USB vs PS/2 Keyboard: latency, bandwidth -> inputs per second? - Page 11 - Overclock.net - An Overclocking Community

Forum Jump: 

USB vs PS/2 Keyboard: latency, bandwidth -> inputs per second?

Reply
 
Thread Tools
post #101 of 131 (permalink) Old 10-08-2010, 01:51 AM - Thread Starter
New to Overclock.net
 
dan_k's Avatar
 
Join Date: Oct 2010
Posts: 15
Rep: 1 (Unique: 1)
Thank you Crazy9000, that's fantastic, and much appreciated.


A.) Passmark keyboard test

~2ms = 500 keys per second = 8 keys per frame @60fps

This should be quite enough, even for 2-players Street Fighter game. -- I wonder if that keyboard has special drivers, or somehow keyboard encoder has convinced Windows to change default polling rate from 125Hz to 500Hz. Did you get any special drivers with it, and did you actually install them?

---------------------------------------------------------------

B.) USB sniffer test

#1 - THREE KEYS DOWN
missing two packets, skipping...

#2 - THREE KEYS UP
3.680077 - 3.653514 = 0.026563
0.026563/3 = 0.008854 ~ 9ms

#3 - FOUR KEYS DOWN
8.965333 - 8.949534 = 0.015799
0.015799/4 = 0.003950 ~ 4ms

#4 - FOUR KEYS UP
11.445492 - 11.429372 = 0.016120
0.016120/4 = 0.004030 ~ 4ms


Ok, so, I have seen about dozen of these USB traces by now, yours is the first one with such strangely uneven distribution of lag between the packets. Usually packets are sent at steady rate, or fluctuate a little bit, but in your case there is some impossibly super small latency, except for the last packet where this lag is huge and interestingly evens things up at the end.

So, I took the time of the 1st packet in a batch minus last packet and then divided by the number of keys being reported to get average latency per packet. Unfortunately this data sample does not quite match with the 1st test, so I do not really know what to say, except that both programs are under slight suspicion now.

And lastly, your keyboard too uses one packet per "state change", i.e. instead of to send "4 keys are down" in one packet it sends one packet for each key, but that might be just the way it is supposed to be, I don't know that yet, though it sure is a waste.
dan_k is offline  
Sponsored Links
Advertisement
 
post #102 of 131 (permalink) Old 10-08-2010, 08:10 AM
I might have tacos tonite
 
Crazy9000's Avatar
 
Join Date: Mar 2006
Location: Seattle, Washington
Posts: 22,289
Rep: 489 (Unique: 344)
The keyboard is just using the default windows drivers.

If something was missing from that log, maybe I did something wrong, IDK. I ran the test a few times then, since I was figuring out how to apply the filter (didn't realize you had to do that first). Here's another one that's done right from the start.
LL

Crazy9000 is offline  
post #103 of 131 (permalink) Old 10-08-2010, 05:52 PM - Thread Starter
New to Overclock.net
 
dan_k's Avatar
 
Join Date: Oct 2010
Posts: 15
Rep: 1 (Unique: 1)
I think you are doing everything right, I often get first packets missing too, but this time you also have a packet missing in last two sets with 4 keys, so the sequence has three instead of four packets. Nevertheless I consider that data to be valid since we only really need to know the time of the 1st packet in a batch and of the last, i.e. the time taken for all the data to be sent.



# - THREE KEYS DOWN
batch of three packets missing...

# - THREE KEYS UP
3.298223 - 3.273155 = 0.025068
0.025068 / 3 = 0.008356 ~ 8ms

# - FOUR KEYS DOWN
7.869663 -7.849103 = 0.020560
0.020560 / 4 = 0.005140 ~ 5ms

# - FOUR KEYS UP
9.667415 - 9.641065 = 0.026350
0.026350 / 4 = 0.006586 ~ 7ms


Similar results as the last time, bit worse maybe. The problem is that in all previous tests USB trace data matched with Passmark keyboard test, but in your case one test says 2ms, while the other shows 4ms to 8ms. I don't know which one is closer to truth, but Passmark keyboard test should be closer to measuring what I actually want to know, which is the maximum performance of keyboard buffer & keyboard controller.

I was hoping to extract that information indirectly from such data samples as USB trace, but there seem to be so many things interdependent on each other that is hard to find out what part of hardware or software is the actual bottleneck, especially since this bottleneck might not be at the same place for everyone, and this makes all the numbers slightly more meaningless.
dan_k is offline  
Sponsored Links
Advertisement
 
post #104 of 131 (permalink) Old 10-12-2010, 04:38 AM
New to Overclock.net
 
EvanPitts's Avatar
 
Join Date: Sep 2010
Location: The Hammer
Posts: 1,484
Rep: 87 (Unique: 73)
Quote:
Originally Posted by De-Zant View Post
MAN... I just LOVE my PS/2 keyboard....
Do you love your PS/2 computer though?

Thought so - long since obsolete, seeing they haven't made one in a decade and a half...
EvanPitts is offline  
post #105 of 131 (permalink) Old 10-12-2010, 05:57 AM
 
Join Date: Apr 2008
Location: Planet Earth
Posts: 4,124
PS/2.

Retro is in. Floppies for decreased latency. The PS/2 Model 25.

The BirthPlace of the PS/2 Port
ripster is offline  
post #106 of 131 (permalink) Old 08-21-2011, 02:03 PM
New to Overclock.net
 
QuackPot's Avatar
 
Join Date: Dec 2010
Posts: 568
Rep: 12 (Unique: 12)
So do USB keyboard may any difference when they use a PS/2 adaptor?

Edit: I know this is an old thread but the post above mine are relevant and interesting to the topic at hand.

JF-AMD on Bulldozer:

Quote:
Originally Posted by JF-AMD go_quote.gif
IPC will be higher
Single threaded performance will be higher

That is all we can say at this point.
QuackPot is offline  
post #107 of 131 (permalink) Old 08-21-2011, 02:11 PM
I might have tacos tonite
 
Crazy9000's Avatar
 
Join Date: Mar 2006
Location: Seattle, Washington
Posts: 22,289
Rep: 489 (Unique: 344)
Quote:
Originally Posted by QuackPot;14665676 
So do USB keyboard may any difference when they use a PS/2 adaptor?

Edit: I know this is an old thread but the post above mine are relevant and interesting to the topic at hand.

If they support ps/2 adapter, its the same as having a ps/2 connector on the end.

Crazy9000 is offline  
post #108 of 131 (permalink) Old 08-21-2011, 02:47 PM
 
Join Date: Apr 2008
Location: Planet Earth
Posts: 4,124
Why would anyone use PS/2 connectors in this day and age of USB 3.0 and cheap USB hubs?

This Tardis one is quite appealing.
tardisusbhub01pr7.png
ripster is offline  
post #109 of 131 (permalink) Old 08-21-2011, 02:53 PM
Retired Staff
 
Riou's Avatar
 
Join Date: Aug 2009
Location: duffle bag
Posts: 13,795
Rep: 817 (Unique: 626)
Quote:
Originally Posted by ripster;14666062 
Why would anyone use PS/2 connectors in this day and age of USB 3.0 and cheap USB hubs?
You cannot cold boot a computer running on BIOS with a USB keyboard.

Riou is offline  
post #110 of 131 (permalink) Old 08-21-2011, 02:55 PM
 
Join Date: Apr 2008
Location: Planet Earth
Posts: 4,124
True.

That was exciting. My Corsair 600T has a power button on the top.
corsair_se_white_600t_07_thumb.jpg

Hey, look at all those USB ports!
ripster is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off