Hello all,
I want to use IPC example from
ipc_1_24_03_32/packages/ti/sdo/ipc/examples/multicore/evm667x/notify_multicore.c
for sending notifications from one core to another with the
purpose to organize parallel multicore processes.
For this I measured the time which takes notification
Core0->Core1->...->Core6->Core7->Core0
I have got unexpected huge time about 37700 [cycles].
Then I have measured time which takes "shortest" notification
Core0->Core1->Core0
I have got again unexpected huge time about 8600 [cycles].
The simple exchange of notifications over shared memory and without using of IPC
(which uses harware resources as I think)
gives the time about tens of cycles and this is acceptable for many cases.
Please suggest me how to use IPC example to get notification time about tens of cycles?
I used:
Code Composer Studio 5.2.1.00018
mcsdk_2_01_02_06
MDSEVM6678L/MDXEVMPCI boards
Source codes of the modified example is in the attachment.
Best regards.