[Issue]: ERROR:root:Driver not initialized (amdgpu not found in modules) with GPU Radeon Instinct MI50

[camzilla1050]

Problem Description

Can't use GPU Radeon Instinct M150 32GB connected with pcie / Thunderbolt 3 on mini-desktop:

I followed official rocm documentation multiple times to install rocm with various versions. I try now with 6.2.4 version but no way, same issues,:

rocm-smi
ERROR:root:Driver not initialized (amdgpu not found in modules)

sudo dmesg
[ 3965.483045] amdgpu 0000:08:00.0: amdgpu: trn=2 ACK should not assert! wait again !
[ 3965.487042] amdgpu 0000:08:00.0: amdgpu: trn=2 ACK should not assert! wait again !
[ 3970.453973] xgpu_ai_mailbox_trans_msg: 1654 callbacks suppressed
[ 3970.453991] amdgpu 0000:08:00.0: amdgpu: trn=2 ACK should not assert! wait again !

Operating System

Ubuntu 24.04.4 LTS

CPU

Intel(R) Core(TM) i5-7260U CPU @ 2.20GHz 4 cores

GPU

[AMD/ATI] Vega 20 [Radeon Pro VII/Radeon Instinct MI50]

ROCm Version

6.2.4

ROCm Component

hiplibsdk,opencl

Steps to Reproduce

No response

(Optional for Linux users) Output of /opt/rocm/bin/rocminfo --support

ROCk module is NOT live, possibly no GPU devices

Additional Information

lshw
 *-display
        description: Display controller
        product: Vega 20 [Radeon Pro VII/Radeon Instinct MI50]
        vendor: Advanced Micro Devices, Inc. [AMD/ATI]
        physical id: 0
        bus info: pci@0000:08:00.0
        version: 01
        width: 64 bits
        clock: 33MHz
        capabilities: pm pciexpress msi bus_master cap_list rom
        configuration: driver=amdgpu latency=0
        resources: irq:17 memory:a0000000-a01fffff memory:c4100000-c417ffff memory:c4180000-c419ffff