE1D27BAA4AC1F5AIRSMSX108gercor_" MIME-Version: 1.0 Subject: [dpdk-ci] Minutes of DPDK Lab Meeting, February 6th X-BeenThere: ci@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK CI discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 08 Feb 2018 11:24:36 -0000 --_000_26FA93C7ED1EAA44AB77D62FBE1D27BAA4AC1F5AIRSMSX108gercor_ Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Test Infrastructure: - Patrick sent email to the ci@dpdk.org list proposing = a JSON schema that all vendor's test cases should produce to allow easy ins= ertion into the database. See http://dpdk.org/ml/archives/ci/2018-February/= 000166.html. Let Patrick know if you have any comments on this. - Beginning to look at email reporting. This is a balance between providing= sufficient notification of results versus not overloading people with emai= l. Proposed that email is sent per vendor/test, and that it's only sent for= failures (results outside of specified tolerance). - Discussed what happens if a developer is notified that his patch degrades= performance on an architecture that they does not have access to. How do t= hey debug? Agreed that it will be the role of the maintainer for the affect= ed part of DPDK to make a recommendation on how the issue can be resolved. - Shreyansh will follow up with Patrick to make sure that the database sche= ma allows for storing relative results, not just absolute performance numbe= rs. We agreed previously that vendors should have the option to publish a) = no results, b) relative performance results versus previous test runs or c)= absolute performance data. The end goal is to get to option c for all plat= forms, but while hardware and software is being tuned for a new platform we= may need to use a or b. Hardware Availability: - For the Intel hardware, the DTS issue has been resolved and the hardware = should be ready to ship around the end of this week. - For Mellanox, they're still seeing some issues with the single core perfo= rmance tests using TRex and DTS. Ali will follow up with Qian on this to se= e if she has any suggestions. - For NXP, the hardware is now in their lab. DTS integration is expected to= take about a month. Next Meeting: - Tuesday Feb 20th. --_000_26FA93C7ED1EAA44AB77D62FBE1D27BAA4AC1F5AIRSMSX108gercor_ Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable

Test Infrastructure:

- Patrick sent email to the ci@dpdk.org list proposing a JSON schema that all vendor’s test c= ases should produce to allow easy insertion into the database. See http:/= /dpdk.org/ml/archives/ci/2018-February/000166.html. Let Patrick know if= you have any comments on this.

- Beginning to look at email reporting. This is a= balance between providing sufficient notification of results versus not ov= erloading people with email. Proposed that email is sent per vendor/test, a= nd that it’s only sent for failures (results outside of specified tolerance).

- Discussed what happens if a developer is notifi= ed that his patch degrades performance on an architecture that they does no= t have access to. How do they debug? Agreed that it will be the role of the= maintainer for the affected part of DPDK to make a recommendation on how the issue can be resolved.

- Shreyansh will follow up with Patrick to make s= ure that the database schema allows for storing relative results, not just = absolute performance numbers. We agreed previously that vendors should have= the option to publish a) no results, b) relative performance results versus previous test runs or c) absolute p= erformance data. The end goal is to get to option c for all platforms, but = while hardware and software is being tuned for a new platform we may need t= o use a or b.

 

Hardware Availability:

- For the Intel hardware, the DTS issue has been = resolved and the hardware should be ready to ship around the end of this we= ek.

- For Mellanox, they’re still seeing some i= ssues with the single core performance tests using TRex and DTS. Ali will f= ollow up with Qian on this to see if she has any suggestions.

- For NXP, the hardware is now in their lab. DTS = integration is expected to take about a month.

 

Next Meeting:

- Tuesday Feb 20th.

 

--_000_26FA93C7ED1EAA44AB77D62FBE1D27BAA4AC1F5AIRSMSX108gercor_--