Improve polling speed and reduce latency

[DvaMishkiLapa]

Anything less than 6000 doesn't change latency, but it does put a load on the CPU. I don't know how it works. By making the polling rate 144hz in a variable the author was clearly guided by something, but I can't find what. Either I don't understand how the speed is actually adjusted.

In any case, before the minimum from the client was 6000, I accidentally changed it to 4000 for some reason. I have not been able to get a delay of less than 7ms. It feels like it's some kind of limitation of the WiFi adapter.

I've read VitaDev and VitaSDK, but I can't find anything about WiFi settings, or any ways to configure it.

I tested the evening with this, pooling-rate above 8000 was starting to noticeably worsen latency.
Anything below 6000 just makes the PS Vita warm the air, without affecting the data rate in any way.

[santarl]

Interesting, I think we can look into this further after we refactor the code

[saidsay-so]

Anything less than 6000 doesn't change latency, but it does put a load on the CPU. I don't know how it works. By making the polling rate 144hz in a variable the author was clearly guided by something, but I can't find what. Either I don't understand how the speed is actually adjusted.

In any case, before the minimum from the client was 6000, I accidentally changed it to 4000 for some reason. I have not been able to get a delay of less than 7ms. It feels like it's some kind of limitation of the WiFi adapter.

I've read VitaDev and VitaSDK, but I can't find anything about WiFi settings, or any ways to configure it.

I tested the evening with this, pooling-rate above 8000 was starting to noticeably worsen latency. Anything below 6000 just makes the PS Vita warm the air, without affecting the data rate in any way.

Well, I'm not particularly proud of this part of the code…
I was just trying to evaluate the influence of the polling rate on the input lag and how it correlates with the refresh rate (not a smart choice in retrospect).

The heartbeat logic could be refactored to have a more standard behavior, like you suggested.