DIE temperature reading cause CPU usage increase

RE: DIE temperature reading cause CPU usage increase

Naeem Maroof — Thu, 15 Aug 2024 18:47:16 GMT

here it is still running for about 9 hours.

RE: DIE temperature reading cause CPU usage increase

Naeem Maroof — Thu, 15 Aug 2024 10:17:42 GMT

Restarted (re-flashed) with the delay of 10ms (instead of 1000ms)

will continue for several hours.

RE: DIE temperature reading cause CPU usage increase

Naeem Maroof — Thu, 15 Aug 2024 10:12:19 GMT

I have run the test and will try it for several hours

RE: DIE temperature reading cause CPU usage increase

LongQi — Thu, 15 Aug 2024 10:00:51 GMT

Can you try to run longer time? It may happen anytime. it is not predictable.

RE: DIE temperature reading cause CPU usage increase

Naeem Maroof — Thu, 15 Aug 2024 09:58:50 GMT

Hi Zhang,

There is no update at my end.

I thought you might have tested as I have shown you.

RE: DIE temperature reading cause CPU usage increase

LongQi — Wed, 14 Aug 2024 02:02:13 GMT

is there any update?

RE: DIE temperature reading cause CPU usage increase

LongQi — Mon, 05 Aug 2024 05:33:09 GMT

Could you try longer time like 24 hours because the issue may happen anytime and I have not find any pattern about it.

RE: DIE temperature reading cause CPU usage increase

Naeem Maroof — Fri, 26 Jul 2024 15:58:40 GMT

Hi Zhang,

What do you mean by "any other testing"?

I have showed you that the code and hardware is working correctly.

I have also informed you that it could be the impact from the thread analyzer itself.

I do not have any more comments on it.

Regards,

Naeem

RE: DIE temperature reading cause CPU usage increase

LongQi — Wed, 24 Jul 2024 09:23:13 GMT

I understand there is no as such limit to reading the die-temperature as it is just like reading from the register.

I mean is there any other testing I can do to identify the root cause since you are saying no limit for the reading?

RE: DIE temperature reading cause CPU usage increase

Naeem Maroof — Wed, 24 Jul 2024 08:42:27 GMT

What I meant was that there is no as such limit to reading the die-temperature as it is just like reading from the register.

As you have described, I think then it should be something from the thread analyzer that is increasing the cpu usage.

I just ran the code a bit longer, and was getting same response at the expected time.

RE: DIE temperature reading cause CPU usage increase

LongQi — Wed, 24 Jul 2024 01:39:55 GMT

I am using thread analyzer to monitor the cpu and stack usage.

CONFIG_THREAD_NAME=y

CONFIG_THREAD_ANALYZER=y

CONFIG_THREAD_ANALYZER_AUTO=y

CONFIG_THREAD_ANALYZER_RUN_UNLOCKED=y

CONFIG_THREAD_ANALYZER_USE_PRINTK=y

it's not always 5737, sometimes it'll last very long time.

Are you referring to sensor_sample_fetch and sensor_channel_get for START task and DATARDY?

Is there any other testing I can do to identify the issue?

RE: DIE temperature reading cause CPU usage increase

Naeem Maroof — Tue, 23 Jul 2024 14:04:09 GMT

Hi Zhang,

I don't think there is any limit on die-temp reading, just that every temperature measurement initiated must start by triggering START task, and DATARDY event is generated to indicate that the TEMP register could be read.

So, requesting a measurement, and reading the register when ready does not have a limit on it, except that excessive operations might increase the die temperature.

I want to ask:

What kind of graph is this that you have shared?

How do you measure the cpu usage?

Does your thread always stop at counter 5737 if you reset?

I am not sure what is happening at your end, but I can successfully run the code, and I have done 15000+ readings by just speeding up (using lower time for sleep).