5KHz acquisition control

Chris Johns chrisj at rtems.org
Tue Jun 8 02:04:43 UTC 2010


On 8/06/10 9:45 AM, Joel Sherrill wrote:
> On 06/07/2010 06:43 PM, Matt Rippa wrote:
>> Hi folks -
>>
>> It sounds like one way forward would be to benchmark my
>> performance with a higher resolution system clock. I'm leaning
>> towards an external interrupt source because my worry is
>> excessive, unnecessary, processor load. With that said it may
>> not be an issue with my particular system requirements and may
>> in fact be a benefit as Till and Eric suggest. Is there some
>> guideline or limit for cranking up the system clock resolution?
>> (System failure being an obvious limit ;)). I'll play around
>> and see, but it looks like a safe bet is to have external timers
>> available just in case.
>>
> Run ticker with progressively smaller and smaller
> ticks until it quits running. That's too low. :)
>

Add to this a single task that sits in a loop incrementing a counter for 
a minute then prints the result. As you raise the frequency you can see 
how it effects the available task time.

Chris



More information about the users mailing list