[Bug] RTX 5060 Laptop – Dynamic Boost limited to 65 W instead of 75 W on Linux (Lenovo ThinkPad T1g Gen 8)

Đang mở
  • Có gì không ổn?
#1,469 0 bình luận 0 reaction 0 người được giao Xem trên GitHub

Maintainer thường phản hồi trong vòng 1 ngày

Chưa có ai nhận issue này.

Đánh giá

Độ khó
5/5
Thời gian dự kiến
Hơn một tuần
Mức phù hợp với người mới
25/100
Loại issue
Lỗi
Độ rõ ràng
Khá rõ ràng
Mức độ hoạt động
Sôi nổi
Công nghệ
c, linux
Lĩnh vực
backend, operating-systems

Hướng nghiên cứu

Start by comparing related issues #1393 and #1162, then investigate the reported assertion in platform_request_handler.c:855 and the nvidia-powerd Dynamic Boost path. The issue provides reproduction steps and power readings; done means identifying why Linux caps this laptop’s GPU at 65W and whether the driver or firmware path needs a change to reach 75W.

Do mô hình lập chỉ mục viết ra từ nội dung của issue.

Mô tả

bug
NVIDIA Open GPU Kernel Modules Version

NVIDIA 615.78.08

Please confirm this issue does not happen with the proprietary driver (of the same version). This issue tracker is only for bugs specific to the open kernel driver.
  • I confirm that this does not happen with the proprietary driver package.
Operating System and Version

Arch Linux

Kernel Release

Arch Linux

Please confirm you are running a stable release kernel (e.g. not a -rc). We do not accept bug reports for unreleased kernels.
  • I am running on a stable kernel release.
Hardware: GPU

GeForce RTX 5060 Laptop GPU

Describe the bug

System Information

Component | Details -- | -- Laptop | Lenovo ThinkPad T1g Gen 8 (21TD0002GE) GPU | NVIDIA GeForce RTX 5060 Laptop GPU (Blackwell) CPU | Intel Core Ultra 7 255H BIOS | N4EET24W (1.10) EC Firmware | 1.8 OS | Arch Linux Kernel | 7.2.11-arch1-1 NVIDIA Driver | 615.78.08 Kernel Module | NVIDIA Open Kernel Module nvidia-powerd | 2.0 (build 1) Power Supply | Lenovo Thunderbolt 5 Dock, 140W

During sustained GPU load, nvidia-smi reports:

  • Performance State: P0

  • SW Power Cap: Active

  • Current Power Limit: 65.00W

  • Default Power Limit: 60.00W

  • Max Power Limit: 75.00W

  • Average Power Draw: 64.93W

  • GPU Temperature: 67°C

  • HW Thermal Slowdown: Not Active

  • SW Thermal Slowdown: Not Active

This confirms that the GPU is being restricted by a software power cap rather than thermal throttling.

Kernel Error

The NVIDIA Open Kernel Module repeatedly logs:

NVRM: GPU0 nvAssertFailedNoLog: Assertion failed: PRH failed to update thermal limit! @ platform_request_handler.c:855

This suggests a potential issue involving the Platform Request Handler and the firmware interface, although a causal connection to the 65W limit has not been established.

Dynamic Boost and Firmware

The GPU reports:

Notebook Dynamic Boost: Supported

The NVIDIA power daemon starts successfully:

  • nvidia-powerd version:2.0 (build 1)

  • DBus Connection is established

All required CPUFreq interfaces exist for every CPU:

  • scaling_max_freq

  • base_frequency

  • cpuinfo_max_freq

CPU frequency scaling uses intel_pstate.

The NVIDIA D-Bus configuration is installed at /usr/share/dbus-1/system.d/nvidia-dbus.conf.

The NVIDIA ACPI device is present:

  • Device: NVDA0820:00

  • Path: \_SB_.NPCF

  • Status: 15

During sustained GPU load, the CPU scaling_max_freq values remain unchanged.

Troubleshooting Performed

The following actions did not resolve the issue:

  • Switching between balanced and performance platform profiles.

  • Restarting nvidia-powerd after changing platform profiles.

  • Disabling TLP completely and rebooting.

  • Verifying all required CPUFreq interfaces.

  • Running sustained GPU compute workloads using gpu-burn.

Attempting to set the limit manually with sudo nvidia-smi -pl 75 returns:

Changing power management limit is not supported in current scope for GPU: 00000000:01:00.0.

