Intel Compiler

Introduction

Intel compiler suite, primarily for Fortran, C, C++.

Web Page

Official web page

Supporting Information

Caveat for C++ compilation

In some instances when using the Intel C++ compiler ( icpc) support for preprocessing or libstdc++ from the GCC compiler set may be required. You are advised to load a version of GCC to support this, but note that the system default GCC does not include support for C++ and is not sufficient.

Caveat when compiling with MPI

The Intel MPI compiler set is recommended. However, using mpicc, mpic++, and mpif90 will , by default, use the GCC compilers. To ensure that the Intel compilers are used in support of Intel MPI, use mpiicc, mpiicpc and mpiifort respectively.

The MPI module you load should match the intel/compiler/version module you load, i.e. intel/mpi/version.

In some instances you may wish to use one of the alternative MPI versions, and specific commands exist with those to use a particular C, C++, or Fortran implementation with them. See the documentation for those alternative MPI implementations.

Using the Math Kernel Library

The Intel MKL is recommended to support optimised versions of BLAS, LAPACK, FFTW3, and so on. If you can make use of this, then you should load a module intel/mkl/version where version matches the compiler version you have loaded.

Using the Intel Threading Building Blocks

You may wish to make use of the Intel TBB when using threaded code.

Using Cilk Plus

Intel Cilk Plus is the easiest, quickest way to harness the power of both multicore and vector processing.

Intel Cilk Plus is an extension to the C and C++ languages to support data and task parallelism.

Using Intel Integrated Performance Primitives

Intel IPP offer developers high-quality, production-ready, low-level building blocks for image processing, signal processing, and data processing (data compression/decompression and cryptography) applications.

Using Intel Data Analytics Acceleration Library

Intel DAAL can be used to accelerate data analytics, post-processing, etc. The version used should match the compiler version.

Running compiled programs

For 2016.3.210 an intel/runtime/64/2016.3.210 package exists which wraps MKL, library and MPI support into one convenient module load for your scripts. For other versions the individual components used (including TBB, DAAL, etc.) should be loaded as required.

Analysing programs

A number of analysis tools are available:

A number of compiler diagnostics are also available.

Versions

2017.1.132

Flavours

32

Architecture

x86_64

Group
module load site-local-unsupported
Build Details

Installed by Cluster Vision.

Usage
module load unsupported intel/compiler/32/2017.1.132
Example Job Scripts

None required.

Status

Tested.

64

Architecture

x86_64

Group
module load cv-standard
Build Details

Installed by Cluster Vision.

Usage
module load intel/compiler/64/2017.1.132
Example Job Scripts

None required.

Status

Tested.

2018.2.199

Flavours

64

Architecture

x86_64

Group
module load site-local

2017.2.174

Flavours

32

Architecture

x86_64

Group
module load site-local-unsupported
Build Details

Installed by Cluster Vision.

Usage
module load unsupported intel/compiler/32/2017.2.174
Example Job Scripts

None required.

Status

Tested.

64

Architecture

x86_64

Group
module load cv-standard
Build Details

Installed by Cluster Vision.

Usage
module load intel/compiler/64/2017.2.174
Example Job Scripts

None required.

Status

Tested.

2016.3.210

Flavours

32

Architecture

x86_64

Group
module load site-local-unsupported
Build Details

Installed by Cluster Vision.

Usage
module load unsupported intel/compiler/32/2016.3.210
Example Job Scripts

None required.

Status

Tested.

64

Architecture

x86_64

Group
module load cv-standard
Build Details

Installed by Cluster Vision.

Usage
module load intel/compiler/64/2016.3.210
Example Job Scripts

None required.

Status

Tested.