Hi
At first I should tell a breif history about my project:
I would like to implement an algorithm by parallel programming on C6678 I should mentioned that I implemented my algorithm on desktop processor (intel core i7 3930) by Openmp in Visual Studio before.
So, I used "image_proc_omp" as my base project and implemented as below:
1) I made three threads for receiving packets, parsing packets and store data in DDR3 and processing the data.
2) I synchronized the treads by semaphore.
All things worked good until I would like to use OpenMP in my processing thread. OpenMP didn't work because SYS/BIOS didn't support OpenMP.
Now, I find LINUXMCDSK for C66x.
Question 1:
If I use LINUXMCDSK can I use OpenMP in a thread? Does linux handle my problem?
Question 2:
(I want to load my program on core_0 and processing thread go parallel. What is your opinion about my architecture?
Question 2:
Does linux handle cache problems like cache coherency...?
Best Regards