On Sat, Jun 6, 2009 at 2:25 AM, Anthony David<ajdavi...@gmail.com> wrote: > > On Sat, Jun 6, 2009 at 8:05 AM, Dr. David Kirkby<david.kir...@onetel.net> > wrote: >> >> I've downloaded gcc 4.4.0 with a view to building this to try to get rid >> of the internal compiler bug on the T2. >> > > As a general question to the list: > > I understand that sage is designed to be built by just typing make on > any platform.
That's not true. You should read README.txt in the root directory of the Sage tarball, which lists supported platforms. > Is there a general problem when people use compilers > other than gcc and, it seems, a particular gcc version? The following > have been my casual experience. Nobody has ever even once build Sage using any compiler except GCC. And even then, one must not use too buggy of a version of GCC. Sage contains much complicated highly optimized code, which exposes compiler bugs. That said, we definitely would welcome any help from anybody to make it possible to build Sage with other compilers. E.g., there is a project to build Sage using MSVC on Windows: http://windows.sagemath.org > > Attempts to build sage 4.0 on ia64 SLES10SP1 failed using the intel > compiler. Problems with gmp-mpir were overcome by unloading the Intel > compiler module from my shell ( icc 9.1.051). gcc 4,1,2 was then used > on all compiles and it then sailed on happily until polybori when that > failed, > > [The ia64 issues have not been high on the list of priorities because > I do not have a need for my model to run on that architecture. I had a > test box I had reassembled and wanted to exercise it so a sage test > build seemed like a good idea.] I regularly build Sage on itanium using GCC-4.3.x. It's a standard supported platform, and all tests pass. Sage built with GCC-4.4 doesn't 100% work in sage-4.0 or earlier, but works fine with sage-4.0.1.rc3 (soon to be released). > > Anthony David > > >> I looked at an install.log of one of my builds of Sage using the Solaris >> binaries from Micheal and see the compiler was configured with: >> >> >> GCC Version >> gcc -v >> Using built-in specs. >> Target: sparc-sun-solaris2.10 >> Configured with: ../gcc-4.3.2/configure >> --prefix=/home/mabshoff/sparc-solaris-toolchain/ >> --enable-languages=c,c++,fortran --with-gnu-ld >> --with-ld=/home/mabshoff/sparc-solaris-toolchain/bin/ld --with-gnu-as >> --with-as=/home/mabshoff/sparc-solaris-toolchain/bin/as >> --with-gmp=/usr/local/gmp-4.2.3/sparc-SunOS-gcc-4.3.1-abi32/ >> --with-mpfr=/usr/local/mpfr-2.3.2/sparc-SunOS-gmp-4.2.3-gcc-4.3.1-abi32/ >> >> >> >> >> Is it possible to move those >> >> /usr/local/gmp-4.2.3/sparc-SunOS-gcc-4.3.1-abi32/ and >> /usr/local/mpfr-2.3.2/sparc-SunOS-gmp-4.2.3-gcc-4.3.1-abi32/ >> >> to somewhere on the T2, so I can configure gcc to use them and build gcc >> 4.4.0? >> >> >> If those are known to work (on most machines anyway), I'd rather use >> them than download the sources and compile gmp and mpfr. >> >> >> >> > >> > > > > -- William Stein Associate Professor of Mathematics University of Washington http://wstein.org --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
