Cherry removes debounce almost entirely!

Everything about latency. Tips, testing methods, mouse lag, display lag, game engine lag, network lag, whole input lag chain, VSYNC OFF vs VSYNC ON, and more! Input Lag Articles on Blur Busters.
Sparky
Posts: 682
Joined: 15 Jan 2014, 02:29

Re: Cherry removes debounce almost entirely!

Post by Sparky » 18 May 2015, 08:28

BTRY B 529th FA BN wrote:I do notice a difference in overall machine performance between a 1k hz polling keyboard VS one plugged in via PS/2 in some games. Specifically CPU intensive games
If USB polling as a significant impact on system performance it should show up in normal benchmarks, hell, it should show up in task manager. I can't see a difference.

Do you have a test condition that consistently produces the issue?

Glide
Posts: 280
Joined: 24 Mar 2015, 20:33

Re: Cherry removes debounce almost entirely!

Post by Glide » 18 May 2015, 22:01

Sparky wrote:
BTRY B 529th FA BN wrote:I do notice a difference in overall machine performance between a 1k hz polling keyboard VS one plugged in via PS/2 in some games. Specifically CPU intensive games
If USB polling as a significant impact on system performance it should show up in normal benchmarks, hell, it should show up in task manager. I can't see a difference.

Do you have a test condition that consistently produces the issue?
With very high CPU usage in some applications, my 500/1000Hz mice will be unusable (cursor updates once every few seconds) while 125Hz mice seem to respond far quicker to inputs.

I know that's not USB vs PS/2, but I do wonder if USB3 with its interrupt support would solve this problem.

flood
Posts: 929
Joined: 21 Dec 2013, 01:25

Re: Cherry removes debounce almost entirely!

Post by flood » 18 May 2015, 23:46

yea but for keyboards it's a different situation as it doesn't send 1000 state changes a second

Sparky
Posts: 682
Joined: 15 Jan 2014, 02:29

Re: Cherry removes debounce almost entirely!

Post by Sparky » 19 May 2015, 12:11

Glide wrote:
Sparky wrote:
BTRY B 529th FA BN wrote:I do notice a difference in overall machine performance between a 1k hz polling keyboard VS one plugged in via PS/2 in some games. Specifically CPU intensive games
If USB polling as a significant impact on system performance it should show up in normal benchmarks, hell, it should show up in task manager. I can't see a difference.

Do you have a test condition that consistently produces the issue?
With very high CPU usage in some applications, my 500/1000Hz mice will be unusable (cursor updates once every few seconds) while 125Hz mice seem to respond far quicker to inputs.

I know that's not USB vs PS/2, but I do wonder if USB3 with its interrupt support would solve this problem.
That sounds more like a thread priority issue, and I remember that happening frequently back when I used a system with a single core CPU. I can't remember it happening recently, but it's a lot harder to fully saturate a 4 core system.
Last edited by Sparky on 19 May 2015, 12:19, edited 2 times in total.

Glide
Posts: 280
Joined: 24 Mar 2015, 20:33

Re: Cherry removes debounce almost entirely!

Post by Glide » 19 May 2015, 12:16

Sparky wrote:That sounds more like a thread priority issue, and I remember that happening frequently back when I used a system with a single core CPU. I can't remember it happening recently, but it's a lot harder to fully saturate a 4 core system.
It is difficult, but some applications can still manage it. I'm using a quad-core system.

Sparky
Posts: 682
Joined: 15 Jan 2014, 02:29

Re: Cherry removes debounce almost entirely!

Post by Sparky » 19 May 2015, 12:27

Glide wrote:
Sparky wrote:That sounds more like a thread priority issue, and I remember that happening frequently back when I used a system with a single core CPU. I can't remember it happening recently, but it's a lot harder to fully saturate a 4 core system.
It is difficult, but some applications can still manage it. I'm using a quad-core system.
Which programs? It really doesn't sound like a USB problem, it sounds like a general system responsiveness and resource prioritization problem. If the program you have an issue with has a linux version, does it happen there too?
flood wrote:yea but for keyboards it's a different situation as it doesn't send 1000 state changes a second
Not quite. Every poll, the keyboard responds with the whole set of currently pressed keys. It's PS/2 keyboards that only send state changes.

Glide
Posts: 280
Joined: 24 Mar 2015, 20:33

Re: Cherry removes debounce almost entirely!

Post by Glide » 19 May 2015, 12:32

Sparky wrote:
Glide wrote:
Sparky wrote:That sounds more like a thread priority issue, and I remember that happening frequently back when I used a system with a single core CPU. I can't remember it happening recently, but it's a lot harder to fully saturate a 4 core system.
It is difficult, but some applications can still manage it. I'm using a quad-core system.
Which programs? It really doesn't sound like a USB problem, it sounds like a general system responsiveness and resource prioritization problem. If the program you have an issue with has a linux version, does it happen there too?
Things like video encoding and certain audio decoders when batch-processing a large number of files can cause it to happen.

Sparky
Posts: 682
Joined: 15 Jan 2014, 02:29

Re: Cherry removes debounce almost entirely!

Post by Sparky » 19 May 2015, 12:44

Glide wrote:
Sparky wrote:
Glide wrote:
Sparky wrote:That sounds more like a thread priority issue, and I remember that happening frequently back when I used a system with a single core CPU. I can't remember it happening recently, but it's a lot harder to fully saturate a 4 core system.
It is difficult, but some applications can still manage it. I'm using a quad-core system.
Which programs? It really doesn't sound like a USB problem, it sounds like a general system responsiveness and resource prioritization problem. If the program you have an issue with has a linux version, does it happen there too?
Things like video encoding and certain audio decoders when batch-processing a large number of files can cause it to happen.
And something like Prime 95 doesn't? It sounds like those programs are using a higher priority than they should be.

Glide
Posts: 280
Joined: 24 Mar 2015, 20:33

Re: Cherry removes debounce almost entirely!

Post by Glide » 19 May 2015, 13:33

Sparky wrote:And something like Prime 95 doesn't? It sounds like those programs are using a higher priority than they should be.
No, it doesn't. Priority is "normal" for their processes, and only 500/1000Hz mice are badly affected.

flood
Posts: 929
Joined: 21 Dec 2013, 01:25

Re: Cherry removes debounce almost entirely!

Post by flood » 19 May 2015, 14:27

windows is just shitty :D

Post Reply