Introduction to OpenMP/OpenACC
OpenACC is a new open parallel programming standard for GPUs with simple directives — similar to the OpenMP standard for programming the CPUs for shared memory computer architectures.
By exposing parallelism to the compiler, these programmer-inserted directives assist the compiler to map computation onto the GPUs without code modifications. The use of directives greatly reduces the effort required for the parallelization of new as well as existing codes for GPU applications.
Because OpenMP and OpenACC share many of the same directive concepts and, because OpenMP and OpenACC may be used together in the same code for multiple CPU and GPU applications, we will cover both the introduction to OpenMP as well as OpenACC.
The OpenACC API supports C/C++ and Fortran. This tutorial covers some of its features along with examples in C and Fortran and hands-on exercises.
Prerequisites: This course is for those who are new to parallel programming with OpenACC. Experience in C or Fortran programming is required. Understanding of parallel programming is helpful. Knowledge of OpenMP is not required.
If you try to sign up for this tutorial and find that it is full, please send email to email@example.com and we will do our best to accommodate you.