Timestamp TULIP Virtual Port Feeds

[rocket.scientist]

Can each line in the various serial feeds be timestamped somehow - console, tracer and HP-il scope? I am doing a lot of correlation between different feeds right now. It is hard to match up where a HPIL transaction happens in the tracer feed. Maybe make this an option that can be turned on / off?

If this is not the best spot for a feature request, please redirect me.

Thanks for the consideration.

[MeindertKuipers]

Best is to add an issue to the TULIP DevBoard GitHub page. But I can do that if you do not have an account there.

In the console I do not think a timestamp adds any value. For the traces it certainly does. I can add the samplecount (first column in the tracer) to the HP-IL scope, so there is a easy cross reference. There will be some jitter in the counter (HP-IL scope data is generated in the non-critical core) but I estimate this is not too much.
A clock will be a bit of a hassle, and there is not battery backed RTC on the device. A disadvantage of the sample counter is that it resets after every power event.

[rocket.scientist]

(10-25-2024 07:57 AM)MeindertKuipers Wrote:  Best is to add an issue to the TULIP DevBoard GitHub page. But I can do that if you do not have an account there.

In the console I do not think a timestamp adds any value. For the traces it certainly does. I can add the samplecount (first column in the tracer) to the HP-IL scope, so there is a easy cross reference. There will be some jitter in the counter (HP-IL scope data is generated in the non-critical core) but I estimate this is not too much.
A clock will be a bit of a hassle, and there is not battery backed RTC on the device. A disadvantage of the sample counter is that it resets after every power event.

Using the samplecount will help immensely, thank you. Right now I am using the cdcident command to bookmark the feeds, and that is w-a-y too granular.

I will add an issue in the github.

-Pat

[MeindertKuipers]

(10-25-2024 12:20 PM)rocket.scientist Wrote:  

(10-25-2024 07:57 AM)MeindertKuipers Wrote:  Best is to add an issue to the TULIP DevBoard GitHub page. But I can do that if you do not have an account there.

In the console I do not think a timestamp adds any value. For the traces it certainly does. I can add the samplecount (first column in the tracer) to the HP-IL scope, so there is a easy cross reference. There will be some jitter in the counter (HP-IL scope data is generated in the non-critical core) but I estimate this is not too much.
A clock will be a bit of a hassle, and there is not battery backed RTC on the device. A disadvantage of the sample counter is that it resets after every power event.

Using the samplecount will help immensely, thank you. Right now I am using the cdcident command to bookmark the feeds, and that is w-a-y too granular.

I will add an issue in the github.

-Pat

And using the cdcident in the trace wil not help you much. The trace is delayed depending on filters, performance and trace depth.
Got the issue.