• Locked Locked
  • Replies 2 replies
  • Subscribers 98 subscribers
  • Views 324 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

OMAP3530 Power Estimation Spreadsheet

Oscar Herranz
Intellectual 770 points
Other Parts Discussed in Thread: OMAP3530

Hi guys!

First of all, I would like to note some topics about my particular situation: I'm studying power consumption at the BeagleBoard (based in omap3530) and the variations of that power consumption by means different scenarios at the platform (running a program only at the ARM, running a program that envolves both cores, ARM and DSP, ...). So, the power consumption is measured by a software tool provided by Agilent, which obtains the current supplied for the power supply. In order to verify the results measured by the software tool , I'm using the power estimation spreadsheet, where I write the values of the %CPU of both cores. In my application I'm able to obtain these %CPUs, therefore I obtain the power results quickly. Comparison between both scenarios causes different conclusions.

Focus in the ARM behaviour. First, at the power estimation spreadsheet, the maximum variation (0 up to 100 %CPU ARM) of power is nearly to 1 W @OPP6, ARM Subsytem: Cortex A8 + Neon. However, at the real situation (power consumption measured by power supply), that maximum variation is almost of 0.5 W!! The procedure to obtain this result is to run a program that produces a high CPU load (near to 100%), i.e. a video decoding task. Besides of that, I have estimated power consumption of the BeagleBoard, so I can obtain that maximum variation.

I have eliminated the action of any peripheral of the BB and the spreadsheet too, taking in account just the action of ARM...

What could be happening? I have checked all parameters, measuring many different situations but I have not found the key so far... Any suggestion would be appreciated!

Regards

Herry ALow

  • TI__Guru 65405 points

    Hi Oscar,

    The power consumption of the BeagleBoard consists of the sum of consumptions of all elements on the board not only OMAP3530. Could you get the values not only for variation but measured values too. You should pay attention on the power circuit and tps.

    BR

    Tsvetolin Shulev

  • in reply to Cvetolin Shulev-XID
    Intellectual 770 points

    Thanks for the answer, Tsvetolin.

    Of course the power consumption of BB consists of the rest of elements on the board, but I can isolate the power consumption only the ARM core (procedure isn't perfect, but it's ok for a good approach).

    By the other hand, at the ARM subsystem section (in the omap3530 spreadsheet), there are 4 four possible configurations: N/A, ARM+NEON, Dhrystone standalone and Saxpy loop under Linux. In my case, an embedded OS (Angstrom-Linux) is running at the ARM, but I'm not using the NEON. What is the correct configuration for me?? I do not know the influence of NEON over the ARM power consumption...

    If anyone have had some kind of experience with spreadsheet, please let me know. We can discuss results and many things...

    Thanks a lot

    Regards