threading

[Paul Rubin]

Sturla Molden <sturla.molden at gmail.com> writes:
> When should we use C++ or Fortran instead of Python? Ever?

When performance matters? 

> There is a reason scientists are running Python on even the biggest
> supercomputers today.

They use Python as a scripting wrapper around numerics libraries written
in compiled languages.

> Hardware might be expensive, but not compared to human resources.

Sure, and that implies that making people wait for computer results
costs you.  If a C++ program can run in 2 seconds while the equivalent
Python script would take a minute, the Python version loses the user's
attention and they have to burn time refocusing it.  Or if the C++
version takes 12 hours and the Python version takes a month, they can
start the C++ version at the end of their workday and have the results
when they arrive at the office the next morning.  Then they spend the
day examining the results and adjusting the program for the next run.

Most of my stuff is in Python but there are times when your stuff just
has to run fast.