[QUOTE=Fred;424706]Thanks all for your thoughts and ideas. I think I have the problem resolved, although it's still a little confusing. Turns out I think it was an issue with my mouse. I have a logitech wireless M510 mouse. When I switch to a different wireless mouse (had a logitech M310 handy), the problem goes away. Very strange, but easy fix.

You would think it's something just up with that mouse (I ruled out batteries, etc), but the original mouse behaved perfectly as long as Prime95 wasn't crunching away. Strange, but glad it's resolved.[/QUOTE]

I have an old crappy pair of Logitech webcams and when I have them both plugged in and streaming basic 640x480 video, sure enough my CPU usage spikes horribly.

They're USB and the drivers may or may not play a part in that, or just that they're both flooding USB with a lot of traffic or who knows what.

I can only imagine what would happen if I were running Prime95 at the same time because if it's interrupt driven, there will be some mild system thrashing going on.

With that said, I'd speculate it could be similar to your mouse problem with the M510, but the M310 is better.

Like Mini-Geek says, maybe there are some settings for the mouse about the rate... if you move the mouse an inch and it generates a ton of activity because the dpi is cranked way high (for gaming purposes or who knows what), I could see that doing strange things if the CPU is otherwise highly active.

One way an application could affect the interrupt latency is if it pushes the interrupt service code out of the cache(s) and increases the time to load the service code from RAM. Another possible effect (in the reverse direction which would improve response time) is that keeping the CPU active, and out high of C-states, will make everything happen faster. I think in both of these cases the difference is micro-seconds and wouldn't be noticeable to the user. So what else could it be?

[QUOTE=retina;424721]One way an application could affect the interrupt latency is if it pushes the interrupt service code out of the cache(s) and increases the time to load the service code from RAM. Another possible effect (in the reverse direction which would improve response time) is that keeping the CPU active, and out high of C-states, will make everything happen faster. I think in both of these cases the difference is micro-seconds and wouldn't be noticeable to the user. So what else could it be?[/QUOTE]

Perhaps a driver that also runs at idle priority, so it'll be fighting for scraps of time with a CPU heavy app? Doesn't seem likely though.

It really does not seem that this thread is an appropriate target for Title Morphing. A new user is involved, and it is a technical issue.

[QUOTE=Fred;424706]Thanks all for your thoughts and ideas. I think I have the problem resolved, although it's still a little confusing. Turns out I think it was an issue with my mouse. I have a logitech wireless M510 mouse. When I switch to a different wireless mouse (had a logitech M310 handy), the problem goes away. Very strange, but easy fix.

You would think it's something just up with that mouse (I ruled out batteries, etc), but the original mouse behaved perfectly as long as Prime95 wasn't crunching away. Strange, but glad it's resolved.[/QUOTE]

I switched from a M310 to a M510 and have never had an issue, so it could well be a faulty M510.
I am using:-
Windows 7
Setpoint Control Centre version 6.67.83
Driver version 5.90.41

I suppose another possibility is that the driver is not fully compatible with Windows 10.

Prime95 stresses a pc more than almost anything else. It is fairly normal to have a slightly sluggish pc when it is running. This was more of an issue on single core pcs. It can also effect pcs with more cores. My old Athlon 64 3800 was terrible. Memory bandwidth starved pcs can be bad ones. Memory bandwidth would be an issue with that cpu although we haven't had loads of people complain about it. What speed is your memory?

[QUOTE=henryzz;424840]Prime95 stresses a pc more than almost anything else. It is fairly normal to have a slightly sluggish pc when it is running. This was more of an issue on single core pcs. It can also effect pcs with more cores. My old Athlon 64 3800 was terrible. Memory bandwidth starved pcs can be bad ones. Memory bandwidth would be an issue with that cpu although we haven't had loads of people complain about it. What speed is your memory?[/QUOTE]

DDR3-1600, I'm sure that's not helping anything.

Honestly at this point thought, the M310 is working perfect with Prime95 happy chugging away, so I just assume not troubleshoot any further. At some point, the ROI of time vs. a $20 mouse just aint there ;)

[QUOTE=Fred;424855]DDR3-1600, I'm sure that's not helping anything.

Honestly at this point thought, the M310 is working perfect with Prime95 happy chugging away, so I just assume not troubleshoot any further. At some point, the ROI of time vs. a $20 mouse just aint there ;)[/QUOTE]

I've run Prime95 on current systems with DDR2-800 speeds and while it's nowhere near what I can get out of other systems, I haven't seen any other performance issues in the process.

I wouldn't make a blanket statement that you'll never have an issue when Prime95 is running, but it's more likely that there's just some other hardware, driver, app, whatever that simply doesn't play nice with others. Maybe a particular driver that expects there will always be ample CPU for it to do it's magic, or a PSU that works great as long as your not stressing the system and reaching peak TDP for CPU/memory, or a CPU that does fine in lower c-states but chokes when pushed to the max.

One thing is for sure... if your system is marginal in any large or small way, there's a good chance Prime95 will reveal that. Which makes it an awesome stress tester.