|The PS/2 ports are long obsolete garbage, USB is superior in all respects.
USB is just a middle part, it's a pipe connecting two ends. More importantly there is keyboard encoder on one end, and there is keyboard controller and buffer at the other end. As many know there are various limitation on keyboard end, such as scanning rate, blocking, ghosting... and also, what is not well known, different devices can utilize USB protocol in different ways, but what is not known at all are limitations on computer end, where USB packets are read by the host OS and stored in keyboard buffer. -- USB is good for that USB disk drive, and such, but for *KEYBOARD* with the default polling rate of 125Hz, then USB obviously sucks.
* Passmark keyboard testhttp://www.passmark.com/products/keytest.htm
With this diagnostic software for USB I get minimum of 7-8ms delay between each key, even if I press three or four of them in the same time, for PS/2 this delay is 4-5ms, PS/2 wins there.
Would you like to perform some tests, and perhaps reconsider your misdirected statement?
|I thought that's what we were doing? Here's one of my sources of info. Please be specific what research you are looking at.
Anyway I think we reached consensus in this post. The OP went with USB.
That source confirms my fears, it says - "USB keyboard is slow".
The bottleneck of processing USB packets and placing them in keyboard buffer is something USB protocol and keyboard encoder can do nothing about, all this depends on Operating System and perhaps legacy support built in hardware, and I am yet to see some proof that polling keyboard 1000 times per second instead of 125 can actually make a difference as to how many keys per second can be stored in keyboard buffer and practically available to software application.
Without performing more tests and finding out actual real world numbers the only consensus I can reach, based on my own devices I tested, is that USB for keyboards/mouse sucks, but I would surely prefer to have more experimental information to make any such conclusion at all.Edited by dan_k - 10/5/10 at 5:01pm