rymut/matlab-mex-intel18

XML configuration files for compiling MATLAB MEX-files using Intel Parallel Studio XE 2018 (Beta).

Matlab XML configuration for Intel Parallel Studio XE 2018 compiler
XML configuration files for compiling MATLAB MEX-files using Intel Parallel Studio XE 2018 (Beta).
Tested on Windows 10 64-bit with MATLAB R2017 and Microsoft Visual Studio 2017 Professional with Intel Parallel Studio XE 2018 (Beta).

Configuration should also work fine with final release of Intel Parallel Studio XE 2018.

Install

First copy the files (intel_c_18_vs2017.xml, intel_cpp_18_vs2017.xml, intel_fortran_18_vs2017.xml) to MATLABROOT\bin\win64\mexopts (operation might require administrator privileges).

Configure

Next run mex -setup C, mex -setup C++, and mex -setup FORTRAN in MATLAB.
Select 'Intel Parallel Studio XE 2017 with Microsoft Visual Studio 2017 (C)' as C compiler, 'Intel Parallel Studio XE 2018 for C++ with Microsoft Visual Studio 2017' as C++ compiler, and 'Intel Parallel Studio XE 2018 for Fortran with Microsoft Visual Studio 2017' as FORTRAN compiler.

eval(['mex -setup:''' ...
matlabroot '\bin\win64\mexopts\intel_c_18_vs2017.xml'' C'])
eval(['mex -setup:''' ...
matlabroot '\bin\win64\mexopts\intel_cpp_18_vs2017.xml'' C++'])
eval(['mex -setup:''' ...
matlabroot '\bin\win64\mexopts\intel_fortran_18_vs2017.xml'' FORTRAN'])
Check if everything is working

Test the new settings with a sample MEX-file included with MATLAB:

% copy example file yprime.c
copyfile(fullfile(matlabroot,'extern','examples','mex','yprime.c'),'.','f')
copyfile(fullfile(matlabroot,'extern','examples','mex','yprimef.f'),'.','f')
copyfile(fullfile(matlabroot,'extern','examples','mex','yprimefg.f'),'.','f')
% make copy yprime.cpp of yprime.c file
copyfile('yprime.c', 'yprimecpp.cpp')

% compile command for C file using Intel C Compiler
mex -n -largeArrayDims yprime.c
% compile file
mex -v -largeArrayDims yprime.c
% test function correct ans is [2.0000, 8.9685, 4.0000, -1.0947]
T=1; Y=1:4; ansc = yprime(T,Y)

% clear mex
clear mex;

% compile command for CPP file using Intel C++ Compiler
mex -n -largeArrayDims yprimecpp.cpp
% compile file
mex -v -largeArrayDims yprimecpp.cpp
% test function correct ans is [2.0000, 8.9685, 4.0000, -1.0947]
T=1; Y=1:4; anscpp = yprimecpp(T,Y)

% compile command for CPP file using Intel FORTRAN Compiler
mex -n -largeArrayDims yprimef.f yprimefg.F
% compile file
mex -v -largeArrayDims yprimef.f yprimefg.F
% test function correct ans is [2.0000, 8.9685, 4.0000, -1.0947]
T=1; Y=1:4; ansfortran = yprimef(T,Y)