High ACPI.sys DPC Latency (>34,000µs) on new ASUS ROG Strix Scar 16 (2025) - Causing 0x7E NVIDIA BSO

[Slytha]

Hello Intel Team & Community,

I am performing a deep-dive analysis of a critical BSOD on a new ASUS ROG Strix Scar 16 (2025) laptop, and all evidence points to a massive DPC latency stall in ACPI.sys. I am posting here to see if this is a known issue with the platform's chipset/firmware interaction and to request an investigation from the Intel side.

System Configuration

• Laptop: ASUS ROG Strix Scar 16 (2025)
• CPU: Intel Ultra 9 275HX
• GPU: NVIDIA RTX 5090 Laptop GPU
• BIOS: ASUS v327
• OS: Windows 11 Pro (Fresh Install)
• External Monitor: 4K 240Hz (via HDMI 2.1, using DSC)

The Problem: BSOD & Latency

The system is experiencing SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (7e) BSODs, pointing to nvlddmkm.sys (NVIDIA driver).

The minidump shows the crash is an Integer divide-by-zero (c0000094). This is happening because a driver function (div eax, r12d) executed while a required register (r12d) was still 0. This indicates a severe race condition.

The root cause of this race condition was found using LatencyMon:

• ACPI.sys (ACPI Driver for NT)
• DPC Routine Execution Time: ~34,665 µs

This 34,000µs stall in the ACPI driver is the core problem. It is freezing power-state transitions long enough to cause other kernel-mode drivers (like nvlddmkm.sys and dxgkrnl.sys) to fail.

Analysis & Question for Intel

This ACPI.sys latency is directly related to the system's firmware (BIOS) and its interaction with the platform's power management. The crashes are most frequent when:

1. Switching between iGPU and dGPU.
2. Driving a high-bandwidth external 4K 240Hz monitor that uses Display Stream Compression (DSC).

My hypothesis is that the ASUS BIOS (v327) is improperly handling an ACPI power-state request (perhaps related to the P-cores/E-cores, the Intel chipset, or the Thunderbolt/Display controller) when the dGPU is active, especially with DSC.

My question is: Is the Intel team aware of or investigating any ACPI.sys latency issues on this new platform, especially in conjunction with ASUS firmware? Could this be related to Intel Dynamic Tuning Technology (DPTF), chipset drivers, or the platform's handling of GPU power states via ACPI calls?

Any insights you can provide or any data I can collect to help you investigate would be appreciated. This appears to be a platform-level firmware bug that is causing downstream driver crashes.

As I am an IT Architect myself, I have tried everything to mitigate or resolve the issue.

A post about my issue: ROG Strix Scar 16 / RTX 5090 – nvlddmkm.sys Crashe... - Republic of Gamers Forum - 1120221

Might be related to a global issue, detailed post & confirmed by Asus: Zephkek/Asus-ROG-Aml-Deep-Dive: A deep dive into the ACPI.sys DPC latency problems on Asus ROG laptops

Alot of other posts about these issues can be found

Hopefully you can read the posts and try and make sense of the issue.

If you need anything i'll try to help.

Kind regards,

KS

[Slytha]

Hi Felipe,

Any progress on the Display External logs issue in Windows event viewer?

Regards

[Felipe_Intel]

Hello Kevin,

Our development team is still working on this. Rest assure once a fix is published for this, I'll let you know. In the meantime, we'd appreciate your patience.

[Slytha]

Hello Felipe,

Thanks for the update. I do have an update regarding the other issues (black screens and stuttering, just for Intels knowledge)

https://rog-forum.asus.com/t5/gaming-notebooks/asus-rog-strix-scar-2025-g18-5090-external-monitor-issues/m-p/1131450#M12847

https://github.com/Zephkek/Asus-ROG-Aml-Deep-Dive/issues/47#issuecomment-3682647483

https://rog-forum.asus.com/t5/gaming-notebooks/randomly-losing-all-usb-devices/td-p/1091235/page/22

It seems to be having something to do with processor's ability to enter deep power-saving states (C-States).

Support has previously stated that engineering "cannot reproduce" the issue. I have most likely identified the specific root cause and a reproducible solution. The instability is most likely caused by a defect in the ACPI implementation regarding Processor Idle (C-State) transitions and maybe USB Selective Suspend (not sure)

• Step 1 (ACPI C-States): Using Power Settings Explorer, I exposed the "Processor idle demote/promote threshold" options. Setting these to 100% prevents the CPU from entering deep C-States.
• Result: Latency spikes disappear (confirmed in latencymon), audio stutter stops, and black screens cease.
• Step 2 (USB Power): Disabling "Allow the computer to turn off this device" for all USB Controllers in Device Manager. (optional)
• Result: Most likely prevents a USB bus collapse that might previously accompanied the black screens or some kind.

Additional Firmware Defects (dGPU Mode) I must also report that the dGPU (Ultimate) Mode is currently functionally broken on this BIOS.

• Visual Artifacts: White static/snow appears on the internal display (initialization failure).
• Performance: GPU usage drops to 60-80% causing severe FPS loss compared to Optimus mode.

Please escalate this to the needed teams. The solution is not to force users to disable power-saving features permanently. The solution is for ASUS to fix the ACPI C-State and USB Power tables and the dGPU mode in the next BIOS update so the hardware can transition power states without crashing the platform.

