> Hi all, > > Anybody know where to get packages for some common > scientific programs such as: > > R + packages > python + packages... numpy, scipy, pylab > fftw > some sort of tex distribution + packages > octave > > I had my first exposure to Solaris as an undergrad > working in a computational biology lab. I know it's > in pretty wide use academically. It would be nice if > scientific packages were available. Right now I'm > trying to use pkgsrc, but I think having a formal > scientific repository is a nicer solution. So I guess > this post is part question and part suggestion.
I've built R and octave on SPARC, and I'm pretty sure I've built octave on x86 too, but perhaps I am mistaken. There is expected to be a port of Sage http://www.sagemath.org/ soon, but it does not compile yet. That package is quite unique in that it includes lots of other packages, such as octave, R etc in it. I've got R in sage to build - again possibly only on SPARC. I've just checked, and these bits have built ok in Sage, so I doubt you would have too much trouble getting the sources of the respective packages to build. (Not all are scientific, but some are) atlas-3.8.1.p3 blas-20070724 bzip2-1.0.4 conway_polynomials-0.2 cython-0.9.8 dir-0.1 eclib-20080310.p5 elliptic_curves-0.1 extcode-3.1.alpha0 f2c-20070816.p0 flint-1.0.13 fortran-20071120.p5 freetype-2.3.5.p0 gd-2.0.33.p5 gdmodule-0.56.p4 givaro-3.2.12rc0 gmp-4.2.2.p1.fake gnutls-2.2.1.p1 graphs-20070722 gsl-1.10.p1 iml-1.0.1.p11 ipython-0.8.2.p0 lapack-20071123.p0 libfplll-2.1.6-20071129.p4 libgcrypt-1.4.0.p2 libgpg_error-1.6.p0 libpng-1.2.22.p7 linbox-1.1.6rc0.p1 matplotlib-0.91.1.p5 mercurial-1.01.p0 mpfi-1.3.4-cvs20071125.p7 mpfr-2.3.1 networkx-0.36.p1 ntl-5.4.2.p4 numpy-20080104-1.0.4.p5 opencdk-0.6.6 pari-2.3.3.p0 pexpect-2.0.p1 polybori-0.3.1.p4 prereq-0.3 pycrypto-2.0.1.p2 python-2.5.2.p3 python_gnutls-1.1.4.p3 quaddouble-2.2.p9 readline-5.2.p3 sage_scripts-3.1.alpha0 scons-0.97.0d20071212 singular-3-0-4-2-20080611.p2 sqlite-3.5.3.p3 symmetrica-2.0.p2 sympy-0.6.0 termcap-1.3.1.p0 twisted-8.0.1.p1 zlib-1.2.3.p3 zodb3-3.7.0.p1 I've had Mathematica running on Solaris x86 on an Intel chip, but is not supported, but I managed to hack it by changing some libraries. Depending on what licenses you have, you might want to try that. If you have a Mathematica site license, they might give you a password for x86. I did this years ago - although officially Solaris (SPARC) was not considered a home computer, therefore our license did not cover it, Wolfram made an exception. But unless you are already using Mathematica, I would keep clear of a closed source program like that. Or you end up in the position I'm in, where I've used MMA a lot, but feel I want to use an open-source one due to financial reasons. One of my own scientific programs http://atlc.sourceforge.net/ has not been tested on Solaris x86, but given it runs on Linux, Solaris, HP-UX, tru64, Unicos, FreeBSD, NetBSD, Irix, Unixware, SCO ... I'd be very surprised if it did not build on Indiana. It's been run on a Sony Playstation and a Cray supercomputer in the past! It would be nice to see a web page listing scientific programs running ok on Solaris, and what (if any) changes/hacks need to be done to get them running. I personally think Sage will be the one to have, since that has all the functionality of its constituent parts, which is one hell of a lot of functionality. This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
