Return-Path: Received: from mail-it0-f52.google.com ([209.85.214.52]:36163 "EHLO mail-it0-f52.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751157AbeDQOvA (ORCPT ); Tue, 17 Apr 2018 10:51:00 -0400 Received: by mail-it0-f52.google.com with SMTP id n81-v6so9837498ita.1 for ; Tue, 17 Apr 2018 07:51:00 -0700 (PDT) Subject: Re: [PATCH v8 0/9] NFSD support for async COPY To: Olga Kornievskaia , "J. Bruce Fields" Cc: Olga Kornievskaia , Christoph Hellwig , linux-nfs , Jorge Mora References: <20180413170158.17589-1-kolga@netapp.com> <20180414072202.GA6514@infradead.org> <20180416214522.GC2634@parsley.fieldses.org> <20180417065203.GA15145@infradead.org> <20180417134141.GB10291@parsley.fieldses.org> <20180417135708.GD10291@parsley.fieldses.org> From: Anna Schumaker Message-ID: Date: Tue, 17 Apr 2018 10:50:57 -0400 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8 Sender: linux-nfs-owner@vger.kernel.org List-ID: On 04/17/2018 10:13 AM, Olga Kornievskaia wrote: > On Tue, Apr 17, 2018 at 10:08 AM, Olga Kornievskaia wrote: >> On Tue, Apr 17, 2018 at 9:57 AM, J. Bruce Fields wrote: >>> On Tue, Apr 17, 2018 at 09:43:45AM -0400, Olga Kornievskaia wrote: >>>>> On Apr 17, 2018, at 9:41 AM, J. Bruce Fields wrote: >>>>> Christoph said: "I'd like to see numbers for actual, real use cases." >>>>> Yes, I'd love to see that too. >>>>> >>>>> Is it possible to get testing on real hardware that we'd expect people >>>>> to use, with sufficient details about the setup that someone else could >>>>> reproduce the results? >>>> >>>> The numbers we presented was on real hardware. If you have any more questions regarding the setup please ask. >>> >>> Argh, sorry, you had to remind me once already: >>> >>> https://www.ietf.org/mail-archive/web/nfsv4/current/msg15379.html >> >> Yes that was for the "inter" copy. >> >> Besides those numbers, Jorge also did the numbers for the asynchronous >> vs synchronous intra copy. Are you interested in them as well? >> Hopefully Jorge still has them. > > I found this in my emails. There is degradation for file size <- 8MB > and increase in performance for >16MB Right now we have an arbitrary sychronous copy limit of 4MB. Could we bump that to 8MB, and then have the server choose to do copies less than this limit as synchronous? As for theconfig optionquestion, my preference is to at least have the client side enabled by default. Anna > > Here are the numbers (each test is run 10 times and the average is displayed): > > /home/mora/logs/nfstest_ssc_20171015174051.log: synchronous SSC > /home/mora/logs/nfstest_ssc_20171015171323.log: asynchronous SSC > > Performance degradation for file size <= 8MB: > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 1KB file: 0.162018680573 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 1KB file: 0.636984395981 seconds > 74.56% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 2KB file: 0.177096033096 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 2KB file: 0.583548688889 seconds > 69.65% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 4KB file: 0.207866406441 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 4KB file: 0.782712388039 seconds > 73.44% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 8KB file: 0.201173448563 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 8KB file: 0.692702102661 seconds > 70.96% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 16KB file: 0.216166472435 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 16KB file: 0.56289691925 seconds > 61.60% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 32KB file: 0.185425901413 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 32KB file: 0.65691485405 seconds > 71.77% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 64KB file: 0.238445782661 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 64KB file: 0.525265741348 seconds > 54.60% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 128KB file: 0.166219258308 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 128KB file: 0.69602959156 seconds > 76.12% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 256KB file: 0.192826724052 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 256KB file: 0.556627941132 seconds > 65.36% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 512KB file: 0.230020523071 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 512KB file: 0.496951031685 seconds > 53.71% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 1MB file: 0.225317955017 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 1MB file: 0.50447602272 seconds > 55.34% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 2MB file: 0.233719205856 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 2MB file: 0.570275163651 seconds > 59.02% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 4MB file: 0.285868096352 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 4MB file: 0.656079149246 seconds > 56.43% > performance degradation by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 8MB file: 0.509396600723 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 8MB file: 0.696055078506 seconds > 26.82% > performance degradation by async > > > Performance improvement for file size >= 16MB: > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 16MB file: 0.960887002945 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 16MB file: 0.817601919174 seconds > 14.91% > performance improvement by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 32MB file: 1.58274662495 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 32MB file: 1.24578406811 seconds > 21.29% > performance improvement by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 64MB file: 2.54690010548 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 64MB file: 1.58639280796 seconds > 37.71% > performance improvement by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 128MB file: 4.2253510952 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 128MB file: 2.58433949947 seconds > 38.84% > performance improvement by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 256MB file: 7.71305201054 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 256MB file: 4.43859291077 seconds > 42.45% > performance improvement by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 512MB file: 14.6445164919 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 512MB file: 8.18448216915 seconds > 44.11% > performance improvement by async > > /home/mora/logs/nfstest_ssc_20171015174051.log: PASS: SSC copy of > 1GB file: 28.7619983673 seconds > /home/mora/logs/nfstest_ssc_20171015171323.log: PASS: SSC copy of > 1GB file: 16.1399238825 seconds > 43.88% > performance improvement by async > -- > To unsubscribe from this list: send the line "unsubscribe linux-nfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html >