Today, I gave a talk at CWI on how to become more efficient with complex computations in our scientific work. I discussed how I have approached the need for distributed computation (to scale up towards larger and more complex problems) and the problem of organising the scientific workflow when doing experimental work.
I promised listeners that they would get
- hands-on information on getting results from large-scale, "embarrassingly parallel" computations,
- ... without actual parallel programming,
- ... little ssh effort
- ... and using the programming language of your choice.
- Plus, some tools to keep track of experiments and data.
Here are the slides (direct link to PDF):