After replacing the GPU with MetaX GPU, the new GPU cannot retrieve information in HDM, and smi shows it as unavailable

2026-01-18 14:47:00 Published
  • 0 Followed
  • 0Collected ,9Browsed

Network Topology

This fault isMetaXGPU-C500X-64GB but most models of MetaX GPU are affected by this issue

Problem Description

Cross-testing


After replacing the GPU again, the new GPU remains unavailable under the system

Process Analysis

Testing and analysis revealed that the new GPU was not DOA. Investigation found that the firmware version of the GPU at factory shipment was 1.71.0, and our spare parts also had the same version.

The firmware version of the GPU at the customer site was 1.20.3.

The root cause of the failure was firmware version mismatch.

Solution

The GPU firmware versions need to be consistent. However, since the newly replaced GPU is unavailable, the firmware cannot be refreshed using the normal method. Use the following method to refresh the firmware
1. Remove all original GPUs, leaving only the spare GPU. At this point, the GPU will be properly identified and available. Then proceed with the normal firmware refresh
2. If the new GPU shows NotAvailable (please update vbios) when using the command mx-Sm1 -l, you can directly upgrade the firmware. If it only shows Not Available, you need to use the command metalink_train 0 to change the GPU to the aforementioned please update vbios state before upgrading the firmware.



Please rate this case:   
0 Comments

No Comments

Add Comments: