Re:Modern Python Performance Considerations
by Robotbeat on Thursday May 05, 2022 @04:25PM
(#31278008)
Fortran >:D
Re:Modern Python Performance Considerations
by whimsicalism on Thursday May 05, 2022 @04:28PM
(#31278033)
For low-level, fair. I only know of people in astronomy academia who actually use it nowadays though.
There are good technical reasons to use Fortran even in 2022. For example, it avoids** is aliasing (pointers / references / etc.). This allows for some kinds of optimizations impossible in most other languages.
It's used in a bunch of small niches, but it has users beyond just astronomy.
** A million disclaimers apply.
Re:Modern Python Performance Considerations
by whimsicalism on Friday May 06, 2022 @10:37AM
(#31285645)
I am aware that hypothetically fortran code is the fastest possible. That said, I am not sure how great this aliasing difference is in practice.
There is a reason why most BLAS implementations have been rewritten into C.
Re:Modern Python Performance Considerations
by BenjiWiebe on Thursday May 05, 2022 @08:22PM
(#31279753)
Signal processing stuff is still sometimes written in it.
More, given that no language competes at high-level numerics with Python outside of Julia and numerics in general only adds C++.