On Mon, 3 Jun 2019 at 08:41, E. Madison Bray <erik.m.b...@gmail.com> wrote:
> Well, this would be a problem building OpenBLAS specifically :) > > I see this same problem mentioned here as well. Maybe it will help: > https://github.com/JuliaLang/julia/issues/30696 > > I'm not sure I understand the patch on that issue that mentions > cross-compiling to a 32-bit architecture. You're clearly not doing > that, and you do have the NO_AVX512 macro defined. it is first trying to build openblas for skylake architecture, and errors out. This could indicate an old as, a.k.a. gas (GNU assembler), which usually comes from binutils. Do you really have a skylake CPU, so it is not a misconfiguration by openblas? What is your as version? Then it tries to build for some dumb architecture, and errors out, perhaps as this option is not tested much, and so it broken in some way. > > I think the race condition issue seems more likely. I've also had > problems in the past with OpenBLAS's build system and running in > parallel. Try running just `make -j1 openblas` to build openblas by > itself first. > > Failing that, you can try using an existing system BLAS by running > `./configure --with-blas=atlas` and also setting SAGE_ATLAS_LIB to the > correct path (typically just SAGE_ATLAS_LIB=/usr/lib). It's > unfortunate that both steps are required; something we're working to > fix. > > Despite the spelling "--with-blas=atlas" this doesn't necessarily have > to use ATLAS BLAS, as it can detect other generic BLAS libs in many > cases. > > On Mon, Jun 3, 2019 at 1:20 AM Charles Campbell <cecn...@gmail.com> wrote: > > > > Hello: > > > > Dima: good catch; I have cp aliased to: /bin/cp -i. I've now unaliased > it and commented the alias command out of the file that normally sets it > (the latter step as a "paranoia" step) > > > > E.M.Bray: I've never had any success building sage, so no old stuff was > lying around. I simply got the tarball for sage 8.7 and extracted the > contents. > > > > I did a make distclean, then make -j10 which resulted in the log file > I've provided. > > Next try: make distclean; configure; make -j10 . Didn't work, either, > although there are some differences between the two log files. I've > attached it, too. > > > > Hope you can help! > > Charles Campbell > > The latest problem now appears to be with openblas; I've attached the > associated log. > > > > > > On Thursday, May 30, 2019 at 3:51:57 PM UTC-4, Charles Campbell wrote: > > > >> Been trying to build sage 8.7 under Scientific Linux (a redhat > derivative); actually, I've tried to do so on and off for several years > with no success ever. I've installed openblas.x86_64 already. > >> > >> From make: > >> > >> The following package(s) may have failed to build (not necessarily > >> during this run of 'make all-start'): > >> > >> * package: openblas-0.3.5.p0 > >> log file: > /home/cecbkup/SW/SAGE/sage-8.7/logs/pkgs/openblas-0.3.5.p0.log > >> build directory: > /home/cecbkup/SW/SAGE/sage-8.7/local/var/tmp/sage/build/openblas-0.3.5.p0 > >> > >> * package: pari_seadata_small-20090618.p0 > >> log file: > /home/cecbkup/SW/SAGE/sage-8.7/logs/pkgs/pari_seadata_small-20090618.p0.log > >> build directory: > /home/cecbkup/SW/SAGE/sage-8.7/local/var/tmp/sage/build/pari_seadata_small-20090618.p0 > >> > >> I have no idea where pari... comes from; yum -y list shows > >> > >> opari2.x86_64 1.1.2-3.el7 epel > >> paris-traceroute.x86_64 0.92-6.el7 epel > >> > >> Neither of which are "pari", although perhaps "opari..." is close. > >> > >> Regards, > >> Charles Campbell > > > > -- > > You received this message because you are subscribed to the Google > Groups "sage-devel" group. > > To unsubscribe from this group and stop receiving emails from it, send > an email to sage-devel+unsubscr...@googlegroups.com. > > To post to this group, send email to sage-devel@googlegroups.com. > > Visit this group at https://groups.google.com/group/sage-devel. > > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/b609243e-4f7e-45c9-8665-1757825ebfb0%40googlegroups.com > . > > For more options, visit https://groups.google.com/d/optout. > > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To post to this group, send email to sage-devel@googlegroups.com. > Visit this group at https://groups.google.com/group/sage-devel. > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/CAOTD34Ysh1y6-9u%2BXpqKqfBTCnNu7KhUn7AnvBZm93OhrmewWw%40mail.gmail.com > . > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/CAAWYfq0zb3xt1UUiNDrTtm9kagmhyr5dwUUBt3yinA6R%3DxFWfg%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
