Issue Activating HMM Feature on NVIDIA RTX A4500 with CUDA Toolkit 12.4 on Debian Bookworm

Hello NVIDIA Community,

I am reaching out for assistance with an issue I’m experiencing in activating the Heterogeneous Memory Management (HMM) feature on my system. Below are the specifics of my setup:

  • Operating System: Debian Bookworm, running Kernel version 6.6.13+bpo-amd64 from the backports.
  • GPU and Driver Details (from nvidia-smi -q):
    • Driver Version: 550.54.14
    • CUDA Version: 12.4
    • GPU Model: NVIDIA RTX A4500 (Ampere architecture)

Despite meeting what I believe are the necessary conditions for HMM activation (recent Linux Kernel, updated NVIDIA driver, and the latest CUDA Toolkit), the HMM feature does not seem to be enabled. Here’s a snippet from my NVSMI log for clarity:

==============NVSMI LOG==============

Timestamp                                 : Wed Mar  6 11:55:48 2024
Driver Version                            : 550.54.14
CUDA Version                              : 12.4

Attached GPUs                             : 1
GPU 00000000:01:00.0
    Product Name                          : NVIDIA RTX A4500
    Product Brand                         : NVIDIA RTX
    Product Architecture                  : Ampere
    Display Mode                          : Disabled
    Display Active                        : Disabled
    Persistence Mode                      : Enabled
    Addressing Mode                       : None

Given this information, I would like to know if there are any reasons why my GPU might not support HMM, or if there are additional system checks I could perform. Additionally, is there a way to force the activation of this feature, or any specific configurations I might be missing?

Any guidance or insights from the community would be greatly appreciated. Thank you for your time and help!