Strange CPU utilization percentages

TennesseeTony

Elite Member
Aug 2, 2003
4,204
3,631
136
www.google.com
Typically I can run BOINC at 99% CPU Utilization on any machine, effectively turning off that last thread, to keep the system responsive and allow the GPU to have a little something to play with. And with such a setting I usually see a predictable, calculable, CPU Utilization percentage.

My hexcore i7 (12 thread) can be expected to use 8.33% (100% divided by 12) per thread, and 99% setting cuts the utilization down to about 96%, as background processes bump it up some.

I've noticed on my Xeon systems, that disallowing several threads to be used for BOINC still results in my CPU utilization to be shown as 100%. This holds true for the Workstations as well as the newest systems. On BumbleBee and WUSS I have to disable 4 threads (BOINC set to 88%) to get under 100% utilization.

Further, I've noticed, by using Kill-a-Watt meters, that my power usage goes UP (only 1-3 watts more) when I finally disable enough threads in BOINC to obtain <100% CPU Utilization.

And in practice, I've managed to observe that identical systems, running the same projects, do indeed score significantly different depending on if the CPU is maxed out or not. Bee was set to 99% (one thread disabled), WUSS was set to 92% (three threads disabled and still showing 100%, but Turbo'ing higher), and over a four day period WUSS scored about 19,000 higher in Universe than Bee. Worse yet for Bee, he has a 2% overclock. :)

(Oh, and I just discovered you can keep track of your individual computer's results in FreeDC. How cool is that!!?)

In summary: Having used numerous monitoring tools, the best I can come up with as an explanation, is that the Xeon's are hard locked to the TDP, or thermal design power, and standard (or should I say, overclockable) desktop CPUs are not. With less cores running, the remaining cores turbo to higher clocks, keeping the TDP maxed out. The higher power consumption observed at 98-99% vs 100% can possibly be attributed to the turbo clocks not having to switch up and down to maintain TDP.

So I believe the CPU Utilization on a (Windows) Xeon, is based on TDP, instead of threads in use.

Has anyone else experienced this, or can either confirm or deny my findings and assumptions?

EDIT: Would disabling Turbo result in predictable percentages?

EDIT: Xeon's running WinTen, Hex-core i7 still on Win7.


&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9604;&#9604;&#9604;&#9604;&#9604;&#9604;&#9604;&#9604;&#9604;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9604;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9604;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;
&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;
&#9616;&#9608;&#9608;&#9608;&#9608;&#9612;&#9472;&#9600;&#9608;&#9608;&#9608;&#9604;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9604;&#9608;&#9608;&#9608;&#9600;&#9472;&#9616;&#9608;&#9608;&#9608;&#9608;&#9612;
&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9604;&#9472;&#9472;&#9600;&#9608;&#9608;&#9608;&#9604;&#9472;&#9472;&#9472;&#9604;&#9608;&#9608;&#9608;&#9600;&#9472;&#9472;&#9604;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;
&#9472;&#9616;&#9608;&#9608;&#9600;&#9608;&#9608;&#9608;&#9604;&#9472;&#9472;&#9600;&#9608;&#9608;&#9608;&#9604;&#9608;&#9608;&#9608;&#9600;&#9472;&#9472;&#9604;&#9608;&#9608;&#9608;&#9600;&#9608;&#9608;&#9612;&#9472;
&#9472;&#9472;&#9608;&#9608;&#9608;&#9604;&#9600;&#9608;&#9608;&#9608;&#9604;&#9472;&#9472;&#9600;&#9608;&#9608;&#9608;&#9600;&#9472;&#9472;&#9604;&#9608;&#9608;&#9608;&#9600;&#9604;&#9608;&#9608;&#9608;&#9472;&#9472;
&#9472;&#9472;&#9616;&#9608;&#9604;&#9600;&#9608;&#9604;&#9600;&#9608;&#9608;&#9608;&#9472;&#9604;&#9472;&#9600;&#9472;&#9604;&#9472;&#9608;&#9608;&#9608;&#9600;&#9604;&#9608;&#9600;&#9604;&#9608;&#9612;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9604;&#9600;&#9608;&#9604;&#9608;&#9608;&#9472;&#9608;&#9608;&#9604;&#9608;&#9608;&#9472;&#9608;&#9608;&#9604;&#9608;&#9600;&#9604;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9472;&#9600;&#9608;&#9608;&#9608;&#9604;&#9600;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9600;&#9604;&#9608;&#9608;&#9608;&#9600;&#9472;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9608;&#9604;&#9472;&#9600;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9600;&#9472;&#9604;&#9608;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9604;&#9472;&#9472;&#9472;&#9472;&#9604;&#9608;&#9472;&#9608;&#9608;&#9608;&#9472;&#9608;&#9604;&#9472;&#9472;&#9472;&#9472;&#9604;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9604;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9604;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9604;&#9604;&#9604;&#9604;&#9604;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9472;&#9600;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9608;&#9608;&#9608;&#9600;&#9472;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9600;&#9608;&#9472;&#9608;&#9608;&#9608;&#9472;&#9604;&#9604;&#9604;&#9604;&#9604;&#9472;&#9608;&#9608;&#9608;&#9472;&#9608;&#9600;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9600;&#9608;&#9612;&#9616;&#9608;&#9608;&#9608;&#9608;&#9608;&#9612;&#9616;&#9608;&#9600;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;
&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9608;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;
 
Last edited:

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
In summary: Having used numerous monitoring tools, the best I can come up with as an explanation, is that the Xeon's are hard locked to the TDP, or thermal design power, and standard (or should I say, overclockable) desktop CPUs are not. With less cores running, the remaining cores turbo to higher clocks, keeping the TDP maxed out. The higher power consumption observed at 98-99% vs 100% can possibly be attributed to the turbo clocks not having to switch up and down to maintain TDP.

