OMEN command center causing core 0 temp spikes on laptop

While playing games I noticed I would get constant periodic stuttering due to my CPU reaching my throttlestop alarm temp and switching profiles to lower clock speeds.

After a little investigation I discovered that only one of my cores, core 0, was spiking into the upper 90s while the rest of my cores were holding steady in the 80s. (i set the affinity for the game i was playing to 2-15 to rule it out).

Here you can see core 0 on top and core 7 on bottom. they are within a few degrees of each other until there is a temperature spike on core 0 every 30 seconds or so.


This tells me that there is a process running on core 0 exclusively that is loading the CPU every 30 seconds and causing the temps to spike and thermal throttling the machine.

I did some more digging with windows ADK and discovered that OmenCap.exe was the culprit. Its loading my CPU on core 0 to nearly 100% every 30 seconds.


OmenCap.exe is loaded by OMEN command center to display CPU and GPU temps. You can disable it by going into your services and disabling "HP OMEN HSA service". Command center wont be able to display your temps but the fan curve will still work properly (i think the fans get their temp info from the bios).

After disabling this service all my cores stay within a few degrees of each other and I have no more thermal throttling.

 

WHOA!

 

Hmm. What happens if you reenable the service?

 

I take it you had the same problem?

Spikes come back.

I also noticed that display fusion and "windows 10 firewall control" were causing similar, but smaller, spikes. My theory is that when they all happened to fire at once I would get the highest temp spikes causing me to throttle.

Nothing i did could get rid of the display fusion spikes so i uninstalled it but windows 10 firewall control has a "low cpu" option.

Also, if you use msi afterburner it uses less cpu cycles if you disable low level monitoring.

 

I probably do however I don't have the level of knowledge you do here. I just finished a thermal repaste to Thermal Grizzly Kryonaut and HOLY MOLY! I ran 3DMARK Time spy and gained almost 1000 points just stock everything but the repaste. 800+ points in Cinabench just now. NEVER went past 78*c on the CPU!!!!! She's already running better. Now to run the Division 2 and test some more. I'm speechless at the difference with the good thermal paste. I appreciate the tips and the inspiration!

Also I'm not seeing any spikes at all now after the repaste through TIme Spy, Cinabench, and FIrst Strike so far.

 

truly thank you for this as now im able to open command center much quicker , may I ask u a question it would really help me in my case ? what kind of cpu do u have is it the i9 9880h ? if its can u check whats the max speed u can get from the cpu as it shown in other websites it should boost to all cores at 4.1ghz , in my case my max is 3.8ghz ( you will need to check via hwinfo to know the max or use throttlestop during cinebench or whatever benchmark ) its good and everything but I would like to know if u also suffer from this issue or its only me ?

 

guys did you see this? just updated and its now called omen gaming hub and has an undervolting tool! It's causing restarts I'm sure because of Throttlestop...

 

HP has been absolutely killing it lately with their hardware, software, and listening to customer feedback. Props to them.

 

This is great stuff, thanks. I considered replicating it but your skill level is much higher than mine; started reading up on xperf and getting it in to WPA and rapidly lost interest in the project. Since you already did the work, I've done some experiments where terminating the omen command center background doesn't get rid of these blips. However, I had a pretty serious problem without Omencap--after a clean reboot, ended it and it benchmarked great under Throttlestop control. Temps were fine but around the time where turbo timer would kick in, it actually dropped the fan to zero and processor watts down to 45, behavior I haven't seen before. I assume this is Intel behavior and not HP, with the HP stuff turned off. There are several layers of control to this machine I don't understand, and Throttlestop is able to do most of it to my satisfaction with the HP layer enabled. I'm going to try troubleshooting this some more and report back. But I completely agree those 30 second interval blips are a real nuisance. There's no affinity attached to it, is there another method to limit its CPU usage?

 

Okay I've played with this some more with Omen HSA service disabled. The temps still work in OGH. It's difficult to disentangle as there are still some 30 second spikes that are associated with Chrome, but they go away over time (and if it's closed).
The problem I've been having since all these major updates is a dropping of PL1 to 45W. After a reboot it's not there. I installed Death Stranding last night and Steam promptly crapped down to 3Ghz and took its sweet time. I haven't figured out the trigger for it yet, today it was after two sleep/wake cycles. Touching OGH or Throttlestop doesn't change it, it can't be gotten rid of without rebooting. I'm afraid this is some new Intel microcode or the F35 BIOS itself. If this keeps up I'll be tempted to go back, but F33 was a very quick update. F35 took half an hour and seemed to flash dozens of things on this computer, I'm not sure flashing F33 will revert all that. I'll be severely disappointed if going forward I have to reboot the computer every time I want full performance. This could also be a fluke in OGH not staying on top of the performance mode selection.
Frustrated. I feel like I'm relearning how this machine works.