geekhack
geekhack Community => Keyboards => Topic started by: dmbr on Fri, 16 August 2013, 21:02:33
-
I just attempted to overclock my CM Quickfire Rapid to a 1000hz polling rate like its longer sibling the "Pro", using HIDUSBF. The keyboard functions fine now, but I'd like to confirm that the OC is working properly; however, I have only been able to find software that confirms the polling rate of mice.
Anyone aware of a program that could help me out?
Thanks!
-
Overclock it more, bro! If you can't notice the speed difference, it's TOO slow.
-
try hitting a key 1000 times per second and see if it registers all 1000 times.
-
I have no answer but just curious what difference does it make?
-
I'm a competitive gamer, so input delay is the enemy. I'd like my movements to register in-game as fast as possible. With a stock 125hz polling rate and less-than-perfect FPS, the delay can be as high as ~30ms, which is enough to matter.
-
Most games are running VSync locked. 60Hz. No more, and often MUCH less!
If you want to talk latency - LCD monitors its about 8ms, add 10ms for USB, let alone whatever packet latency you have 50+ms...
This is why PS/2 beats everything else. Its so incredibly fast - nothing else can match it. (Mabey a Real-Time app....)
-
No serious FPS gamer uses v-sync, and I myself have a 120hz monitor with a 1ms response time.
P/S2 is nearly twice as slow as stock USB, so not sure where you're coming from there...
Anyway, not looking for a debate, just an answer.
-
I'm a competitive gamer, so input delay is the enemy. I'd like my movements to register in-game as fast as possible. With a stock 125hz polling rate and less-than-perfect FPS, the delay can be as high as ~30ms, which is enough to matter.
polling rate doesn't matter.
there is a mechanical limit in debouncing which is in the ~10 ms range. I also suggest you explore how PS/2 works: interrupts beat USB polling.
-
USB protocol analyzer?
-
USB protocol analyzer?
^This.
Other than looking at the quoted spec, the only way I see to actually verify the polling rate would be getting a protocol analyzer and having it looked at.
Additionally, I don't want to get into the merits of PS2 vs USB, but it's true that cherry switches take some time to debounce (roughly 10ms or so, I believe dropping this to 8ms leads to some chatter) so upping the polling rate will not gain you much as the keyboard's latency will be determined more by the matrix scan rate, which is limited by the debounce time of the switches. I suspect this is not the case with capacitive switches such as topre or ibm flip plate (model F), s well as hall effect switches.
There is a benchmark that I believe tries to measure the keyboard somewhat, perhaps passmark or something? I haven't sued it and dont know much.
-
I don't think a USB analyzer will work.
Because polling happens at the keyboard's controller unit. If you do not type anything, the polling still occurs but may not have sent anything via USB.
-
From a razer FAQ page:
How does Razer measure the polling rate (ultrapolling) and response rates of their keyboards and mice?
Razer uses digital scopes and USB analyzers connected directly to the motherboard to obtain real and accurate measurements of polling rates.
-
Even with debounce, there can be an improvement in overclocking. You see, debounce impacts how long after the key is pressed that the keyboard can register when the keyboard can consider that key unpressed. Debounce does not impact how long after the key is pressed that the keyboard can report to the USB controller that it has been pressed. To summarize, a well programmed debounce filter should never add latency to the initial key press, but will add latency to a quick key release after a keypress.
That being said, you have to be really really good and have a great network connection for the 3.5 ms on average that you gain going from 125 Hz to 1000 Hz to matter.