From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 03AC3ECAAA1 for ; Fri, 28 Oct 2022 08:34:49 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229908AbiJ1Ier (ORCPT ); Fri, 28 Oct 2022 04:34:47 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:33798 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229629AbiJ1Ien (ORCPT ); Fri, 28 Oct 2022 04:34:43 -0400 Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 861EB1C42D for ; Fri, 28 Oct 2022 01:34:41 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1666946082; x=1698482082; h=from:to:cc:subject:references:date:in-reply-to: message-id:mime-version; bh=I7+l1li97AG1W4nCdyskwIa37THe0xzjljy5/l3afeQ=; b=KQRhi/nIPXbyx53oWRDRqu2cA9MhLAeG0TQrztfQ6uyb6ItQ1zSs/lLC 4YI/UkxHBHGuM2VXpxE1h7uqs6oPKNpCBYBDb24T3j6TejBUnsEAnAXR4 2ZCGmOShB3ZZfcljjPTW/+hNPlRb7lBrn5s7FToQErET+iaO1S7H5ELe3 OvMvgYtTmXwo3uOlXFErMDJYuHv6QerA76VOL7nj1KZP+3k/FAD15IPCD GG21PPh5yIiMiqO2YknRAzZMG1Hekl1bpS8AeVR+8RK7Azh1SB0sh0Pcl 7erpAHPqcebpFmiCkvmYt5jrLOoeQGJCPv1QMsMejY2carzpYhe2U2lkH Q==; X-IronPort-AV: E=McAfee;i="6500,9779,10513"; a="335083761" X-IronPort-AV: E=Sophos;i="5.95,220,1661842800"; d="scan'208";a="335083761" Received: from orsmga002.jf.intel.com ([10.7.209.21]) by fmsmga101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 28 Oct 2022 01:34:41 -0700 X-IronPort-AV: E=McAfee;i="6500,9779,10513"; a="632708139" X-IronPort-AV: E=Sophos;i="5.95,220,1661842800"; d="scan'208";a="632708139" Received: from yhuang6-desk2.sh.intel.com (HELO yhuang6-desk2.ccr.corp.intel.com) ([10.238.208.55]) by orsmga002-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 28 Oct 2022 01:34:37 -0700 From: "Huang, Ying" To: Bharata B Rao Cc: Aneesh Kumar K V , , , Andrew Morton , Alistair Popple , "Dan Williams" , Dave Hansen , Davidlohr Bueso , Hesham Almatary , Jagdish Gediya , Johannes Weiner , Jonathan Cameron , Michal Hocko , Tim Chen , Wei Xu , Yang Shi Subject: Re: [RFC] memory tiering: use small chunk size and more tiers References: <20221027065925.476955-1-ying.huang@intel.com> <578c9b89-10eb-1e23-8868-cdd6685d8d4e@linux.ibm.com> <877d0kk5uf.fsf@yhuang6-desk2.ccr.corp.intel.com> <59291b98-6907-0acf-df11-6d87681027cc@linux.ibm.com> <8735b8jy9k.fsf@yhuang6-desk2.ccr.corp.intel.com> <0d938c9f-c810-b10a-e489-c2b312475c52@amd.com> Date: Fri, 28 Oct 2022 16:33:48 +0800 In-Reply-To: <0d938c9f-c810-b10a-e489-c2b312475c52@amd.com> (Bharata B. Rao's message of "Fri, 28 Oct 2022 13:34:46 +0530") Message-ID: <87tu3oibyr.fsf@yhuang6-desk2.ccr.corp.intel.com> User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/27.1 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=ascii Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Bharata B Rao writes: > On 10/28/2022 11:16 AM, Huang, Ying wrote: >> If my understanding were correct, you think the latency / bandwidth of >> these NUMA nodes will near each other, but may be different. >> >> Even if the latency / bandwidth of these NUMA nodes isn't exactly same, >> we should deal with that in memory types instead of memory tiers. >> There's only one abstract distance for each memory type. >> >> So, I still believe we will not have many memory tiers with my proposal. >> >> I don't care too much about the exact number, but want to discuss some >> general design choice, >> >> a) Avoid to group multiple memory types into one memory tier by default >> at most times. > > Do you expect the abstract distances of two different types to be > close enough in real life (like you showed in your example with > CXL - 5000 and PMEM - 5100) that they will get assigned into same tier > most times? > > Are you foreseeing that abstract distance that get mapped by sources > like HMAT would run into this issue? Only if we set abstract distance chunk size large. So, I think that it's better to set chunk size as small as possible to avoid potential issue. What is the downside to set the chunk size small? Best Regards, Huang, Ying