Tristram Scott wrote: > Hi Dave, > > I might be interested in working on this, but only if there is a chance > that it will be possible to get it to compile using the SunStudio > compilers. I see that at the moment the only work being done is under gcc. > > Anyway, I have sun hardware, ranging from a SunFire 280R (similar to > your Blade 2000) to a couple of T1000 machines, as well as quite a > number of x86 machines, both AMD and Intel based. They are all running > Solaris 10, mostly close to the latest release. > > As a starting point, should I just grab the source code and type gmake > at the prompt, or do you know of a number of steps required to make it > work? The gcc I have installed is the one in /usr/sfw/bin/gcc: > gcc version 3.4.3 (csl-sol210-3_4-branch+sol_rpath) > > I guess I could look at building a different version of gcc, if > required, but that always seems like hard work, and the Sun compilers > seem to give me in the order of 20% speed improvement. > > Regards, > > Tristram
Hi Tristram, than you for replying to my request for help with the Solaris port of Sage, that I posted on a few relevant newsgroups (sci.math.symbolic, comp.unix.solaris, comp.soft-sys.math.maple and comp.soft-sys.matlab) I'm forwarding your reply to sage-devel mailing list. I suggest you subscribe to that. Sage will not currently build with Sun's compiler suite. The Sage source would need some changes, but as I said on my newsgroup post, it will not reliably build on gcc either, though several people, myself included, have managed to build it. William Stein can tell you more about his plans for Sage (I'm sure he will reply to you very soon). I believe both use of the Sun compiler and 64-bit are aims. I think particularly with 64-bit, the Sun compiler would be better. I would like to get it working with Sun's compiler too. There are obvious advantages. The problem is, Sage makes use of a large number of bits of different software. I've just done as 'ls' of the standard packages below. You can see there are a quite a few. Some unfortunately are better written than others. Some authors take the view that 'everyone uses gcc'. As you are no doubt aware, the Sun compilers are more fussy than gcc in what they accept. Clearly one useful contribution would be to find fixes in code which is a bit lax, and permitted by gcc but not by the Sun compiler. Then there can be one of two approaches. a) Apply a patch to the sources which allow it to work in Sage. b) Let the developers of the package known it does not compile with Sun's compiler and send them a fix. Probably the most useful thing to do is to build the latest gcc 4.4.0 (it can be built from the gcc in /usr/sfw/bin), download the latest sources, and report problems. Then fix those problems where you can. Here's a partial list of the packages installed in Sage. Getting them all to build with Sun's compiler would be great, but is not an easy task. atlas-3.8.3.p2.spkg atlas-3.8.3.p2.spkg.orig blas-20070724.spkg boehm_gc-7.1.p1.spkg boost-cropped-1.34.1.spkg cddlib-094f.spkg conway_polynomials-0.2.spkg cvxopt-0.9.p7.spkg cython-0.11.1.p0.spkg deps docutils-0.5.spkg dsage-1.0.1.spkg ecl-9.4.1.spkg eclib-20080310.p7.spkg ecm-6.2.1.p0.spkg elliptic_curves-0.1.spkg examples-4.0.1.alpha0.spkg extcode-4.0.1.alpha0.spkg f2c-20070816.p1.spkg flint-1.2.4.p2.spkg flint-1.2.4.p3.spkg flintqs-20070817.p4.spkg fortran-20071120.p5.spkg freetype-2.3.5.p0.spkg gap-4.4.10.p11.spkg gd-2.0.35.p1.spkg gdmodule-0.56.p5.spkg genus2reduction-0.3.p5.spkg gfan-0.3.p4.spkg ghmm-20080813.p0.spkg givaro-3.2.13rc2.spkg gnutls-2.2.1.p1.spkg graphs-20070722.spkg gsl-1.10.p1.spkg iml-1.0.1.p11.spkg ipython-0.9.1.spkg jinja-1.2.spkg jmol-11.6.16.p0.spkg jquery-1.2.6.p0.spkg jqueryui-1.6r807svn.p0.spkg jsmath-3.6b.p1.spkg lapack-20071123.p0.spkg lcalc-20080205.p2.spkg libdist_filelist libfplll-3.0.12.p0.spkg libgcrypt-1.4.3.p0.spkg libgpg_error-1.6.p0.spkg libm4ri-20090128.spkg libpng-1.2.35.spkg linbox-1.1.6.spkg matplotlib-0.98.5.3rc0-svn6910.p3.spkg maxima-5.16.3.p2.spkg mercurial-1.1.2.spkg moin-1.5.7.p2.spkg mpfi-1.3.4-cvs20071125.p7.spkg mpfr-2.4.1.spkg mpir-1.2.p0.spkg networkx-0.99.p1-fake_really-0.36.spkg newest_version notes.txt ntl-5.4.2.p7.spkg numeric-24.2.txt numpy-1.2.0.p0.spkg opencdk-0.6.6.spkg palp-1.1.p1.spkg pari-2.3.3.p0.spkg pexpect-2.0.p3.spkg polybori-0.5rc.p7.spkg polytopes_db-20080430.spkg pycrypto-2.0.1.p3.spkg pygments-0.11.1.spkg pynac-0.1.7.spkg pyprocessing-0.52.spkg python-2.5.4.p1.spkg python_gnutls-1.1.4.p3.spkg quaddouble-2.2.p9.spkg r-2.6.1.p22.spkg readline-5.2.p6.spkg rubiks-20070912.p8.spkg sage-4.0.1.alpha0.spkg sage_scripts-4.0.1.alpha0.spkg scipy-20071020-0.6.p4.spkg scipy_sandbox-20071020.p3.spkg scons-0.97.0d20071212.spkg setuptools-0.6c9.spkg singular-3-0-4-4-20090511.spkg sphinx-0.5.1.spkg sqlalchemy-0.4.6.p0.spkg sqlite-3.5.3.p3.spkg sqlite-3.5.3.p3.spkg.orig sqlite-3.5.3.p4.spkg symmetrica-2.0.p4.spkg sympow-1.018.1.p6.spkg sympy-0.6.4.spkg tachyon-0.98beta.p8.spkg termcap-1.3.1.p0.spkg tinyMCE-3.2.0.2.p0.spkg twisted-8.1.0.p2.spkg weave-0.4.9.spkg zlib-1.2.3.p4.spkg zn_poly-0.9.p0.spkg zodb3-3.7.0.p1.spkg --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
