Overclock.net banner

1 - 15 of 15 Posts

·
Registered
Joined
·
2,411 Posts
Discussion Starter · #1 ·
Here are some xVelocity plots of mine. They have some crazy data points way above 5 m/s. Is this ok? I don't see this in anybody else's MouseTester graphs.





 

·
Registered
Joined
·
1,653 Posts
yea, the few points out of range make the graph useless.
would be nice if the program would allow you to delete those, so you could actually see a curve.

xCount vs. Time, should still show you curve tho.
 

·
Registered
Joined
·
657 Posts
Well, those readings still are indicative of a problem somewhere in your system. My notebook spat out crap like that with every mouse, and mousing feels horribly swampy and imprecise on it. I bet it has to do with polling issues. If polls are missed the host will gather a load of counts during the next poll and that will make the tester think you covered a lot of ground in one reading.
Change USB ports, polling rate and observe your polling rate with mouserate.exe. You don't want your mouse in an external hub or USB 3.0 slot.
 

·
Registered
Joined
·
2,411 Posts
Discussion Starter · #4 ·
Quote:
Originally Posted by HAGGARD View Post

Well, those readings still are indicative of a problem somewhere in your system. My notebook spat out crap like that with every mouse, and mousing feels horribly swampy and imprecise on it. I bet it has to do with polling issues. If polls are missed the host will gather a load of counts during the next poll and that will make the tester think you covered a lot of ground in one reading.
Change USB ports, polling rate and observe your polling rate with mouserate.exe. You don't want your mouse in an external hub or USB 3.0 slot.
You're right! I dropped the polling rate from 1000Hz to 500Hz, and the outliers went away (see below). But should I be satisfied with this? I'm kinda not. This mouse should be able to report at 1000Hz. What could be wrong?

 

·
Registered
Joined
·
657 Posts
Try changing USB ports first. See if you get healthy readings @ 1kHz on any of your backpanel ports.
The next problem could be your processor. Try the high performance energy mode, maybe disable speedstep and C states in your BIOS. Check for background tasks that could interfere with the processor being able to issue polls consistently (got an USB keyboard or other USB devices?). Update your USB controllers (chipset drivers) and BIOS.
What's the mouse, your processor, OS?
 

·
Registered
Joined
·
779 Posts
I had some issues (not nearly this bad however) @ 1000Hz until I got my IRQ resources sorted out. Here are posts I made to guide you through it:

http://www.overclock.net/t/1433882/gaming-and-mouse-response-bios-optimization-guide-for-modern-pc-hardware-2014-r0ach-edition/1000_20#post_23334527
http://www.overclock.net/t/1433882/gaming-and-mouse-response-bios-optimization-guide-for-modern-pc-hardware-2014-r0ach-edition/1020_20#post_23338500

If that's still a no go, you may want to try uninstalling/reinstalling the latest firmware and/or software for your mouse. You could also consider uninstalling any non- default Microsoft USB controller drivers and disable USB 3.0 implementations in your motherboard BIOS.

Or maybe you're just a god and can actually move your mouse in short bursts of 150+m/s
 

·
Premium Member
Joined
·
8,764 Posts
If I keep power savings on, mousetester will show a blip of 7-10+ m/s before the test every time. If I go into the power options and set the minimum processor % to 100 the strange results go away entirely.
 

·
Registered
Joined
·
2,411 Posts
Discussion Starter · #8 ·
Ok, I've managed to improve the results, but I haven't totally fixed all the problems. HPET was incorrectly set in my BIOS. I have a 64-bit OS, but HPET was set to 32-bit mode. I fixed that, and my results have improved..



XDL5z7e.png


I will get a result like this maybe 1/10 times.



Here's a summary of the things I tried:

  • The first thing I did was undo the HIDUSBF stuff I did many years ago to use the WMO, but that had no effect.
  • Next, I checked all my USB ports and tweaked power options, but that had no effect.
  • Next, I checked for IRQ conflicts, but didn't really see any.
  • Finally, changed HPET mode from 32-bit to 64-bit in BIOS to match my 64-bit OS.
 

·
Registered
Joined
·
1,653 Posts
i get this too with alot mice, i wonder if there is actually someone who get always clean results without some dots showing at unrealistic m/s from time to time.
 

·
Registered
Joined
·
657 Posts
Seeing as how the HPET is the only thing you changed that did do something - do you have your OS use platformclock? cmd.exe -> bcdedit /set useplatformclock true

Other than that, BIOS/chipset driver updates and disabling clocking rules/C states.
 

·
Mouse reviewer
Joined
·
3,103 Posts
Quote:
Originally Posted by cuad View Post

Ok, I've managed to improve the results, but I haven't totally fixed all the problems. HPET was incorrectly set in my BIOS. I have a 64-bit OS, but HPET was set to 32-bit mode. I fixed that, and my results have improved..



XDL5z7e.png


I will get a result like this maybe 1/10 times.



Here's a summary of the things I tried:

  • The first thing I did was undo the HIDUSBF stuff I did many years ago to use the WMO, but that had no effect.
  • Next, I checked all my USB ports and tweaked power options, but that had no effect.
  • Next, I checked for IRQ conflicts, but didn't really see any.
  • Finally, changed HPET in BIOS
Quote:
Originally Posted by thuNDa View Post

i get this too with alot mice, i wonder if there is actually someone who get always clean results without some dots showing at unrealistic m/s from time to time.
That one wrong result is completely normal, I get it too occasionally. IMHO nothing to worry about anymore.

But it's interesting to see that switching HPET did something for you in terms of mouse response that is measurable (considering you only changed HPET and not anything else?)
 

·
Registered
Joined
·
2,411 Posts
Discussion Starter · #12 ·
Quote:
Originally Posted by Ino. View Post

But it's interesting to see that switching HPET did something for you in terms of mouse response that is measurable (considering you only changed HPET and not anything else?)
That's right. I only changed HPET mode from 32-bit to 64-bit to match my 64-bit OS. HPET itself had always been on, and useplatformclock had always been set to true. I did try other things before doing that, but they had no effect.
 

·
Registered
Joined
·
406 Posts
I assume these inflated velocity values are due to incorrect poll timing. If this is true, they will not cause a spike on the xCount vs. Time graph. You will likely also see a distinctly lower value at the same time on the Interval vs. Time graph.
 

·
Registered
Joined
·
2,411 Posts
Discussion Starter · #14 ·
New development, guys. I disabled multi-threading in my BIOS, and the problem came back (see below). So then I was like "To hell with it.", and I pressed the button on my motherboard to clear the CMOS. Now I'm getting decent graphs even with multi-threading disabled and HPET set to 32-bit mode. I need to do more testing...

With multithreading disabled in the BIOS.



With multithreading enabled in the BIOS.



After clearing CMOS, with multithreading disabled and HPET set to 32-bit mode.

 

·
Registered
Joined
·
2,411 Posts
Discussion Starter · #15 ·
Quote:
Originally Posted by povohat View Post

I assume these inflated velocity values are due to incorrect poll timing. If this is true, they will not cause a spike on the xCount vs. Time graph. You will likely also see a distinctly lower value at the same time on the Interval vs. Time graph.
You got it. Here is one spike on the velocity graph between 140ms and 150ms. On the interval graph, there's an aberration occurring at the same time.

Velocity graph.


Interval graph.
 
1 - 15 of 15 Posts
Top