+1 On Thu, Sep 24, 2015 at 5:28 PM, Stefan Karpinski <ste...@karpinski.org> wrote:
> This conversation is getting pretty tiresome. There are programs where > Matlab is already as fast as it's possible to be. If all you're doing is > computing a big matrix product, for example, then all any language is just > going to call BLAS. Julia is not going to be any faster than Matlab for > that, and neither, for that matter, is C or Fortran. There are plenty of > programs where this is not the case, however, and in Julia you can get as > fast as C without having to write C. > > On Thu, Sep 24, 2015 at 5:15 PM, Marcio Sales <marciole...@hotmail.com> > wrote: > >> That is a meaningless comparison. First, you are not comparing loops, you >>> are comparing matrix inversion. Second, neither Matlab nor Julia will >>> natively perform a matrix inversion well. They are both going to use an >>> external library (LAPACK) so what you are testing is the library, not the >>> language. For example, Matlab almost certainly ships with Intel's MKL (math >>> kernel libraries) which includes a call for matrix inversion that was >>> carefully written by Intel employees to run as fast as possible on Intel >>> CPUs. Julia cannot ship with MKL, but if you care enough it is probably >>> possible to link to it. >>> >>> Daniel. >>> >> >> Well, then did you just say that if one keeps using matlab in the way it >> was meant to (matrices operations), there's no way Julia can beat it >> currently (performancewise)? >> >> >> >
