Clock of private (SCU) timers

Bastian,

To check the speed of the clocks running on your system, you can use the omapconf tool: https://github.com/omapconf/omapconf/wiki

Simply download and "make" the omapconf source code and push the resulting binary into /system/bin.  You can run the command "omapconf audit clkspeed" to see the clock frequencies.

Regards,
Gina 

Gina,
not quiet what I need as I'd like to know the speed after u-boot runs and the relations of the clock.
But I will check the sources and try compile the relevant parts for our RTOS.

Bastian,

If you are interested in checking the clock tree / relations of the clock, you might want to check out the Clock Tree Tool, available for download here: http://www.ti.com/general/docs/wtbu/wtbudocumentcenter.tsp?templateId=6123&navigationId=12037

This tool is a Java-based stand-alone application to run on your PC.  It allows you to visualize the device clock tree, as well as see the effect of clock tree changes on the PRCM registers and vice-versa.

Regards,
Gina 

Hi Gina,

Neither the clock tree tool or omapconf will be able to answer this question. The timers in question are the those in the A9 MP core and these tools don't go to that level of detail (AFAIK). According to the ARM documentation the timer clock is the PERIPHCLK [1] and is a divided down clock from the main A9 functional clock [2]. So the clock will be based on the A9 clock, but I am unable to determine what the divider is. So we need to figure out what divider is used for OMAP.

Cheers
Jon

[1] http://infocenter.arm.com/help/topic/com.arm.doc.ddi0407i/CCHCHHDF.html
[2] http://infocenter.arm.com/help/topic/com.arm.doc.ddi0407i/CACEBBFE.html

Hi Jon
right, Gina's answer was not quite what I needed. But the clock-tool helped me to fix the speed of the CPU to exact 1GHz.
I did some measurements using a GPTIMER at sys-clock (on the Phytec Board 19.2MHz).
I found some strange results:
My speed-test loop _always_ takes 51748 ticks of the private timer no matter if I clock the core with 1GHz or 500MHz.
But counting 1s, I get 941,000,000 for 1GHz or 470,000,000 for 500MHz.

So I am slightly confused what the private timers gets as clock. Could it be some bus-cycles?

Anyway, I think I can live with the GPTIMERs.

Hi
made a mistake: I got an overflow during measurement.
The private timer seems to tick 1/34 of the core clock or bus cycles ?!