This post is older than 2 years and might not be relevant anymore
More Info: Consider searching for newer posts

What is the meaning of "application processing" ?

Hello, I use sdk 8.0.0 and s110. I read this blog

And I did not understand some terms;

  1. Application processing(for example 2ms each adv interval)
  2. Number of packets to send in each connection interval(for example 1 packet)

Where and how do we organize these terms? How do we change it?

Parents
  • Application processing is the when ARM Coretx-M MCU executes code of your application FW. In other words what you compile into binary and is processed at given interrupt or other trigger.

    Number of packets is linked to mechanism how BLE Link Layer works. If you understand what Connection interval is and that normally (unless packet is lost or one of the sides uses Slave/Master Latency) two PDUs are exchanged (one Master -> Slave and one Slave -> Master) then it should be easy to understand that if both sides are able to execute more than one of such PDU-pair. If any of sides have more then one packet ready before connection event is scheduled and the opposite side reflects on MD bit in PDU header (More Data) then another PDU pair is exchanged until there are no more data o one of the sides have limit within the interval (these are stack and configuration specific).

  • If you have questions to a blog post then post it under it in the discussion. If it is just dummy experiment like this you can easily place loops full of some dummy instructions (like NOP) and simulate MCU processing by the app. Because nRF5x chps have fixed frequency of ARM Cortex-M core it's quite easy to simulate certain occupation time. Indeed it's more difficult to solve opposite problem: measure processing time of real application code. But that's what I discussed above. If the simulation is the only problem then it's easy and solved I guess.

Reply
  • If you have questions to a blog post then post it under it in the discussion. If it is just dummy experiment like this you can easily place loops full of some dummy instructions (like NOP) and simulate MCU processing by the app. Because nRF5x chps have fixed frequency of ARM Cortex-M core it's quite easy to simulate certain occupation time. Indeed it's more difficult to solve opposite problem: measure processing time of real application code. But that's what I discussed above. If the simulation is the only problem then it's easy and solved I guess.

Children
No Data
Related