Throttling Issues

[Vox_Maris]

First of all, I use HP 17-cb0005nt laptop (Windows 10) with Intel i7-9750H and RTX 2070 and 16 GB RAM.

Since last year, I have been having problems with my computer. It started with stuttering every now and then, and advanced to screen glitches.

I eventually figured out that this was caused by my CPU thanks to some people’s help on a discord PC server.

Watching it while playing games or using high CPU usage software, I saw that my CPU would first use up to 75-80 W (Called Turbo Mode) as a 45 W CPU. After roughly 30 seconds, it would then locked at 45 W (Power Throttling). The power throttle, especially when it went between 75 and 45 randomly, would cause my entire computer struggle, even if only half my resources were used by a program.

My Bios did not have an easy way to change any of the power settings but OMEN by itself has a place to undervolt. I undervolted, it did nothing. Changed every single power setting, reformatted my drives, went between Windows 10 and 11, changed driver versions a few times for both my CPU and GPU, at one point found that my SSD was borked but changing it did nothing, I repasted it with a higher quality paste…

Eventually I ended up using ThrottleStop to completely remove the power limitation- which stopped all of my power throttling issues! Yay!

Then my computer started heating up like crazy and would thermal throttle- which caused me CPU issues once again. I removed the thermal limits just to test. No problem for 90% of the time. It runs smoothly, passes stress tests, but then my temperatures hit 100 C while using high usage programs and the laptop crashes to protect itself. (for some reason stress testing did not bring it to crash)

My fans are set to maximum, and thermal paste is fresh as I said.

My CPU seems to insist on taking more wattage than it can actually handle, which causes it to essentially accelerate, hard press the brakes, then accelerate again, causing the whole car to shake. When I cut the breaks- obviously I crash into a wall.

I did do tests with all my other computer parts with both third party software and HP’s own, CPU seems to be the only one having issues currently.

[Malnutrition]

Below is the order that usually gives the best “area under the curve” (consistent performance without hitting 100 °C or stuttering).

[ol]
[li]Undervolt as far as silicon allows (core + cache together).[/li]Typical 9th-gen sweet spot: −80 mV to −125 mV.
Do it in ThrottleStop (FIVR tab), not OMEN Hub—the latter sometimes reverts after sleep. Monitor stability with TS bench and CB R23 loop. A solid undervolt often trims 8-10 °C.
[li]Lock in moderate power limits that the cooler can actually hold.[/li]
[ul]
[li]TPL ➜ set PL1 = 50 W, PL2 = 60 W, Turbo Time = 28 s.[/li][li]Check “Lock” and “MMIO Lock”. That keeps MSR/MMIO values fixed even when the EC pokes them; it won’t beat the EC forever, but it stops the rapid oscillation most users notice. [/li][/ul]
[li]Cap the all-core turbo ratio 1–2 bins lower.[/li]In the FIVR window set Max Turbo Ratio for 6-core = 38× (instead of 40×). The chip now needs < 1.15 V for full load, shaving another ~5 °C yet costing only ~3 % performance.
[li]Maximise cooling the cheap way.[/li]
[ul]
[li]Vacuum or blow out the fins—HP Omen 17 intakes clog quickly.[/li][li]Raise the rear 1 cm (even a book helps) or use a cooling pad that blows into the bottom grill; expect −3 °C.[/li][li]Replace 0.5 mm stock thermal pads on the VRM/GPU side with 1 W/m K pads if they’ve squashed flat. It stops heat soaking into the CPU heat pipe.[/li][li]Liquid-metal repaste adds another 8-10 °C headroom but is risky—only if you’re comfortable insulating the die perimeter.[/li][/ul]
[li]Co-operate with the GPU.[/li]Games are where the 9750H/RTX 2070 share a 230 W adapter. Undervolt or set a 150-W power target in MSI Afterburner; every watt you pull from the GPU becomes thermal budget the CPU can use. (On some Omens the EC eases its 45 W clamp when the dGPU isn’t at full tilt.)
[li]Validate.[/li]
[ul]
[li]HWInfo64 ➜ Sensors only → log 20 min of your typical workload.[/li][li]Look for: CPU Package Power flat at ~48-52 W, Core Max ≤ 92 °C, no “PL1/PL2” or “Therm” flags flickering.[/li][li]If temps still spike, drop PL1 another 3 W or lower the 6-core ratio one more bin.[/li][/ul]
[/ol]

[Vox_Maris]

Set it to -100 mV offset (It was -80 earlier). Set the power limits as instructed (50 and 60)

Ratios I had set to 40 while default is 43. I set them to 38 now.

Already have a cooler.

Sadly Afterburner is not available for my GPU (blocked by manufacturer)

I’m back to experiencing power throttle and weirdly enough it is throttling itself to 35-40 rather than 45 as it did before. Temperature is 80-85 C consistent and goes up to 97.

