Skip to content

GX650PY (Asus ROG Zephyrus Duo 16, 2023) - severe stutter, dxgkrnl.sys/nvlddmkm.sys DPC spikes up to 125ms - possibly related to ECLV/ACPI GPE issue #67

Description

@DarkHawkSouls

System

  • ASUS ROG Zephyrus Duo 16 (GX650PY, 2023)
  • AMD Ryzen 9 7945HX
  • NVIDIA GeForce RTX 4090 Laptop GPU
  • 32GB RAM, Windows 11 Home, version 25H2 (build 10.0.26200)
  • BIOS: GX650PY.324 (13-06-2025) - latest available on ASUS support site as of this post
  • Armoury Crate GPU Mode: Ultimate (dGPU only)

Symptom
Severe stutters under GPU load, especially in games like forza horizon 6, genshin impact, Neverness to everness, where winds meet,Microsoft Flight Simulator , VR (Meta Quest 3),etc. During each stutter: GPU clock, GPU power, GPU voltage, and CPU package power all dip simultaneously. Mouse and audio briefly freeze. FPS drops. No BSOD, no reboot. No WHEA/Display/Kernel-Power/PCIe errors in Event Viewer.

Issue started roughly 7-8 months ago after months of normal operation.

Diagnostics performed:

  • LatencyMon captures across multiple separate test sessions consistently show high execution times in both dxgkrnl.sys (DirectX Graphics Kernel) and nvlddmkm.sys (NVIDIA driver) - these are two separate captures from two different runs, not simultaneous readings from a single session, but the issue shows up in both drivers across repeated testing. One run captured dxgkrnl.sys spiking to 125.44ms; a separate run (on driver 610.74) captured nvlddmkm.sys at 8.66ms. I only have one screenshot attached per driver since they were caught on different test passes, but both are reproducible and represent the same underlying stutter symptom, not a one-off.
  • Windows Performance Analyzer (WPA): CPU Usage and DPC/ISR traces show clustered spikes in dxgkrnl.sys, nvlddmkm.sys, and (with a USB hub attached) USBXHCI.SYS, correlating with stutter windows.
  • Ruled out: USB hub as sole cause (removing it reduced USBXHCI.SYS DPC time from ~21ms to ~2.4ms, but stutter persisted), NVIDIA driver version (rolled back 610.74 -> 596.49 via DDU clean install, stutter persisted), PCIe Link State Power Management (already disabled), dual-GPU/ScreenPad+ compositing overhead (ScreenPad+ is hardwired to the AMD Radeon 610M iGPU; disabling it did not resolve the issue), Windows Memory Integrity/VBS (in progress).
  • 3DMark Time Spy/Steel Nomad and OCCT GPU/Power tests complete without errors; temps normal (~80-87C max).

Question / why I'm posting:
Found this repo via a Reddit/ASUS forum thread describing the ECLV/ACPI GPE self-rearm bug. My symptom profile (simultaneous GPU/CPU power dips, mouse/audio freeze, no OS-level errors) matches closely, but my traces show the stall surfacing in dxgkrnl.sys/nvlddmkm.sys rather than directly in ACPI.sys. Not sure if this is the same underlying mechanism (ACPI stall propagating up into the graphics scheduler) or a separate issue with a similar symptom - flagging in case it's useful data, or in case GX650PY needs to be added to the list of affected/tracked models.

Also confirmed: the hidden "ACPI L6D PME Handling" BIOS setting mentioned elsewhere in this repo is not exposed in the GX650PY BIOS menu (tried Ctrl+F1 hidden-menu unlock with no effect).

Image

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions