Return-Path: linux-nfs-owner@vger.kernel.org Received: from mail-qc0-f181.google.com ([209.85.216.181]:55559 "EHLO mail-qc0-f181.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754356AbaGAUNU (ORCPT ); Tue, 1 Jul 2014 16:13:20 -0400 Received: by mail-qc0-f181.google.com with SMTP id x13so8832553qcv.40 for ; Tue, 01 Jul 2014 13:13:19 -0700 (PDT) From: Jeff Layton Date: Tue, 1 Jul 2014 16:13:17 -0400 To: "J. Bruce Fields" Cc: Jeff Layton , Trond Myklebust , Christoph Hellwig , Linux NFS Mailing List Subject: Re: [PATCH v2 000/117] nfsd: eliminate the client_mutex Message-ID: <20140701161317.0761839e@tlielax.poochiereds.net> In-Reply-To: <20140701200135.GE14728@fieldses.org> References: <1403810017-16062-1-git-send-email-jlayton@primarydata.com> <20140630125142.GA32089@infradead.org> <20140630085934.2bf86ba0@tlielax.poochiereds.net> <20140630193237.GA11935@fieldses.org> <20140630162014.20e63e1a@tlielax.poochiereds.net> <20140630163647.5227ac55@tlielax.poochiereds.net> <20140701101034.520b04bb@tlielax.poochiereds.net> <20140701144631.4bf0e9b5@tlielax.poochiereds.net> <20140701200135.GE14728@fieldses.org> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Sender: linux-nfs-owner@vger.kernel.org List-ID: On Tue, 1 Jul 2014 16:01:35 -0400 "J. Bruce Fields" wrote: > On Tue, Jul 01, 2014 at 02:46:31PM -0400, Jeff Layton wrote: > > Ok, I haven't tested that yet, but I did put together a f20 machine on > > bare metal to act as a server. Here are some results from a test that > > forks 128 child processes and has each one create and close 1024 files. > > Client is 3.14.8-200.fc20.x86_64 and using NFSv4.1: > > > > With a stock Fedora kernel on the server running 3.14.8-200.fc20.x86_64: > > > > $ time sudo ./opentest -n 128 -l 1024 /mnt/palma > > > > real 59m5.022s > > user 0m1.401s > > sys 0m17.561s > > > > > > ...with the patched kernel 3.16.0-rc2.jlayton.1+ (basically Bruce's > > for-3.17 branch with the rest of my patches piled on): > > > > $ time sudo ./opentest -n 128 -l 1024 /mnt/palma > > > > real 4m19.060s > > user 0m1.259s > > sys 0m13.059s > > > > ...so around a factor of 13x speedup on parallel creates. Again, this > > is not a terribly scientific benchmark, but it does suggest that this > > patchset helps immensely. > > Neat-o. > > It might also be interesting to compare with NFSv3? > > I wonder what a first good test of the smp scalability would be. > > --b. Yeah, it is neato. And just to be sure that it wasn't some other random change between 3.14 and 3.16, I popped off all of my patches, rebuilt nfsd.ko and reran the test with everything else the same: $ time sudo ./opentest -n 128 -l 1024 /mnt/palma real 58m50.564s user 0m1.513s sys 0m18.963s ...so it really does appear that these changes make the difference. -- Jeff Layton