Advice and background information is given on typical issues that may arise when threading an application using the Intel Fortran Compiler and other software tools, whether using OpenMP, automatic parallelization or threaded libraries.
This article is part of the Intel® Modern Code Developer Community documentation which supports developers in leveraging application performance in code through a systematic step-by-step optimization framework methodology. This article addresses: Thread level parallelization.
Are the Intel Fortran run-time libraries thread safe?
Vectorization essentials to effectively use feature in the Intel® Xeon product family
No longer does Moore’s Law result in higher frequencies and improved scalar application performance; instead, higher transistor counts lead to increased parallelism, both through more cores and thr
SIGSEGV on Linux and SIGBUS on MacOS Root Causes
When confronted with nested loops, the granularity of the computations that are assigned to threads will directly affect performance. Loop transformations such as splitting and merging nested loops can make parallelization easier and more productive.
Vectorization is one of many optimizations that are enabled by default in the latest Intel compilers. In order to be vectorized, loops must obey certain conditions, listed below. Some additional ways to help the compiler to vectorize loops are described.
One key to attaining good parallel performance is choosing the right granularity for the application. Granularity is the amount of real work in the parallel task. If granularity is too fine, then performance can suffer from communication overhead.