Polyhedron Fortran compilers benchmark

[Mikhail_K_1]

I'm sorry - I sent erroneously my message to Fortran/Windows forum, and repeat now. The Polyhedron Fortran benchmark is absent know,

and in particular is absent well known html:

http://www.polyhedron.com/compare0html

Do somebody have relative new (not too old) corresponding data ? I myself is sure that Intel Fortran benchmark results are the best, but I need this

data to prepare short information in our grant proposal.

Mikhail Kuzminsky

[TimP]

I'd be a little embarrassed to justify a selection on the basis of the (apparently withdrawn) Polyhedron benchmarks, as the story is much more complicated than that.

For example, if your project makes use of MKL libraries, or vector math functions, that alone might justify choice of Intel compiler.

gfortran has made great strides, with recent support for AVX2. In many important cases, it shows optimum performance.  However, it still shows a significant performance deficit for vectorizable dot_product (one of the few differences which might have showed up in Polyhedron, perhaps not in a useful form).  Intel's own comparisons tend not to use equivalent compile options nor use competing compilers from the same year of release.

ifort has good support for the new feature omp parallel do simd clause which is ignored by other popular compilers.  That is most likely to be useful on server CPUs with a large number of cores.

ifort may well be considered as having a superior offering of support vs. licensing fees, with the combination of these forum sites and the problem reporting site.

[Steven_L_Intel1]

The Polyhedron benchmarks can now be found at http://www.fortran.uk/compiler-comparisons/

What happened was that Polyhedron ceased operations. John Appleyard, a cofounder of Polyhedron and the maintainer of the benchmarks, is now with a new company "Fortran UK", and he tells me he intends to keep the benchmarks going. I have encouraged him to do so, as they are a great service to the Fortran community.