Archive for April, 2007

Educause 2007 – eResearch

Attended my one and only day of educause today. Random thoughts and impressions follow:

Middleware, and the importance of it in enabling researchers who don’t have the skills to “build”, to be involved.

Emergence of electronic publishing as the ‘primary copy’ – ability to link to raw data etc. So many advantagesw over print.

Data storage – copies on CD, copies on desktop are neither sustainable nor maintainable. Central repositories (or institutional repositories).

IBM and HP computing infrastructure – what’s the name of the new processor with it’s set of ‘slave’ processors?

Funding regime in australia – NCRIS etc.

DART, ARCHER… attempts at producing eResearch middleware.

Scientific workflows, it’s all about data.

Expertise shortage, and coupled lack of career paths, formal positions for workers who straddle the ICT and science domain areas.

Data standards (data curation), annotation. Data publishing (to make public) seperate from journal publishing.

Interoperability of new MS office formats – xml, easy conversion between other open formats.

Advertisements

Leave a Comment

format statements – g95/ifort

ifort is quite happy with
write(fp,'(15(F))')
g95 is somewhat more demanding, and requires the length of the output
write(fp,'(15(F10.5))')
Just one of those things.

Comments (1)

Sph phase transitions – the state of play

Nugent and Posch’s generation of condensed drops, and two phase mixtures as far as I know hasn’t been topped, saved for Melean and Sigalotti, who used various schemes (and artificial viscosity, and ‘adaptive kernel estimation’) to improve the homogeneity and reduce the size of the interface respectively.

It turns out that simple von Neumann Richtmyer artificial viscosity doesn’t prevent the ringed structures in low temperature droplets.

I’m tempted to skip straight past the other methods to Hoover’s ‘density gradient’ potentials.

Leave a Comment

MPICH on ubuntu

Installation was relatively straightforward. Don’t unzip the source into the directory you want to install into. sudo the configure, make, and make install commands. Make sure the paths are set up for all compilers you want mpich to pick up. Add the /installdir/bin directory to the path in the system’s environment file (or however else you want to add it).

Comments (8)

Intel fortran compiler on Ubuntu

I reinstalled dash. The promise of faster startup time was too good. I’m still a little miffed at the dev’s decision to make my life harder, but I might as well make the best of it.

To make sure ifort will still work, you need to make sure the files ifort, and ifortvars.sh use #!/bin/bash, otherwise you get a cryptic error message when you call ifort.

Comments (5)

Ubuntu environment variables

The easiest way to set an environment variable in Ubuntu is to set it in the /etc/environment file.

Setting it in your bash_profile or bashrc file doesn’t work for me. There’s a Debian historical reason behind it, and no doubt more details. But if you want a quick way to set your LD_LIBRARY_PATH for the intel fortran compiler, that’s it.

My environment file looks like this:


PATH="~/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:
/usr/bin/X11:/usr/games:/opt/intel/fc/9.1.036/bin:
/usr/local/mpich-1.2.7p1/bin:PATH"
LD_LIBRARY_PATH="/opt/intel/fc/9.1.036/li"
LANG="en_AU.UTF-8"

note: this is an old post and I don’t find myself using Ubuntu a lot now, especially as an computer administrator. If it turns out this is wrong or out of date let me know.

Leave a Comment

Workflow

Papers to read, literature reviews to write, code to tweak, simulations to run, results to analyse, graphs to graph…

I’m experiencing a min-crisis of workflow at the moment, one of those moments where you start making spreadsheets to track progress, writing lists, and vainly attempting to tidy up cluttered file systems.

I find that when I’m driving towards a particular development goal, I get into a zone – I imagine if I was a sculptor, my studio would end up with tools everywhere, chips all over floor, and a shiny finished sculpture in the middle of the pile of rubbish.

So how do I resolve this? Do I get caught up in meticulously planning and documenting a workflow and a file management system (knowing full well that it will go out the window when I’m on fire). Do I just discard the old workspace and start with a clean one?

Leave a Comment

Older Posts »