DSP BIOS RTA does not work reliably with same application running from RAM, but running from FLASH is ok.

What version of BIOS are you using?

Thanks,

Chris

Hello Chris! 

Sorry I missed that...it is version: bios_5_41_07_24

Thanks,

Krishna

With the versions of BIOS and CCS that you're using, you should at least be seeing data when you halt the target.

See the information on stop mode here:

http://processors.wiki.ti.com/index.php/BIOS_5_Real-Time_Analysis_(RTA)_in_CCSv4

Are you seeing anything in the Raw Logs view when you halt the target?

Thanks,

Chris

Chris,

Sometimes when I halt the target I see incomplete data in the Raw Logs.  Most times I don't see anything at all...

If I keep repeating the test with a restart, run and then Halt  or a complete CCS restart run and then Halt, I see the Raw Logs sometimes.

However, I never see the CPU load data...which is the main thing I am interested in at the moment.  CPU load graph does not work at all.

Can you please advice?

Thanks,

Krishna

A couple things:

• The Statistics view and CPU Load graph are not supported by stop-mode. 
  • http://wiki.tiprocessors.com/index.php/BIOS_5_Real-Time_Analysis_(RTA)_in_CCSv4#Supported_Views
• Stop mode should be 100% reliable for retrieving the raw logs. When you halt the target, it should pick up all of the records currently sitting on the target (all of the records, but not the STS objects). If you think that stop mode isn't working correctly, can you include a screenshot of the raw logs view and point out what you think is missing?

To get the CPU load graph to work, you need RTDX to be functioning properly to support run mode. RTDX is no longer supported, so there's not much we can do there except to suggest workarounds (such as toggling the 'Stream RTA Data' button off and back on when it seems like RTDX has failed).

Thanks,

Chris

Chris,

Thanks for your feedback.  Please see the attached file with the screen shots...5545.dsp_bios_screen_shots_04182011_rev1.docx

Let me know if there is a problem in seeing it.

I think at this point it will be extremely productive if we can have a quick phone call.  I can walk through the screen shots with you and ask you a couple of questions.  Is that possible? 

Thanks,

Krishna

Hi Krishna,

Thanks for the screenshots, they're helpful. I'd probably only resort to a call if I felt there was a solvable bug here (i.e., not an issue with RTDX) and I couldn't reproduce it on my end; hopefully the following is helpful.

For number 3, it looks like there's a problem with RTDX (which is why you're not receiving any STS or CPU load events), but you're receiving stop mode data successfully, so that's good. The gap in the LedPrdCount events (from 14 to 76) is probably due to the logger wrapping and overwriting older events.

There are known issues with RTDX and BIOS 5; RTDX is known to fail intermittently when reloading the program, or when repeatedly running to a breakpoint. It works best when you start with a fresh debug session.

If RTDX is being too flaky, and you're interested in the CPU load, one thing you could try is manually servicing the cpu load STS object, and logging the current load with LOG_printf. You could then retrieve the load values from the 'trace' log with stop mode. See the following article for calculating CPU load on the target.

http://processors.wiki.ti.com/index.php/Calculating_DSP/BIOS_5_CPU_Load

Thanks,

Chris

Hi Chris,

Thanks for the info...I would still like to request a short call because I would like to ask you some questions on interpreting some of the data that I am seeing in the first screen shot, for example.

Also, please note that I am working with the same application but it is operating to run from FLASH in one case and RAM in the other case.  The DSP BIOS data...raw logs etc., are as expected when running from FLASH, the issues occur when running from RAM.  Can you please shed some light on this...why the difference?  do you think it is still RTDX related?  Does running from RAM cause the RTDX more frequently or something?

Thanks,

Krishna

Chris,

Thanks for the feedback.  Yes please call me 281-285-8432.  I have a meeting at 3:00pm CST and it is 2:15pm now.

If we miss each other, please leave me a call back number or propose a time for tomorrow.

Thanks,

Krishna