Expected Behavior

When sufficient CPU and system power budget is available, Dynamic Boost should be able to increase the GPU power limit from 60W to the supported maximum of 75W, as observed under Windows on the same hardware.

Actual Behavior

Dynamic Boost increases the GPU power limit by only 5W, from 60W to 65W.

The remaining 10W of available Dynamic Boost headroom cannot be utilized under Linux.

Questions

  1. Why does nvidia-powerd only increase the GPU power limit by 5W?

  2. Could PRH failed to update thermal limit! indicate a firmware communication problem?

  3. Is the 65W ceiling determined by the NVIDIA driver, nvidia-powerd, or the Lenovo SBIOS/NVPCF interface?

  4. Is any additional platform-specific configuration required to enable the full 75W?

Related Issues

A comparison with the proprietary NVIDIA kernel module is not applicable because the RTX 5060 Blackwell GPU requires NVIDIA Open Kernel Modules.

To Reproduce

Steps to Reproduce

  1. Boot Arch Linux on a Lenovo ThinkPad T1g Gen 8 (RTX 5060 Laptop GPU) with NVIDIA Open Kernel Module 615.78.08.

  2. Connect the laptop to AC power using the Lenovo 140W Thunderbolt 5 docking station.

  3. Set the platform profile to performance:

    echo performance | sudo tee /sys/firmware/acpi/platform_profile

  4. Enable NVIDIA Dynamic Boost:

    sudo systemctl enable --now nvidia-powerd.service

  5. Verify that the service is active:

    systemctl status nvidia-powerd

    The daemon starts successfully and establishes its D-Bus connection.

  6. Start a sustained CUDA GPU workload using gpu-burn:

    ./gpu_burn 120

  7. While the GPU workload is running, inspect power management information:

    nvidia-smi -q -d PERFORMANCE,POWER,TEMPERATURE

  8. Observe the following values:

    • Performance State: P0
    • SW Power Cap: Active
    • Current Power Limit: 65.00W
    • Default Power Limit: 60.00W
    • Max Power Limit: 75.00W
    • GPU utilization: approximately 100%
    • Power consumption: approximately 65W
    • No active thermal slowdown

  9. Stop NVIDIA Dynamic Boost:

    sudo systemctl stop nvidia-powerd

  10. Repeat the same GPU workload.

    The GPU power consumption drops to approximately 57–60W.

  11. Restart NVIDIA Dynamic Boost:

    sudo systemctl start nvidia-powerd

  12. Repeat the workload.

    The GPU power limit returns to exactly 65W, but does not increase to 75W.

  13. Check kernel messages:

    sudo journalctl -k -b | grep -iE 'PlatformRequestHandler|PRH|SBIOS|NVRM'

    The following assertion is observed:

    NVRM: GPU0 nvAssertFailedNoLog: Assertion failed: PRH failed to update thermal limit! @ platform_request_handler.c:855

Additional Verification

  • Repeat with TLP completely disabled: still 65W.
  • Switch between balanced and performance profiles: still 65W.
  • Restart nvidia-powerd after changing platform profiles: still 65W.
  • Verify that all CPUFreq interfaces exist: successful.
  • Test the same laptop under Windows: the GPU reaches 75W.

Result

The GPU is software-limited to 65W under Linux despite a reported maximum limit of 75W and successful Dynamic Boost initialization.

Under Windows, the same GPU reaches 75W.

Bug Incidence

Always

nvidia-bug-report.log.gz

nvidia-bug-report..log.gz

More Info

No response

Ngôn ngữ chính
C
Star
17.5k
Fork
1.9k
Không có pull request nào được merge trong 30 ngày

Chuẩn bị môi trường

Bắt đầu từ đâu

  1. Đọc hết issue, rồi đọc hướng dẫn đóng góp của dự án.
  2. Bình luận trên issue rằng bạn sẽ nhận — tránh hai người làm cùng một việc.
  3. Fork repository và làm thay đổi trên một nhánh.
  4. Mở pull request có tham chiếu số hiệu của issue.

Issue khác của NVIDIA/open-gpu-kernel-modules

Tất cả issue của NVIDIA/open-gpu-kernel-modules

Issue tương tự

Thêm issue về C

Nhận issue mới trong hộp thư của bạn

Bản tóm tắt ngắn những issue GitHub phù hợp với người mới.