Hello community,

I will admit that I am still new to using Linux (Fedora), and am having trouble finding a way to link my MKL library to Eigen. My task is to run a benchmark between Armadillo and Eigen through gcc, based on basic, acml, and then mkl. I have everything working on the Armadillo side, which was easy because all it required was a few changes to the cmake files, and a make install to my directory (eg. arma_withacml_2.4.4 or something similar). When I was looking at the Eigen 3.0.5 tar file, I noticed there were no make or configure files.

I thought I'd try compiling while linked like this:
gcc test_eigenAcml.cpp -o test_eigenAcml -I/opt/eigen3.0.5/include -L{path to acml library} -{designated libs}

or even just set up a module, but I was a bit lost on how this would help, considering the Eigen directory was still the same as my previous benchmark for basic compilations.

Maybe I am just overthiking this, but any advice would be greatly appreciated!

Thank you!
Ryan

ok with a little more research, I did redownload a new version of eigen (eigen-eigen-6e7488e20373.tar.gz) that had the ability to aloow me to do a cmake to a build directory. and I found this link, which looks promising:
http://eigen.tuxfamily.org/dox-devel/To ... elMKL.html
now I just have to find the EIGEN_USE_MKL macro, and make sure it is defined.

Since I haven't found this yet, I'm unsure if I will be able to use ACML, because I am seeing posts about the need of a patch.

Will i need the patch? or will there be a way to work around this?

Thank you!
Ryan

Eigen is a header-only library, for MKL you need to get the devel branch, and compile your test with -DEIGEN_USE_MKL and all the appropriate linker flags.

Thanks,

I managed to get it to compile now, but without any noticable difference in time (.002s~.0004s difference). I'm going to spend more time playing with this, to see if my Wl linker might be failing.

matB = matA*A^T, for matrix size = 500x500, is the function im testing. And was anticipating eigen with MKL would have much better performance

only if you have a CPU with AVX instruction sets, otherwise Eigen is nearly as good as MKL:
http://eigen.tuxfamily.org/index.php?title=Benchmark

Ahhh that makes sense, Thank you!

Just a heads up to anyone who was having same problem as me:

Using Eigen3.1.0 and MKL 10.3.7, I had to create two modules for proper compilation. The first one was an Intel_MKL module, designating the proper libraries and paths. The second was the Eigen module, which is doing the same thing as the Intel_MKL one, only for Eigen.

Example:
-----intel------
#%Module 1.0
2 #
3 # Intel module for use with 'environment-modules' package:
4 #
5
6 ## MKL
7 setenv MKLROOT /opt/mkl10.3.7/mkl
8 setenv MKL_TARGET_ARCH intel64
9 prepend-path LD_LIBRARY_PATH /opt/mkl10.3.7/mkl/lib/intel64
10 prepend-path LIBRARY_PATH /opt/mkl10.3.7/mkl/lib/intel64
11 prepend-path LD_LIBRARY_PATH /opt/mkl10.3.7/lib/intel64
12 prepend-path LIBRARY_PATH /opt/mkl10.3.7/lib/intel64
13 prepend-path NLSPATH /opt/mkl10.3.7/mkl/lib/intel64/locale/%l_%t/%N
14 prepend-path MANPATH /opt/mkl10.3.7/man/en_US
15 prepend-path INCLUDE /opt/mkl10.3.7/mkl/include
16 prepend-path CPATH /opt/mkl10.3.7/mkl/include

-----Eigen-----
1 #%Module 3.2
2 #
3 # Eigen with Intel MKL Blas and Lapack module
4 #
5 prereq intel-x86_64
6 prepend-path INCLUDE /opt/eigen3.1.0_MKL/include/eigen3
7 prepend-path CPATH /opt/eigen3.1.0_MKL/include/eigen3
8 prepend-path LD_LIBRARY_PATH /opt/eigen3.1.0_MKL/lib
9 prepend-path LIBRARY_PATH /opt/eigen3.1.0_MKL/lib
10 setenv EIGEN_ROOT /opt/eigen3.1.0_MKL/
11 setenv EIGEN_INFO eigen3.1.0_MKL

----------------
with these two modules loaded, in order 1.)intel 2.)eigen, you should be able to easily compile your Eigen code using MKL....
eg. g++ code.cpp -o outputfile

Sometimes you will need to add the flag -DEIGEN_USE_MKL_ALL in compilation to turn on the MKL features.

I know this may be common sense for a lot of users on here,
But for the newbies like me this could be extremely helpful.

sounds useful, I copied it in the wiki:
http://eigen.tuxfamily.org/index.php?title=MKL_Tips