You may be aware of this, but my understandig as to why Fortran is column major for arrays is that this lends itself to more efficient compuation of many matrix operations (matrix multiplication). Most other scientific-centric languages are the same. What is funny is that for HPC Fortran and C++ are still the go to languages, yet are opposite in this column vs row major attribute. It gets messier when you start diving into CUDA , BLAS, etc etc, but in the end I think it gets down to what sort of number crunching the language was primarily intended to do (c /c ++ vs Fortran). Indeed, there is more than one way to skin a cat!
That would not be surprising at all. That's the whole point of FORTRAN, it is a language for highly efficient computations. Knowing how it works at the assembly/machine-code level, I'm not sure how that would be beneficial, but I wouldn't be surprised. Thanks for your comments - always good to hear these types of ideas.
You may be aware of this, but my understandig as to why Fortran is column major for arrays is that this lends itself to more efficient compuation of many matrix operations (matrix multiplication). Most other scientific-centric languages are the same. What is funny is that for HPC Fortran and C++ are still the go to languages, yet are opposite in this column vs row major attribute. It gets messier when you start diving into CUDA , BLAS, etc etc, but in the end I think it gets down to what sort of number crunching the language was primarily intended to do (c /c ++ vs Fortran). Indeed, there is more than one way to skin a cat!
That would not be surprising at all. That's the whole point of FORTRAN, it is a language for highly efficient computations. Knowing how it works at the assembly/machine-code level, I'm not sure how that would be beneficial, but I wouldn't be surprised. Thanks for your comments - always good to hear these types of ideas.