Overview

Here you will find information on building a variety of software packages I find useful in my work on linux (both 64 and 32 bit systems) and OS X. This post serves as a useful entry point to the site for those who found there way here. The “Platforms” tab above has information on how I set up different platforms before installing the various packages.

. . . and on RHEL 5

So I now have to try and get my setup working on a RHEL 5 machine with 16 GB of ram and two dual core processors.  There are a few rather annoying problems that this brings up that were not present in my Mac setup.  First and foremost is that there are no universal binaries so if you compile one things for a 64 Bit machine you must compile everything that depends on that with 64 bit support.  This problem is exacerbated by the fact that ROOT does not respect its arch flags throughout its make processes.  That is you cannot cross compile ROOT so you need to trick it into thinking you have a 32 bit system.  Really annoying no?

So we had an installation setup with both 32 bit and 64 bit binaries which both apparently worked (although I never tested this).  Unfortunately a over zealous user or two managed to destroy this setup and it was replaced with a much more confusing setup where everything was in the same directory and some things were 32 bit and some 64 bit.   

The new game plan is that I will compile all the stuff I need in 32 bit first and get that all working in a new 32 bit specific location then compile what I can in 64 bit versions and put that in yet another directory.  Hopefully this will all get done in a day.

Program Notes

Here are some cursory notes on installing each program:

1. python 2.5 – Built into most systems these days including my MBP
2. gsl - Pretty easy to install just using configure and make.
3. fftw - Pretty easy to install but you do need a fortran compiler.
4. cfitsio -  Pretty easy to install just using configure and make. 
5. boost – Pretty easy to install just using configure and make.
6. ROOT – Takes forever but is pretty easy.
7. pyfits – VERY easy to install using the standard python setup.py methods.
8. coords - VERY easy to install using the standard python setup.py methods.
9. ipython - VERY easy to install using the standard python setup.py methods. 
10. numpy - VERY easy to install using the standard python setup.py methods.
11. scipy- Can be temperamental.  If you just use this it should go smoothly.
12. matplotlib – Easy to install but check and make sure it knows where all your libraries are.
13. HippoDraw – Very temperamental.  Keeps giving random errors unless you start from the beginning every time you find a problem.

Programs to be Used

Here is a list of programs for analysis I have installed and the order I think I installed them in:

1. python 2.5
2. pyfits
3. boost
4. cfitsio
5. coords
6. fftw
7. gsl
8. ipython
9. numpy
10. scipy
11. ROOT
12. matplotlib
13. Hippodraw

First post

So this is the first post.

I am trying to get Numpy, Scipy, and matplotlib all to work smoothly again after stupidly trying to update numpy.