TMS320C6678: TI-RTOS timing benchmark for task's priority change with context switch

Niklas Winter

Part Number: TMS320C6678

I am on a TMS320c6678 device. I want to implement a synchronization algorithm which supports active waiting and waiting with sleeping. For this reason, I want to change the priority of tasks during runtime. When I want to put a task to sleep I set the task inactive by setting the priority to -1. When I want to wake the task up, I set the priority to a certain number. Since performance is a crucial part in our algorithm it is import to know some timing benchmarks for this particular use case. I know the ti-rtos manual mentions the timing benchmark "Lower the current task's own priority, context switch". Unfortunately, I was not able to find this particular benchmark among the provided benchmarks (https://software-dl.ti.com/dsps/dsps_public_sw/sdo_sb/targetcontent/bios/sysbios/6_83_00_18/exports/bios_6_83_00_18/packages/ti/sysbios/benchmarks/doc-files/TI_C66_ti_platforms_evm6678_time.html)

over 3 years ago

0 Shankari G over 3 years ago

TI__Mastermind 25535 points

Niklas,

Let me have a look at it and get back.

Regards

Shankari G

0 Shankari G over 3 years ago in reply to Shankari G

TI__Mastermind 25535 points

Niklas,

I understand you are asking the benchmark timing for the following case.

But unfortunately, it is not given, for this case, " Lower the current task's own priority, context switch".

It is given only for the previous case , " Set a Task Priority without a Context Switch " - 140 cycles...

Regards

Shankari G

0 Niklas Winter over 3 years ago in reply to Shankari G

Prodigy 130 points

Yes exactly, I am looking for this specific benchmark. Well, it would have been nice using the results of this particular benchmark as reference. But I guess I can also approximate the total cycles needed for the benchmark "Lower the current task's own priority, context switch" by adding the cycles needed for the benchmark "yield()" and the benchmark "lower the priority without context switch". My approximation should be then in the same range as the actually needed cycles, though, I need to look at the ti-rtos code to say for sure.

0 Shankari G over 3 years ago in reply to Niklas Winter

TI__Mastermind 25535 points

Niklas,

Logically yes, we can add them up.

But cannot be pretty sure.

A good try though.

Regards

Shankari G.

Processors

Processors forum

TMS320C6678: TI-RTOS timing benchmark for task's priority change with context switch