This is all my eyes of the solution.

"Disclaimer: I am a user, not a developer or ASUS engineer. These findings are based on my personal testing and troubleshooting to resolve stability issues on my own unit. I am not responsible for any potential damage or issues that may arise from applying these power modifications to your own hardware; proceed at your own risk." (As i do think it is safe for these changes to make, as long as you monitor your temps at start).

Regards,

Kevin

[Felipe_Intel]

Hello @Slytha, 
Our latest driver 32.0.101.8425 includes a fix for the Error 10 reports. 
Could you please check if this driver reduces the reports of this error in your system? Looking forward for your results. 

[Slytha]

Hello Felipe,

I have installed the driver, cleared the logs and keep you posted on the results.

I'll do some work for a couple of hours and get back to you.

[Slytha]

Hello Felipe,

Yesterday I installed the driver and after half a day of work I got 12 Error 10 reports, so the reports got reduced.

Today I noticed Windows Update reinstalled an older driver, so I did the following (maybe usefull for others):

- Downloaded DDU from official site: https://www.wagnardsoft.com/content/Download-Display-Driver-Uninstaller-DDU-18140

- Download the latest driver: https://www.intel.com/content/www/us/en/download/785597/intel-arc-graphics-windows.html

- Via Group policy disabled driver updates via Win update: Computer Configuration > Administrative Templates > System > Device Installation > Device Installation Restrictions

- Disabled: The "Advanced Settings" Step -> Press Win + S and type Advanced System Settings, then select View advanced system settings. In the small window that opens, go to the Hardware tab. Click the Device Installation Settings button. Select No (your device might not work as expected). Click Save Changes.

- Disconnect Internet (Wifi and ethernet) 

- Reboot into safe mode via (Hold Shift while clicking Restart > Troubleshoot > Advanced > Startup Settings > Restart > Press 4 or 5).

- Ran DDU: Select GPU on the right, then select Intel. Go to Options and make sure the very bottom checkbox ("Prevent download of drivers from 'Windows Update'...") is checked. Click Clean and Restart.

- Once back in normal Windows (still offline), run the Intel installer you downloaded and execute a clean installation and only install the "Intel Graphics Driver", the rest is optional if you have a better Nvidia gpu.

- Reboot the OS

- Recheck GPO and advanced settings are still active to prevent windows update driver installation

- Go to Device Manager and verify the "Driver Version" matches the one you just installed. Once you are sure it's correct, reconnect your internet. Because you set the Group Policy and used DDU's "Prevent Windows Update" toggle, Windows should now leave your driver alone.

- Turn the Internet back on, let windows update search for updates.

Can you let me know when the official WHQL drivers are available?

Thanks!

Regards,

Slytha

[Felipe_Intel]

Hello @Slytha, thanks for the detailed report.

The roll back made by Windows Update is a known situation related to the Operative System and OEM drivers. One way to partially overcome this is to let WU install the driver automatically and then install our latest driver on top of it. No clean installation, no DDU. Just install it. This way, unless a major Windows Update happens (e.g., 24H2 to 25H2) or a new OEM driver is published (and forced through WU), the system should keep that last version you manually installed. So, if you want to try it or if it rolls back again, try: DDU, then let Windows automatically install a driver, then manually install the latest Intel's Download Center driver.

About the WHQL drivers, I can't provide an specific date for when non-released drivers will be available. If you have Intel Driver and Support Assistant installed, it will let you know when a new driver is available, and you can select to be notified/install only WHQL drivers. Intel Graphics Software also has this notifications. 

As explained in this article, the "WHQL" part of the driver is a certification provided by Microsoft and once we provide the driver to them, it takes a while for it to be certified. This is why you may see some non-WHQL drivers be released between WHQLs. Nevertheless, a non-WHQL driver only lacks that certification, meaning the code is tested and validated internally (Intel side) the same way as WHQLs. Rest assure you will install a driver of the same quality no matter if it is WHQL or not. 

Finally, as for the Error 10, I'm glad to hear you are getting just a handful of notifications, instead of the bulk you were having before. This was happening because of how the OS and the driver were handling the logging. This shouldn't be affecting the performance of your system. 

If you are ok with it, we can close this ticket. Let me know your thoughts. 

ACPI.sys DPC Latency issue is being handled by Asus directly and they need to provide the fix for it.

[Slytha]

Hello Felipe,

Sounds logic indeed, I usually do clean installs to be sure, definitely when testing bugfix drivers.

For me this ticket can be closed, I have noticed people still are having tons of errors with the newest driver though: 

https://community.intel.com/t5/Graphics/Intel-Gfx-Display-External-error/m-p/1734780/emcs_t/S2h8ZW1haWx8dG9waWNfc3Vic2NyaXB0aW9ufE1LVjBJRlBFMThKOTEwfDE3MzQ3ODB8U1VCU0NSSVBUSU9OU3xoSw#M148735

Just so you know, it seems that not with everyone (hw specific) the logging is fixed.

Regards and thank you for the resolution.

Slytha

[Felipe_Intel]

Great, thanks for letting me know. That case you linked is being handled by one of my teammates so they are aware of the situation. If you see this problem coming back in the future, please create a new ticket since this one will not be monitored anymore. 

We will proceed to close this one.