From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:43154) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1dcdVT-00075K-NS for guix-patches@gnu.org; Tue, 01 Aug 2017 16:11:08 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1dcdVO-0004pr-LA for guix-patches@gnu.org; Tue, 01 Aug 2017 16:11:07 -0400 Received: from debbugs.gnu.org ([208.118.235.43]:34637) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1dcdVO-0004pU-Ft for guix-patches@gnu.org; Tue, 01 Aug 2017 16:11:02 -0400 Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1dcdVO-0001in-0U for guix-patches@gnu.org; Tue, 01 Aug 2017 16:11:02 -0400 Subject: [bug#27850] gnu: mpi: openmpi: Don't enable thread-multiple Resent-Message-ID: From: Dave Love References: <87wp6thqjc.fsf@i-ulialbion.it.manchester.ac.uk> <87379c4oxs.fsf@gnu.org> <87o9s0bhro.fsf@i-ulialbion.it.manchester.ac.uk> <87lgn3skmm.fsf@inria.fr> <87tw1r9sss.fsf@i-ulialbion.it.manchester.ac.uk> <87bmnzxk50.fsf@inria.fr> Date: Tue, 01 Aug 2017 21:10:23 +0100 In-Reply-To: <87bmnzxk50.fsf@inria.fr> ("Ludovic \=\?iso-8859-1\?Q\?Court\=E8s\?\= \=\?iso-8859-1\?Q\?\=22's\?\= message of "Tue, 1 Aug 2017 19:39:23 +0200") Message-ID: <874ltr9hhs.fsf@i-ulialbion.it.manchester.ac.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: guix-patches-bounces+kyle=kyleam.com@gnu.org Sender: "Guix-patches" To: Ludovic =?UTF-8?Q?Court=C3=A8s?= Cc: 27850@debbugs.gnu.org Ludovic Court=C3=A8s writes: >> Maybe, but what about the non-ABI compatibility I expect there is? (I >> don't know whether there's still any penalty from thread-multiple >> anyhow; I guess not, as I see it's not the default.)=20 > > I propose this because you had written that the =E2=80=9Cperformance pena= lty for > thread-multiple is supposed to be mitigated in the most recent openmpi.= =E2=80=9D > If it=E2=80=99s not, then fine. I don't know the value of "mitigated". I could ask or, better, measure when I get back from holiday (at least micro-benchmarks over Infiniband). >> If anyone's using it seriously, I'd have thought effort would be better >> spent on support for SLURM (as it's in Guix) and supporting >> high-performance fabrics (which I started on). > > You already mentioned openfabrics a couple of times I think. Mentioning > it more won=E2=80=99t turn it into an actual package. :-) It=E2=80=99s = on my to-do > list, I guess it=E2=80=99s on yours too, so we=E2=80=99ll get there. Sure. It's only what seems important. I'll post what I've got, but if someone else is doing it, fine, and I won't duplicate effort. > What do you have in mind for SLURM? There's integration with SLURM (--with-slurm), PBS/Torque, and LSF (or, I guess, Open Lava in the free world). I don't know much about them, but they build MCA modules. Unlike the gridengine support, they link against libraries for the resource managers, so you want them to be add-ons which are only installed when required (not like the Fedora packaging). > As for =E2=80=9Cusing it seriously=E2=80=9D, I think this is a needlessly= aggressive way > to express your frustration. I'm sorry I'm mis-communicating trans-Manche, at least. It wasn't meant like that at all and I'll try to be more careful. Please assume I'm a friendly hacker, even if I have strong opinions, which I hope I can justify! > People *are* using Guix =E2=80=9Cseriously=E2=80=9D in HPC I meant openmpi, not Guix generally. "Seriously" meant applications which are communication-intensive (like the latency-sensitive DFT applications). > already, but (1) different application domains emphasize different > aspects of =E2=80=9CHPC=E2=80=9D, and (2) there=E2=80=99s on-going work t= o improve Guix for HPC > and your feedback is invaluable here. I hope I can give useful feedback, and any criticism is meant constructively. However, I'm not representative of UK HPC people -- happier to use functional Scheme than Python, and believing in packaging for a start! Happy hacking.