cancel
Showing results for 
Search instead for 
Did you mean: 

Disabling integrated GPU: Won't boot

MiniSpip
Level 8

Trying to "force" the iGPU off by changing in the BIOS: Advanced > SA Configuration > Primary Display > PCIE (instead of Auto).

When I do that: The PC reboots, then I get a very long black screen with nothing happening at all. After a long while, it reboots normally but... if I then enter the BIOS, I'll see that it's back to Auto.

The graphics card (RTX4070) is working fine and connected via DisplayPort to the monitor.

Setup: Asus Z790 Hero, with BIOS v1501, 13900K.

PS: Reason why I'm trying to do that is because I have very annoying and unexplainable frame drops in all 3D games and I'd like to try a completely clean install of nvidia drivers, after booting with the iGPU fully disabled. As I suspect there might be something interfering, somehow, somewhere...

1,709 Views
14 REPLIES 14

Jiaszzz_ROG
Customer Service Agent

Hello, @MiniSpip 

Regarding the CPU's integrated graphics, typically, if the motherboard recognizes the installed GPU, it will prioritize the discrete GPU and not use the integrated graphics.
If you have a specific requirement to disable the integrated graphics, we suggest trying to disable it by right-clicking on the "13900K" entry in Device Manager under Display Adapters and then restarting the system.
Issues such as overheating, outdated drivers, insufficient memory, network anomalies, or power supply problems could potentially cause frame drops during gaming.
Could you specify which PCIe x16 slot on the ROG MAXIMUS Z790 HERO your RTX4070 is installed in?
Additionally, have you used any adapters or extension cables, or is any other shared channel occupied?

Thank you.

Due to the ongoing regional holidays, my responses may be slightly delayed.
Thank you in advance for your patience.‌‌‌‌ 🙂

Many thanks for your detailed answer.

Thing is: The integrated graphics does not appear at all in Device Manager, even with "Show hidden devices" and after a rescan. But: The CPU GT Cores are seen by sensors (HWInfo64, AIDA64) and permanently consume ~10W.

The discrete GPU is installed in the first PCIe slot, nothing is installed in the M2_1 socket. My NVMe drives are in M2_2 and M2_3. No riser cable or adapter for the power-in. Nothing is installed in the second and bottom PCIe slots.

As additional information regarding my GPU troubleshooting:

  • Overheating: Ruled out.  GPU (incl. hotspot) and CPU are within very acceptable range, both well below 80°C in heavy games. VRM, PCH temps are fine too, same for SSD and RAM.
  • Drivers: Tried DDU multiple times and different driver versions.
  • Network: The problem occurs in offline games, the ethernet controller isn't doing much during those test gaming sessions.
  • PSU: Have tried so many things that I'm started to think it could be power related. But that is a lot harder to confirm.
  • Others: HAGS on / off in Windows, setting GPU interrupt mode to MSI...
  • BIOS: Tried Resizebar on / off, disabling E-Cores, "legacy game mode"...

It's not really stuttering in games: Randomly, the GPU will stop rendering for 1-2 seconds, while all the rest goes on: Sound for instance is unaffected. In that timeframe, all GPU metrics drop very low (Core Load, mem clock, power draw...). LatencyMon reveals no issues.

Jiaszzz_ROG
Customer Service Agent

Hello, @MiniSpip 

Regarding the frame drop issue during gaming, aside from considering overheating factors, we would suggest verifying the stability and capacity of your memory, as well as the health of the hard drive.
Therefore, we recommend cross-testing hardware components, including the CPU, GPU, PSU, RAM, and system disk.
Concerning software settings, we suggest checking power management settings under Windows, NVIDIA Control Panel configurations, verifying graphics card drivers and VBIOS versions, and adjusting in-game visual settings (considering the need for enabling V-Sync).
Also, it's advisable to assess the test results across different games.

Thank you.

Due to the ongoing regional holidays, my responses may be slightly delayed.
Thank you in advance for your patience.‌‌‌‌ 🙂

Hello @Jiaszzz_ROG ,

Thank you for your reply and follow-up on this.

For the frame drop issue:

  • It really seems to be a CPU Power Management problem.
  • I have ruled out hardware components, via multipe tests, both stability (long runs) and stress; using: Intel processor diagnostic tool, memtest86 (passes completely), Samsung SSD tests, Geekbench, CinebenchR23 etc. Everything passes fine.

Back to framedrop and power management: Yes it occurs across all games, be they DX11, DX12, GPU intensive or not, CPU intensive or not. Forcing "High performance" power plan in Windows drastically reduces but does not eliminate the issue.

It seems the issue stems from the fact that, even with a high power plan where all CPU Cores should operate at 100% all the time, one or two cores will briefly "park" and drop to their lowest frequency (800Mhz down from 5500) ! This is easily observable at the desktop in idle.

Core lowers frequency for a second, at max power planCore lowers frequency for a second, at max power plan

While gaming, this behaviour is probably what causes a sudden spike in frametime, and hence the drop.

