Hallo,
I am student and compare different RTOS for a Cortex-M3 (LPC1768 from NXP) and Keil compiler. At the moment I'm checking the Segger embOS.
To measure the context-switch-time I used the scope example comes with the evalversion of the embOS. My results are 4,66 µs with 72 MHz Core-Clock and 3,2 µs with 100 MHz Core-Clock. If I calculate correctly, they are round 320 CPU-Cycles. Is this a realistic time? What was the Core-Clock by your test for Cortex-M3 shown here?
Furthermore, I want every task placed in an own C-File. But it doesn’t work. I have made a header with the prototype of every task (like I do for any C-function). But if I link the project the Keil linker reports: "Undefined symbol "Taskname" (referred from main.o)" and that for every task. In the main.c file I use just the OS_CREATETASK() with the "symbol" of the Taskname. If I put all tasks together with the main function in one C-File it works fine.
What's my mistake? Do you have an example with the tasks in different files?
Greetings
Martin
I am student and compare different RTOS for a Cortex-M3 (LPC1768 from NXP) and Keil compiler. At the moment I'm checking the Segger embOS.
To measure the context-switch-time I used the scope example comes with the evalversion of the embOS. My results are 4,66 µs with 72 MHz Core-Clock and 3,2 µs with 100 MHz Core-Clock. If I calculate correctly, they are round 320 CPU-Cycles. Is this a realistic time? What was the Core-Clock by your test for Cortex-M3 shown here?
Furthermore, I want every task placed in an own C-File. But it doesn’t work. I have made a header with the prototype of every task (like I do for any C-function). But if I link the project the Keil linker reports: "Undefined symbol "Taskname" (referred from main.o)" and that for every task. In the main.c file I use just the OS_CREATETASK() with the "symbol" of the Taskname. If I put all tasks together with the main function in one C-File it works fine.
What's my mistake? Do you have an example with the tasks in different files?
Greetings
Martin