CUDACasts Episode #3
Write your first OpenACC program.
The video presents the basic steps and the typical process for accelerating code with OpenACC.
- Identify the computationally intensive sections of code you want to offload to the massively parallel GPU.
- Using OpenACC directives, move parallel loop execution to the GPU and verify it is functionally correct.
- Optimize any data movement between the host and device.
Read more about this on the Parallel Forall Blog.