[kishoremg040]

Hi All,

I am a newbie to Linux as well CFD Environments. I am working on to setup an Environment to run distributed parallel OpenFOAM usecase using Intel MPI over the INFINIBAND network. (https://docs.microsoft.com/en-us/azu...c/rdma-cluster).

I was able to compile OpenFOAM using GCC(4.8.5) compilers and OpenMPI.
I now have to compile using Intel ICC to make use of built in RDMA INTEL MPI drivers over INFINIBAND.

I have followed this blog(installation with Intel MPI and Intel compilers) to make necessary ICC related configuration changes. but still unable to succeed.

I see the Intel MPI binaries are built in with the VM image i have chosen.
How do i enable INTEL Compilers on my VM.
Do i need to manually install INTEL ICC compilers on the VMs?

Need help compiling OpenFoam latest/stable version with INTEL icc compilers on Openlogic CentOS 7.3(Linux 3.10.0-514.21.1.el7.x86_64).

Please suggest.

[gtarang]

You have to source the "compilervars.sh" to your .bashrc. In your $HOME/.bashrc
type
source /opt/intel/bin/compilervars.sh

to source intel mpi
type source <the address of the mpivars.sh shown in your locate command>

since mpirun for both openmpi and intelmpi is same, its better if you can setup "module load" or "alias" for intel mpi otherwise you won't be able to use open-mpi

don't worry about the different locations of the compilervars.sh and mpivars.sh. All the files are same and are just symbolic link to each other.