Honored Contributor
Honored Contributori still couldn't understand how the cpu's can share the same address space to be able to execute the same thread's set. i was thinking in copying the code (chosen by scheduler) to cpun memory area to be executed... but that can't solve my problem, because the kernel will not be accessible by others cpu's.
can anyone help me?
thank you!!!
<div align='right'><{post_snapback}> (index.php?act=findpost&pid=13418)
--- quote end ---
--- Quote End --- I have no experience with eCOS, however, IMHO, I don't think that implementing a scheduler with migration (that is, that moves tasks around the CPUs) will result in an efficient implementation on Nios II... Also consider that: - the peripherals seen by the various Nios II are in general different - there is no cache coherency in the Nios II hardware ... and that's basically why we chose for ERIKA a partitioning approach, with transparent handling of the multicore shared data. You can find some more info in the ERIKA Enterprise reference manual, in the "Multtiprocessor support" chapter, available for free at http://www.evidence.eu.com/nios2/literature.asp (http://www.evidence.eu.com/nios2/literature.asp) bye PJ
