IBM System x3650 M5 Installation And Service Manual page 967

409-900-000 NVIDIA User Diagnostic Test Failed
Explanation: NVIDIA User Diagnostic Test Failed.
Severity: Event
Serviceable: Yes
Recoverable: No
Automatically notify Support: No
User response: Complete the following steps:
1. Verify that the GPU is seated in the PCIe slot correctly by reseating the GPU. Then power cycle the system.
2. Verify that the power connectors to the GPU are connected firmly. Then power cycle the system.
3. Run nvidia-smi -q In some cases this will report a poorly connected power cable.
4. Rerun the diagnostics, using the same GPU, on system that is known to be working. A variety of system issues
can cause diagnostic failure.
5. If the problem remains, contact your IBM technical-support representative.
Related Links:
IBM Support Website
v
Latest level of DSA
v
Latest level of BMC/IMM
v
409-903-000 Nvidia::DiagnosticServiceProvider::Bandwidth Test Failed
Explanation: Nvidia GPU Bandwidth Test Failed.
Severity: Error
Serviceable: Yes
Recoverable: No
Automatically notify Support: No
User response: Complete the following steps:
1. Verify that the GPU is seated in the PCIe slot correctly by reseating the GPU. Then power cycle the system.
2. Verify that the power connectors to the GPU are connected firmly. Then power cycle the system.
3. Run nvidia-smi -q In some cases this will report a poorly connected power cable.
4. Rerun the diagnostics, using the same GPU, on system that is known to be working. A variety of system issues
can cause diagnostic failure.
5. If the problem remains, contact your IBM technical-support representative.
Related Links:
IBM Support Website
v
Latest level of DSA
v
Latest level of BMC/IMM
v
409-900-000 • 409-903-000
Appendix C. DSA diagnostic test results
951