Well if one of your numerous monitoring tools is core temp then you can tell us if indeed frequencies go higher up with more cores disabled.

In general yes,that's how turbo works it's pushing the cores while remaining within TDP.
 

TennesseeTony

Elite Member
Aug 2, 2003
4,204
3,631
136
www.google.com
I guess I'm more confused with the way Windows Task Manager is reporting the usage, as you can see in the pic below, the actual usage is much lower than 100%. I've not seen that on desktop, overclockable processors before.

Bumping the usage to 100%, the clockspeeds were jumping around but didn't dip any lower than 2699MHz, and the CPU package temps were 1 degree higher ( clocks were semi-stable in the pic below). The watts used didn't change, but the DRAM wattage decreased by 1 watt.

I'm guessing the milli-seconds lost during a clock change is enough to affect performance. Regardless of what's going on, it certainly appears to perform better, and have more output, when WinTask Manager reports less than 100% load. But.....it's just so hard to accept that disabling 3-4 threads HELPS performance. :)

33mwnrn.png
 

waffleironhead

Diamond Member
Aug 10, 2005
6,917
429
136
Its a function of task manager. I think it uses the processor base frequency as the 0 point. I can load up one core on my laptop which has a base frequency of 2.0ghz, when there is only one loaded core it can ramp up to 2.9ghz...task manager show cpu utilization at 37% instead of the 25% id expect.
 

TennesseeTony

Elite Member
Aug 2, 2003
4,204
3,631
136
www.google.com
That would explain why I've never seen such behavior before, I lock in the processor speed during the overclock process and disallow speed-step/turbo schemes.

Not sure why the difference in output though between the lesser laden machine and fully laden machine. Maybe 4 days wasn't enough of a sampling period.

Ah well, I'll host a little race between them when I next switch projects to find out. ;)
 

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
Its a function of task manager. I think it uses the processor base frequency as the 0 point. I can load up one core on my laptop which has a base frequency of 2.0ghz, when there is only one loaded core it can ramp up to 2.9ghz...task manager show cpu utilization at 37% instead of the 25% id expect.

Task manager shows cpu usage as an avarage over time if you have two cores and run only one thread, but to the max, then task manager would show you 50% on each core because the thread gets thrown around the cores,so each core runs the thread at 100% for one turn and for 0% in the next, avereging at 50%

You run only one thread but windows runs threads too,it's 25% + whatever else is running.
 

waffleironhead

Diamond Member
Aug 10, 2005
6,917
429
136
Task manager shows cpu usage as an avarage over time if you have two cores and run only one thread, but to the max, then task manager would show you 50% on each core because the thread gets thrown around the cores,so each core runs the thread at 100% for one turn and for 0% in the next, avereging at 50%

You run only one thread but windows runs threads too,it's 25% + whatever else is running.

Ok, then can you explain how one task assigned to one core can use more than 25% cpu in 2C4T system?

cpuusage.PNG


cpucorepercent.PNG
 
Last edited:

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
Ok, then can you explain how one task assigned to one core can use more than 25% cpu in 2C4T system?
Run process hacker, it's free, double click on the prime exe and go to the threads tab, it might only run one prime thread but it might have secondary threads as well it will also show you what your threads are doing.
 

waffleironhead

Diamond Member
Aug 10, 2005
6,917
429
136
Process Hacker doesnt show any other associated process or threads being spawned. It does however show only 25% cpu usage for the program. This must just be some strange way that windows 8.1 shows cpu percentage use.
Thanks for the suggestion of using process hacker.
 

VirtualLarry

No Lifer
Aug 25, 2001
56,226
9,990
126
When my MeegoPad T02, with Atom Z3735F 1.33Ghz, throttles down to 0.49Ghz according to Task Manager, still, it only shows around 25% CPU usage, but RealTemp shows up to like 90% CPU usage.

So, IMHO, something a bit screwy with how Task Manager displays CPU utilization, at least on Win10 Pro 32-bit.
 

GLeeM

Elite Member
Apr 2, 2004
7,199
128
106
Tony are the Operating Systems the same on all computers? Win8.1 task manager, to me, seems different than Win7 task manager.
 

TennesseeTony

Elite Member
Aug 2, 2003
4,204
3,631
136
www.google.com
Tony are the Operating Systems the same on all computers? Win8.1 task manager, to me, seems different than Win7 task manager.

WinTen for the Xeon's, still waiting for the hex-core machine to get its upgrade, so it's still Win7.

Fairly important thing to leave out, so thanks for questioning that, and I'll update the first post with an edit.