Re: Matrix Multiplication



On Jan 5, 1:44 am, user923005 <dcor...@xxxxxxxxx> wrote:
On Jan 4, 1:24 pm, Evgenii Rudnyi <use...@xxxxxxxxx> wrote:

I would suggest to switch from words to actions. I do not know newer
Fortran and I am not sure if I want to learn it. Could you please give
me a program in newer Fortran that multiplies two matrices and with
that you are satisfied? Then I compile it with Intel compiler and
compare with my code. Is this sounds reasonable?

I guess that if you make a request to news:comp.lang.fortran for them
to provide a modern and fast F95 Fortran matrix multiply someone will
do it for you. Atlas will still clobber it, of course.

Thanks, probably at some point I will do. If would be interesting to
see it indeed. I just wonder why Gordon prefers to keep silent when it
comes to the code.
.



Relevant Pages

  • Re: Matrix Multiplication
    ... Fortran and I am not sure if I want to learn it. ... me a program in newer Fortran that multiplies two matrices and with ... A Fortran 90 compiler is free to use Atlas, ... A carefully written code like Atlas would be an obvious candidate ...
    (sci.math.num-analysis)
  • Re: Matrix Multiplication
    ... Fortran and I am not sure if I want to learn it. ... me a program in newer Fortran that multiplies two matrices and with ... Atlas will still clobber it, ... early compiler implementations would not have but by now the run ...
    (sci.math.num-analysis)
  • Re: Matrix Multiplication
    ... Fortran and I am not sure if I want to learn it. ... me a program in newer Fortran that multiplies two matrices and with ... Then I compile it with Intel compiler and ... Atlas will still clobber it, ...
    (sci.math.num-analysis)
  • Re: why is Fortran A*B (for matrices) seem to be slow?
    ... on windows 7 with gfortran 4.6 for fortran ... system_clock for matmul: 43.5 ... On the same platform with Matlab R2009B and your original script I get an typical second or subsequent run tic-toc time of 11.9 seconds. ... I was half expecting the Intel compiler to pick that the variable "c" wasn't being used, and so remove the MATMULT and GEMM calls. ...
    (comp.lang.fortran)
  • Re: what is wrong with this code ?
    ... but the principle applies pretty generally to most lists. ... It's actually the now-defunct predecessor of the Intel compiler ... Standard violation in not diagnosing the nonstandard usage. ... The inner is bad Fortran ...
    (comp.lang.fortran)