Hi
I asked some questions in " keystone multicore forum " and they said it is better that I asked from HPC forum, My thread is as below:
Please answer all of my questions from your point of view, again.
Best Regards
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Hi
I asked some questions in " keystone multicore forum " and they said it is better that I asked from HPC forum, My thread is as below:
Please answer all of my questions from your point of view, again.
Best Regards
Question 3:
Can I run omp program on DSP cores via ARM core in " ARM+DSP " device?
I can try to answer your questions as they relate to offloading computation from the ARMs to the DSPs.
In regards to OpenMP, you can use the the OpenMP4 accelerator model or OpenCL to offload code/data from the ARMs to the DSPs. See this link:
Best Regards, Eric
Hi Eric
Please, answer all questions. I think it is better that I explained my target behind my questions.
A:
I worked parallel programming on desktop processor.
IDE: Visual Studio
Compiler: Micro Soft / Parallel Studio(intel)
OS: Win 7
Processor: Core i5 2410
When I worked with desktop processor , The parallel programming was very easy.
B:
I started to work with DSP.
IDE: CCS V5.2
Compiler: GCC
OS: Win 7
Processor: C6678
I faced a lot of new problems that I didn't have them when I worked with desktop processor.
1) RTOS didn't support OMP in a thread.
2) Cache coherency.
3) Managing Communication between CorePacs and other peripherals.
Finally I understand that OMP isn't suitable for parallel programming on DSP and it is better that used EDMA3 for communication and made parallel programming. The model is shown as below:
Now:
I search and study about " ARM+DSP" DSP like 66Ak2h. I want to know if I used OMP4 on ARM processors all the problems on the section " B " are solved?
By another words, Does OS Linux manage all problems especially in HPC mode(like when there is a big matrix(20MB) for processing) by using OMP4?
Best Regards