>From a bash shell where you build Sage: export OPENBLAS_CONFIGURE="TARGET=ATOM" MAKE="make -j5 -l3.5"
should set the important variables. This is the MAKE I use for a very, very old 4 core machine. The OPENBLAS_CONFIGURE variable is alluded to in build/pkgs/openblas/spkg-check but I find it no where else. Somewhere in the openblas code there is a listing of supported architectures. I don't know how many ATOM processor variations there are? And getting the openblas code to work properly is probably an upstream issue. On Tuesday, April 11, 2017 at 2:23:42 PM UTC-5, Ackbach wrote: > > All right! Sage 7.6 compiled correctly, so far as I know. I haven't run > the tests, but it opens into a notebook in Firefox, and correctly computes > > integrate(e^(-x^2),x,-infinity,infinity) > > Is there a way to combine the OPENBLAS_CONFIGURE command with the -j4 > option for using multiple cores to compile? It did take a VERY long time to > compile - I think it was about 8 hours. > > In any case, targeting an ATOM processor seems like a common-enough > occurrence that it deserves a mention in the manual, if it's not there > already (I haven't looked, and please don't shoot me). > > Thanks much! > > Cheers. > > On Tuesday, April 11, 2017 at 6:50:10 AM UTC-4, Ackbach wrote: >> >> 7.5.1 failed to build. Looks like the same error as before. I'll try >> Steven's linked suggestion with 7.6, and see if that flies. >> >> Thanks for all your help, by the way! >> >> One suggestion: if the OPENBLAS_CONFIGURE="TARGET=ATOM" make command >> builds everything correctly, then maybe the next version of Sage could >> automate processor detection and do the right switches on that basis? >> >> Cheers. >> >> On Tuesday, April 11, 2017 at 12:45:44 AM UTC-4, Steven Trogdon wrote: >>> >>> It is curious that 7.5.1 would build but not 7.6. You might get some >>> mileage with >>> >>> OPENBLAS_CONFIGURE="TARGET=ATOM" make >>> >>> See this thread: >>> >>> https://groups.google.com/forum/#!topic/sage-release/3QJoAgg9bgo >>> >>> On Monday, April 10, 2017 at 8:00:51 PM UTC-5, Ackbach wrote: >>>> >>>> I'm afraid I don't have that data. I will say I've never been able to >>>> compile Sage 7.6 the first time. I have definitely been able to compile >>>> 7.5.1 and earlier the first time, though sometimes there are still errors. >>>> I can try to compile 7.5.1 on this machine and get back to you how it >>>> works. >>>> >>>> On Monday, April 10, 2017 at 4:17:38 PM UTC-4, François wrote: >>>>> >>>>> Did it start happening with openblas 0.2.19 or did it suddenly happen >>>>> one release build openblas 0.2.19 successfully and the next didn’t? >>>>> I suspect it is a CPU detection problem, we have seen similar logs >>>>> before on atom chips. >>>>> >>>>> François >>>>> >>>>> > On 11/04/2017, at 07:45, Ackbach <ack...@gmail.com> wrote: >>>>> > >>>>> > Having issues building the same package. Only I get a different log >>>>> file. I always prefer building Sage from scratch for performance, but >>>>> have >>>>> lately been having trouble doing so, as Sage keeps failing to build. >>>>> Background: >>>>> > >>>>> > OS: Debian 8.7, 64-bit. >>>>> > Sage Version: 7.6. >>>>> > HW: Lenovo ThinkPad 11e, 120GB SSD, 8GB RAM, Intel Celeron N2940, >>>>> with 4 cpu cores (as evidenced by the cat /proc/cpuinfo command). >>>>> > >>>>> > I follow the instructions for [installing from source]( >>>>> http://doc.sagemath.org/html/en/installation/source.html), and I >>>>> execute the MAKE='make -j4' make command. The compile gets quite a ways >>>>> into the compile. It finishes compiling maxima, in fact, and then I get >>>>> the >>>>> following error message: >>>>> > >>>>> > Error building Sage. >>>>> > >>>>> > The following package(s) may have failed to build (not necessarily >>>>> during this run of 'make all'): >>>>> > >>>>> > * package: openblas-0.2.19.p0 >>>>> > log file: >>>>> /home/akeister/Downloads/sage-7.6/logs/pkgs/openblas-0.2.19.p0.log >>>>> > build directory: >>>>> /home/akeister/Downloads/sage-7.6/local/var/tmp/sage/build/openblas-0.2.19.p0 >>>>> >>>>> >>>>> > >>>>> > I have attached the indicated log file, where it says to contact >>>>> this group. You can see that the issue is different from the OP, but it's >>>>> still a problem compiling that package. >>>>> > >>>>> > Ideas? >>>>> > >>>>> > Thanks for your time! >>>>> > >>>>> > Cheers, >>>>> > Adrian >>>>> > >>>>> > On Wednesday, October 12, 2016 at 12:43:20 PM UTC-4, Ethan Petersen >>>>> wrote: >>>>> > Hello all, >>>>> > >>>>> > I'm building sage on a Lenovo Thinkpad W540 with Ubuntu 16.04, and >>>>> there was an "Error building OpenBLAS"/"Error installing package >>>>> openblas-0.2.19" with a suggestion to email this google group with the >>>>> attached log file. If anyone has ideas on solving this issue, please let >>>>> me >>>>> know! >>>>> > >>>>> > Thanks, >>>>> > >>>>> > Ethan >>>>> > >>>>> > -- >>>>> > 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+...@googlegroups.com. >>>>> > To post to this group, send email to sage-...@googlegroups.com. >>>>> > Visit this group at https://groups.google.com/group/sage-devel. >>>>> > For more options, visit https://groups.google.com/d/optout. >>>>> > <openblas-0.2.19.p0.log> >>>>> >>>>> -- 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. For more options, visit https://groups.google.com/d/optout.