While there is power throttling, I am not experiencing the screen glitches from before (They were not consistently appearing before I removed power limits a while ago, will have to see after a week or two of usage)

[Vox_Maris]

Yeah I talked too soon it seems. After an hour of usage it started going bad.

[Malnutrition]

Please download MiniToolBox and save it to your desktop.

Run the program by right clicking on it and selecting Run as administrator.
When the program opens select the following boxes:

Flush DNS
Report IE Proxy Settings
Reset IE Proxy Settings
Report FF Proxy Settings
Reset FF Proxy Settings
List content of Hosts
List IP Configuration
List Winsock Entries
List last 10 Event Viewer Errors
List Installed Programs
List Devices (Only Problems)
List Users, Partitions and Memory size

Please post the log in your next reply

---

---

Speccy Scan.

[ul]
[li]Please go here and download Speccy.[/li][li]Install and run the program.[/li][li]Upon Completion:[/li][li]Hit File[/li][li]Publish Snap Shot[/li][li]A link will appear, post that link.[/li][/ul]

---

---

Download GetSystemInfo by Kaspersky.
In GSI, click Start. (takes about 10mins)
It makes a ZIP file on your desktop, drag that to their GSI Parser site.
Once analysed, paste the newly created URL into your post.

[Vox_Maris]

An error occurred in the secure channel support on Speccy. Sending the file as is.

https://www.getsysteminfo.com/report/0f6f64285e2492f2f5f37b50b89e7e40

[Malnutrition]

i will be back to this on sunday i have some things to take care of today.

[Malnutrition]

Unsigned overlay DLLs injected into Discord Code-Integrity blocked ow-graphics-vulkan.dll & OWClient.dll from Overwolf Can cause sudden GPU/CPU stalls and window flicker (the “screen glitches” you mentioned). Uninstall or update Overwolf; if you need it for game overlays, switch to the current v0.278+ driver-signed build.

MSI Afterburner side-by-side errors (VC90 MFC missing) Event ID 33 SideBySide for MSIAfterburner.exe
Afterburner can’t start properly, so RTSS may keep reopening—extra CPU wake-ups every few seconds. Install the “Visual C++ 2008 SP1 (x86) redistributable” (the one containing mfc90.dll).

Repeated NetBT 4311 failures NetBT Event 4311 on every boot and resume NetBIOS driver can’t bind → extra 3-5 s delay negotiating network shares; sometimes spikes CPU when it retries. 1) Update the Realtek GbE driver, 2) Disable NetBIOS over TCP/IP on the ProtonVPN (WireGuard) and any “Microsoft Wi-Fi Direct” adapters, 3) Reboot and clear the event log.

[Vox_Maris]

Already had Visual C++ 2008 SP1 (x86) Redistributable, uninstalled and installed again.
Had the latest Realtek GbE, did the same.
Removed Overwolf, don’t use it that much anyways.

I am not sure if it improved anything with my CPU but disabling NetBIOS seems to have made my connection better. I just used to think it was my provider being bad.

Sadly did not have any improvement in the power or thermal department out of those.

I think I may be looking into a more hardware related issue. Had taken to a repair before but they couldn’t troubleshoot it properly and just gave it back to me untouched.

Any guidance on what I can give to the repair crews to make it easier?

[Malnutrition]

sorry for the delay. @phillpower2 can you please take a look here

[phillpower2]

I eventually figured out that this was caused by my CPU thanks to some people’s help on a discord PC server.

First of all, I use HP 17-cb0005nt laptop

Be happy to take a look at some point if still required but I[ICODE]m under the cosh fulfilling contractual work at the moment I[/ICODE]m afraid, sorry

[Malnutrition]

Flash latest BIOS and EC firmware

Lock sensible power limits directly in BIOS if possible (some revs expose “Long Duration PWR Limit”). If not, staff should set PL1 = 50 W, PL2 = 60 W, Tau = 28 s in ThrottleStop and tick MSR Lock + MMIO Lock so the EC can’t instantly re-write them.

nvidia-smi -pl 150

A 150 W cap on the RTX 2070 leaves ~20 W more AC-adapter and thermal headroom for the CPU without visible FPS loss on 1080p panels. (If Afterburner can be unlocked, the same limit can be set in the curve editor.)

[Vox_Maris]

I didn’t get notifications for some reason. Thanks for all the help. Gave it to repair shop. Let’s see if they can figure it out.

[Malnutrition]

Alright come back and let us know.

[phillpower2]

The OP revisited the forum but left no reply so to avoid us having to ask for an update every couple of days we will close this thread until the OP has any update and/or the time available to work towards resolving the issue, PM a member of staff when you are ready and the thread can be reopened.