« Previous 1 2 3 4
OpenACC Directives for Data Movement
Summary
The powerful OpenACC data directive allows you to control the movement of data in an application, with the purpose of improving performance. In this article, I looked at five data directive “clauses” – copy, copyin, copyout, create, and present – that give you great control over moving data between the host and the accelerator and give the compiler some hints as to what is going on in the code so it can generate better data movement (and, therefore, better performance).
With the two OpenACC directives discussed so far – parallel loop and data (along with its clauses) – you have the basics for getting started writing code using OpenACC.
« Previous 1 2 3 4
Related content
-
Compiler Directives for Parallel Processing
With directive coding, you annotate code with compiler directives to take advantage of parallelism or accelerators. The two primary standards are OpenACC and OpenMP.
-
OpenACC – Porting Code
OpenACC directives can improve performance if you know how to find where parallel code will make the greatest difference.
-
OpenMP – Coding Habits and GPUs
In this third and last article on OpenMP, we look at good OpenMP coding habits and present a short introduction to employing OpenMP with GPUs.
-
Using loop directives to improve performance
OpenACC is a great tool for parallelizing applications for a variety of processors. In this article, I look at one of the most powerful directives, parallel loop. -
OpenACC – Parallelizing Loops
OpenACC is a great tool for parallelizing applications for a variety of processors. In t his article , I look at one of the most powerful directives, loop .