My understanding is that it could affect either.
GPIO:P0 should be pulled high (via 10k), or the key-scan feature of the device may think a key on a keyboard is pressed and it will stay awake to perform further scanning.
Based on my understanding of this thread, it appears that the fine timer could indeed affect power consumption: FineTimer
Thanks! I'll try to work up a new prototype that I can test this with.
For the attached thread, I'm don't see a correlation with power. Most of this discussion seems to revolve around attempting to have very small cycles. What I'm trying to determine is if increasing the fine timer interval from 250ms to 500ms would have a meaningful impact on power consumption - assuming I can get into sleep mode.
I checked with the developers today and most definitely, on the BCM20736/37, there is an impact on power consumption that is proportional to the fine timer setting.
Essentially, if the fine timer fires more often (i.e. wakes up, handles an event, then goes back to sleep), the part will consume more power.
On the BCM20732 (different firmware), this was not as much the case because the firmware/fine timer function was architected differently.