Indeed. Physicists are the same as engineers.
All I care about is 1) code that is readable
enough for most other physicists and 2) wall-clock
time. FORTRAN wins in virtually every instance
I've come across.
That said, there is an emotional response among
physicists using C or C++ for scientific computing. For jobs that aren't so
computationally intensive, C++ can be great.
Many physicists have started using C and C++
extensively, but it's not what I would call
pervasive.
But, if you're doing high performance computing
and need to efficiently use every clock
cycle of your IBM SP, you've no real choice
except FORTRAN. The compilers are just plain
better.
Slashdot Mirror
Indeed. Physicists are the same as engineers. All I care about is 1) code that is readable enough for most other physicists and 2) wall-clock time. FORTRAN wins in virtually every instance I've come across. That said, there is an emotional response among physicists using C or C++ for scientific computing. For jobs that aren't so computationally intensive, C++ can be great. Many physicists have started using C and C++ extensively, but it's not what I would call pervasive.
But, if you're doing high performance computing and need to efficiently use every clock cycle of your IBM SP, you've no real choice except FORTRAN. The compilers are just plain better.