For treating VTK data on Julia, Mohamed Tarek makes related libraries. I really appreciate his works. After Pkg.clone VTKDataTypes.jl & VTKDataIO.jl, you may have to resolve some dependencies.

For reading VTK, you can use the following:

using VTKDataTypes
using VTKDataIO

data = read_vtk("m000000.vts")

data. point_coords is coordinate data, and data.point_data is a dictionary data.

I start to use Julia. Fast & Useful linear algebra. It is a fresh alternative of Matlab. We can use Intel MKL without Intel Compiler. However, we have to bulid Julia as follows (Julia v0.6 & Ubuntu 16.04).

1. 1. Install Intel MKL
   2. Clone Julia repository of git
   3. Add Makefile.user in the main julia source directory with the following

      override USE_INTEL_MKL=1
      override USE_INTEL_LIBM = 0
      prefix=/home/hoge/julia

      (Note that the final line is install directory)

   4. Add the following to deps/tools/common.mk, for ARPACK.

      ifeq  ($(USE_INTEL_MKL), 1)
      ifneq ($(USEIFC),1)
      USE_BLAS_FFLAGS += -ff2c
      endif
      endif
      

      Refs:

      • https://github.com/JuliaLang/julia/issues/17946
      • https://software.intel.com/en-us/articles/mkl-single-dynamic-library-libmkl-rtso-does-not-conform-to-the-gfortran-calling-convention
   5. “make -j4” after “source /opt/intel/mkl/bin/mklvars.sh intel64 ilp64”
   6. We can check results by make testall
   7. I saw an error about the following part of cmdlineargs.jl, but it seems no problem.

      if !is_windows() && VersionNumber(Base.libllvm_version) > v"3.3"
      testdir = mktempdir()
      cd(testdir) do
      rm(testdir)
      @test success(`$exename -e "exit(0)"`)
      end
      end
      

   8. make install