From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ricardo Wurmus Subject: Re: [PATCH] Attempt to fix OpenBLAS on MIPS. Date: Thu, 24 Sep 2015 10:27:01 +0200 Message-ID: References: <87zj2p1r0h.fsf@netris.org> <87vbdd1mox.fsf@netris.org> <20150722171952.GA8708@debian> <87wpxsyn80.fsf@netris.org> <87si8gyn11.fsf@netris.org> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Return-path: Received: from eggs.gnu.org ([2001:4830:134:3::10]:51161) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Zf1s3-00041a-9A for guix-devel@gnu.org; Thu, 24 Sep 2015 04:27:16 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Zf1rz-0006Wx-8Z for guix-devel@gnu.org; Thu, 24 Sep 2015 04:27:15 -0400 Received: from pegasus.bbbm.mdc-berlin.de ([141.80.25.20]:41383) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Zf1ry-0006Vx-Ra for guix-devel@gnu.org; Thu, 24 Sep 2015 04:27:11 -0400 In-Reply-To: List-Id: "Development of GNU Guix and the GNU System distribution." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org Sender: guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org To: Mark H Weaver Cc: guix-devel Hi Mark, >>>> On Wed, Jul 22, 2015 at 05:59:06PM +0200, Ricardo Wurmus wrote: >>>>> The test target appears to only do something if TARGET is not set: >>>>> Since we do pass TARGET=3DSICORTEX the tests are skipped. I don=E2= =80=99t know if >>>>> there=E2=80=99s a good reason for this. >>>> >>>> It looks a bit like a flawed logic to skip tests in cross building >>>> environments. Maybe it is worth asking upstream. >>> >>> I removed this logic (see attached patch), and the tests passed excep= t >>> for one: >> >> Sorry, I didn't look closely enough. Actually, on MIPS the build late= r >> aborted due to 1 hour of silence in the middle of the test suite: >> >> --8<---------------cut here---------------start------------->8--- >> OPENBLAS_NUM_THREADS=3D2 ./sblat2 < ./sblat2.dat >> OPENBLAS_NUM_THREADS=3D1 OMP_NUM_THREADS=3D1 ./dblat3 < ./dblat3.dat >> OPENBLAS_NUM_THREADS=3D1 OMP_NUM_THREADS=3D1 ./cblat3 < ./cblat3.dat >> OPENBLAS_NUM_THREADS=3D1 OMP_NUM_THREADS=3D1 ./zblat3 < ./zblat3.dat >> rm -f ?BLAT3.SUMM >> OPENBLAS_NUM_THREADS=3D2 ./sblat3 < ./sblat3.dat >> building of `/gnu/store/974lvn8v6sxvqhlm57n8x6fiqv6xn6dh-openblas-0.2.= 14.drv' timed out after 3600 seconds of silence >> @ build-failed /gnu/store/974lvn8v6sxvqhlm57n8x6fiqv6xn6dh-openblas-0.= 2.14.drv - timeout >> note: keeping build directory `/tmp/nix-build-openblas-0.2.14.drv-0' >> killing process 16081 >> guix build: error: build failed: build of `/gnu/store/974lvn8v6sxvqhlm= 57n8x6fiqv6xn6dh-openblas-0.2.14.drv' failed >> --8<---------------cut here---------------end--------------->8--- > > Have you been able to try this again with a larger silent timeout value= ? > I would like to know if the patch (+ explicitly enabling tests on MIPS) > is sufficient to fix OpenBLAS on MIPS, or if I should discuss more with > upstream. I=E2=80=99m just pinging again, because this patch still sits in my queue= . Can we use this patch to fix the build of OpenBLAS on MIPS? As OpenBLAS is an (indirect) input of many failing packages on MIPS I=E2=80= =99m interested in seeing this merged. ~~ Ricardo