GPU stable. CPU spike in frametime causes framedropGPU stable. CPU spike in frametime causes framedrop

 

Jiaszzz_ROG
Customer Service Agent

Hello, @MiniSpip 

Based on your description, may I inquire if your CPU is overclocked in the BIOS?
We recommend trying to clear CMOS and rechecking under default BIOS settings.
If the problem persists, please help to confirm the question below.
- games that experience frame drops can be influenced by various factors, such as playing a specific game for a few minutes and the CPU and GPU temperatures at this point, as well as details about the concurrently running programs.
- the images of the setting in BIOS, or whether it is the default
- the brand, model name, and part number of the DRAM currently installed
- the brand and model name of the SSD/HDD and PSU currently installed
- are there any extensions or adapters used to connect any devices, such as GPU or SSD?
- the OS version and OS build
- a comprehensive video recording of the problem, along with capturing third-party software verification results, would be beneficial for further analysis.

Thank you.

Due to the ongoing regional holidays, my responses may be slightly delayed.
Thank you in advance for your patience.‌‌‌‌ 🙂

MiniSpip
Level 8

Thank you @Jiaszzz_ROG  for the follow-up. Please find some answers / comments below:

  • Games: AC Odyssey CPU at 65° and GPU at 48°, AC Valhalla CPU at 54° GPU at ~50°, Cyberpunk77 CPU at 55° GPU at 70°. Please note that the framedrops occurs in games that are neither CPU intensive nor GPU intensive. The random CPU downlocking happens at idle at the desktop with nothing running.
  • Concurrently running programs: at minimum Norton 360*. Otherwise quite a few. The issue also occurs in a completely fresh Windows 10 install and  in a completely new, barebones Windows 11 install.
    *Tested with a selective boot (msconfig), disablig all services except Norton.
  • BIOS: Will provide images later. Settings are on "optimized defaults" except XMP1 is enabled, ASPM power management is OFF otherwise it causes WHEA errors, MCE is set to "disabled enforce all limits".
    Tried different BIOS settings, disabling E-Cores, resizebar, virtualization... Tried different BIOS versions: Was on 1202, and upgraded successively to 1501.
  • DRAM: G.Skill Trident Z5. Highest temp in gaming 52°. Memtest86 passed (all 4 runs, 3h). Framedrop occurs even when RAM is not set to XMP and running at 4800MHz instead of 6000.
  • SSD: Samsung 990Pro, installed with heatsink. Temp in gaming well below 50°. Other OS is on Samsung EVO870.
  • PSU: be quiet! Dark Power 13 850W. Tried different cables between PSU and GPU.
  • Extensions / Adapters: None. Only one PCIE slot is used directly for the GPU.
  • OS: Win10 Pro 22H2, build 19045.3803
  • Video recording: Will try to provide one. Please advise on the best way to share the resulting file ?

I could also provide a bunch of screenshots with HW monitors that show the issue occur. If those would be useful as well, please tell on the best way to share all the screenshots.

Jiaszzz_ROG
Customer Service Agent

Hello, @MiniSpip 

Thank you for your detailed explanation.
Based on your description, all hardware configurations have passed various third-party software tests, temperatures are maintained within the 80-degree Celsius range, and the game type is single-player, so there shouldn't be any network-related issues.
Adjusting BIOS settings (default or XMP I/ASPM power management disabled/MCE set to "disabled enforce all limits"/disabling E-Cores, resizebar, virtualization) hasn't resolved the problem.
Modifying Windows power management only minimizes the frequency of occurrences.
You mentioned the possibility of the CPU peak causing throttling and frequency reduction, which occurs randomly, even during idle, leading to frame drops in games.
Could I be overlooking anything?

Additionally, regarding BIOS settings and the video demonstrating the frame drops, we recommend uploading them to the cloud, setting permissions to public, and then providing us with the link. 
Please also share the part number and the slot position of the G.Skill Trident Z5, as well as the driver version and VBIOS version of the GPU.
And, may I ask if you are using a UPS? If possible, we recommend cross-testing with different CPU and PSU.

Thank you.

Due to the ongoing regional holidays, my responses may be slightly delayed.
Thank you in advance for your patience.‌‌‌‌ 🙂

Jiaszzz_ROG
Customer Service Agent

Hello, @MiniSpip 

After several days of waiting, we have consulted and discussed with the relevant department.
If the dropped frames you are experiencing do not result in issues such as stuttering or lag during gameplay, a simple fluctuation in frame rates falls within the normal range.
However, if there are any impacts beyond frame rate changes, please provide the numerical variations in frame rates and a video of the situation at that moment.

Thank you.

Due to the ongoing regional holidays, my responses may be slightly delayed.
Thank you in advance for your patience.‌‌‌‌ 🙂

Has your issue been resolved? On the contrary, once I enable the IGPU, my motherboard lights up with a yellow light and then reboots.